[TEST] Visual, PSNR, SSIM, VQM tests with RV9, XviD ,DivX, WMV9 ...

[Sagittaire]

Two Towers (Trailer 178 sec)

~~ PSNR, SSIM, VQM Test ~~



| Codec | P-Process | Settings | Size | PSNR | SSIM | VQM |
|------------------|-----------|------------|---------|---------|----------|----------|
| Xvid Devapi4 | No | 570 kbps | 640x272 | 40.7359 | 0.961267 | 0.855857 |
| Divx Kaukura | No | 570 kbps | 640x272 | 40.9452 | 0.963071 | 0.816827 |
| VMV9 | No | 570 kbps | 640x272 | 41.0462 | 0.964233 | 0.810412 |
| RV9 | | 570 kbps | 640x272 | 41.2712 | 0.966207 | 0.809199 |
|------------------|-----------|------------|---------|---------|----------|----------|
| Xvid Devapi4 | No | 715 kbps | 640x272 | 41.7980 | 0.968348 | 0.758332 |
| Divx Kaukura | No | 715 kbps | 640x272 | 41.8806 | 0.968601 | 0.742593 |
| VMV9 | No | 715 kbps | 640x272 | 42.4004 | 0.970695 | 0.717179 |
| RV9 | | 715 kbps | 640x272 | 42.2920 | 0.971928 | 0.723943 |
|------------------|-----------|------------|---------|---------|----------|----------|
| Xvid Devapi4 | No | 950 kbps | 640x272 | 43.0887 | 0.975330 | 0.655166 |
| Divx Kaukura | No | 950 kbps | 640x272 | 43.0920 | 0.974714 | 0.653429 |
| VMV9 | No | 950 kbps | 640x272 | 43.5733 | 0.976320 | 0.640811 |
| RV9 | | 950 kbps | 640x272 | 42.9970 | 0.975566 | 0.656301 |
|------------------|-----------|------------|---------|---------|----------|----------|



Download and compare ... no comment is necessary ...



~~ Configurations codecs used ~~

WMV9
Adjustments: Maximum quality + Complex + 2 pass + Force Post Process Mode = 0


RV9 EHQ
Adjustments: EHQ 85 + MSL 60 sec + 2 pass


XviD Devapi4
Adjustments: Ultra high + h263 + vhq4 + chroma motion + bframe 1/150/75 + 2 pass


DivX Pro Kaukuru
Adjustments: Bframe + Slowest + Bitrate Modulation "Slow Movie" + 2 Pass


~~ Script AVS ~~

Encodage

source=Mpeg2Source("D:\Mes dossiers\B.A\Les deux tours\azerty.d2v")
source=Crop(source,24,88,-24,-72)
source=BicubicResize(source,640,272,0,0.6)
return source


Test PSNR

# --> Video Opening <--

source=AviSource("D:\Mes dossiers\B.A\Les deux tours\Encodage.avs")
source=ConvertToYV12(source)

video=AviSource("D:\Mes dossiers\B.A\Les deux tours\Kaukura-950.avi")
video=ConvertToYV12(video)

# --> PSNR analysis <--
CompareYV12(video,source,"","PSNR-Kaukura-950.txt")


Test SSIM

# --> Video Opening <--

source=AviSource("D:\Mes dossiers\B.A\Les deux tours\Encodage.avs")
source=ConvertToYV12(source)

video=AviSource("D:\Mes dossiers\B.A\Les deux tours\Kaukura-950.avi")
video=ConvertToYV12(video)

# --> PSNR analysis <--
return SSIM(source,video,"SSIM-Kaukura-950.txt")


Test VQM

# --> Video Opening <--

source=AviSource("D:\Mes dossiers\B.A\Les deux tours\Encodage.avs")
source=ConvertToYV12(source)

video=AviSource("D:\Mes dossiers\B.A\Les deux tours\Kaukura-950.avi")
video=ConvertToYV12(video)

# --> PSNR analysis <--
return VqmCalc(source,video,"VQM-Kaukura-950.txt")



Harry Potter II (Trailer 128 sec)

~~ PSNR, SSIM, VQM Test ~~



| Codec | P-Process | Settings | Size | PSNR | SSIM | VQM |
|---------------|-----------|----------|---------|---------|----------|----------|
| DivX SBC | No | 570 kbps | 576*240 | 44.5105 | 0.969443 | 0.745271 |
| Xvid Devapi4 | No | 570 kbps | 576*240 | 44.8596 | 0.972988 | 0.696618 |
| Divx Manihi | No | 570 kbps | 576*240 | 44.9865 | 0.971749 | 0.685860 |
| VMV9 | No | 570 kbps | 576*240 | 45.1909 | 0.974026 | 0.671028 |
| RV9 | | 570 kbps | 576*240 | 45.4853 | 0.975972 | 0.672990 |
|---------------|-----------|----------|---------|---------|----------|----------|
| DivX SBC | No | 715 kbps | 640x272 | 45.0649 | 0.971749 | 0.705900 |
| Xvid Devapi4 | No | 715 kbps | 640x272 | 45.5277 | 0.975403 | 0.651688 |
| Divx Manihi | No | 715 kbps | 640x272 | 45.6118 | 0.975643 | 0.644280 |
| VMV9 | No | 715 kbps | 640x272 | 45.7738 | 0.976034 | 0.630541 |
| RV9 | | 715 kbps | 640x272 | 45.9816 | 0.977080 | 0.644326 |
|---------------|-----------|----------|---------|---------|----------|----------|
| DivX SBC | No | 950 kbps | 720x288 | 45.5201 | 0.975122 | 0.641671 |
| Xvid Devapi4 | No | 950 kbps | 720x288 | 46.0449 | 0.978425 | 0.594206 |
| Divx Manihi | No | 950 kbps | 720x288 | 46.1166 | 0.978254 | 0.591982 |
| VMV9 | No | 950 kbps | 720x288 | 46.2447 | 0.978799 | 0.584267 |
| RV9 | | 950 kbps | 720x288 | 46.4354 | 0.978577 | 0.623178 |
|---------------|-----------|----------|---------|---------|----------|----------|





~~ Configurations codecs used ~~

WMV9
Adjustments: Maximum quality + Complex + 2 pass


RV9 EHQ
Adjustments: EHQ 85 + MSL 60 sec + 2 pass


XviD Devapi4
Adjustments: Ultra high + h263 + vhq4 + chroma motion + bframe 1/150/75 + 2 pass


DivX Pro Manihi
Adjustments: Bframe + Slowest + Bitrate Modulation "Slow Movie" + 2 Pass

DivX3 SBC ffvfw
Adjustments: SADT + chroma motion + 3 Pass



~~ Script AVS ~~

Encodage

source=Mpeg2Source("D:\Mes dossiers\B.A\Harry Potter\azerty.d2v")
source=Crop(source,18,76,-18,-76)
source=BicubicResize(source,640,272,0,0.6)
source=Convolution3D(source,0,3,4,3,4,3,0)
source=ConvertToYUY2(source)
return source


Test PSNR

# --> Video Opening <--

source=Mpeg2Source("D:\Mes dossiers\B.A\Harry Potter\azerty.d2v")
source=Crop(source,18,76,-18,-76)
source=BicubicResize(source,640,272,0,0.6)
source=Convolution3D(source,0,3,4,3,4,3,0)
source=ConvertToYUY2(source)

video=AviSource("D:\...\Kauehi.avi")
video=ConvertToYUY2(video)

# --> PSNR analysis <--
Compare(video,source,"","Kauehi.log")


Test PSNR XviD

# --> Video Opening <--

source=Mpeg2Source("D:\Mes dossiers\B.A\Harry Potter\azerty.d2v")
source=Crop(source,18,76,-18,-76)
source=BicubicResize(source,640,272,0,0.6)
source=Convolution3D(source,0,3,4,3,4,3,0)
source=ConvertToYUY2(source)

video=AviSource("D:\...\XviD.avi")
video=Trim(video,1,0)
video=ConvertToYUY2(video)

# --> PSNR analysis <--
Compare(video,source,"","XviD.log")


Test PSNR RV9

# --> Video Opening <--

source=Mpeg2Source("D:\Mes dossiers\B.A\Harry Potter\azerty.d2v")
source=Crop(source,18,76,-18,-76)
source=BicubicResize(source,640,272,0,0.6)
source=Convolution3D(source,0,3,4,3,4,3,0)
source=DeleteFrame(source,0)
source=ConvertToYUY2(source)

video=DirectShowSource("D:\...\RV9.rmvb",fps=25)
video=ConvertToYUY2(video)

# --> PSNR analysis <--
Compare(video,source,"","test.log")

[bond]

interesting results
on my own test with old rv9 (non-ehq) i got values ~36.5

edit: i also got pretty low values with wmv9 vcm ~36.8

[karl_lillevold]

Very interesting!

For any PSNR measurements with RV9, please use the following in codecProperties, to disable the noisyEdgeFilter. The filter works like this: in case there are black or noisy pixels along the edge (from a crop that's slightly off, or capture noise), RV9 will copy real image edge pixels to cover these pixels. This is good for visual results, but oftentimes reduces PSNR compared to the original.

<codecProperties type="bag">
...
<noisyEdgeFilter type="bool">false</noisyEdgeFilter>
...
</codecProperties>

[Sagittaire]

@ karl_lillevold

update: all information which you wanted

The RV9 is the best ... ;-)

[temporance]

Great results!!

Watch out though: there is a problem with AviSynth's PSNR measurement as done by Compare(). It measures the PSNR for each individual frame and then calculates the arithmetic mean of the frame PSNRs to get a sequence PSNR.

What, you may say, is the problem with taking an average?

Taking the mean of the PSNR for every frame only ever gives a reliable result when all frames of the sequence are of similar complexity (approx same motion, texture, brightness etc.). If frames differ a lot, then the mean PSNR becomes squewed.

Worse, if a codec has very bad rate control, or something else causes a codec to produce variable quality throughout the sequence, then the mean PSNR as computed by AviSynth's Compare() will actually come out better :eek:.

What's the solution? The best approach is to compute one PSNR for the whole sequence by treating the sequence as if it was one huge video frame. The PSNR is then a function of the total noise energy in the whole sequence and no nasty averaging effects come into play. I tried to modify AviSynth to work like this, but I couldn't get it to compile. So if anyone knows how to make this change...

[Eric B]

Originally posted by karl_lillevold
disable the noisyEdgeFilter.
in case there are black or noisy pixels along the edge (from a crop that's slightly off, or capture noise), RV9 will copy real image edge pixels to cover these pixels.

Should this filter be used for black & white old movies ? And EHQ 80 ?

[Sgt_Strider]

@Sagittaire

I am wondering did you use 2 pass bitrate encoding with Wm9 or 1 pass at 100% quality? How about that slider on image quality? Is it set a 1 space before the first on the left or on default? Can you possibly display the time it took to encode each of those clips with the codecs? Thx

[Sagittaire]

@ Sgt_Strider

640*272 resolution ...
average fps ...

RV9 EHQ
EHQ 80 + MSL 60
5.88 fps

XviD
Ultra hight + VHQ4 + bframe + chroma ME
7.65 fps

DivX Tahanea
Slow + bframe
5.42 fps

WMV9
Complexe + Quality max
5.09 fps

[superdump]

XviD 24.06.03 Koepi
Adjustments: Ultra high + h263 + vhq4 + chroma motion + bframe (1/150/75/0) + 2 pass

Could you try with trellis and chroma optimisation enabled in the debug tab and with b-frames set to 3/150/75/0 and update your results.

Much appreciated. :)

[CruNcher]

@ Sagittaire
nice compare but your b-frame setting for Koepi was not well chosen 2/150/75/0 would be much better, and also to compare devapi3 vs rv9 is not fair you should better have compared with devapi4 + gmc and trellis quant, 3 warppoint gmc + vhq is working in their now "thx to syskins work" and gives a little psnr boost.

i'll post my new Compare in the XviD forum no rv9 was tested yet :)

[RadicalEd]

make sure you have postprocessing off in the kauehi decoder, last time I checked (which was awhile ago), DXN managed to sneak the post somewhere in vfw decoding >\

[Sagittaire]

@ superdump

Source Bugsan unite-video


xvid koepi 24-06-03, 2pass: 20000ko (650kbps)
avisynth 2.52
Mpeg2Dec3 1.08
FFdshow 030103
Athlon XP 2000+
---avs------avs------avs------avs------avs------avs---
mpeg2source("E:\Ripping\lotrcutted\lotr.d2v",idct=7)
Crop(4,80,-4,-80)
BicubicResize(640,256,0,0.5)
---avs------avs------avs------avs------avs------avs---
clip1 = mpeg2source("E:\Ripping\lotrcutted\lotr.d2v",idct=7)
.Crop(4,80,-4,-80).BicubicResize(640,256,0,0.5).ConvertToYUY2()
clip2 = directshowsource("xvid_XX.avi",fps=25).ConvertToYUY2()
Compare(clip1,clip2,"YUV","psnr.log")
---avs------avs------avs------avs------avs------avs---
info:
altcc-best = "altcc-h-150-100-50"
exotic = "mpeg vhq4 bf1 cm cmop altcc-best"
pp4 = "ffdshow nic pp4 strength 256"

|---------|---------|
| K140503 | K240603 |
|-----------------------|---------|---------|
| default | 41.9028 | 41.8955 |
| default pp4 simple | 42.2854 | 42.1851 |
| default pp4 walken | - | 42.2744 |
| default pp0 walken | - | 41.8940 |
|-----------------------|---------|---------|
| chroma motion | 41.9698 | 41.9864 |
| global motion comp | 41.8826 | 41.8799 |
| quarter pixel | 41.5587 | 41.5599 |
| lumi masking | 41.8397 | 41.8423 |
| chroma optimiser | 41.8964 | 41.9019 |
| trellis R-D 0 | 41.8684 | 41.8855 |
| mpeg quantization | 42.0297 | 42.0356 |
|-----------------------|---------|---------|
| VHQ 0 bf -1 | 41.9028 | 41.8955 |
| VHQ 1 bf -1 | 42.0766 | 42.0887 |
| VHQ 2 bf -1 | 42.1032 | 42.1782 |
| VHQ 3 bf -1 | 42.1491 | 42.2292 |
| VHQ 4 bf -1 | 42.2199 | 42.3316 |
| | | |
| VHQ 0 bf 0 | | 41.8946 |
| VHQ 1 bf 0 | | 42.1150 |
| VHQ 2 bf 0 | | 42.2359 |
| VHQ 3 bf 0 | | 42.2522 |
| VHQ 4 bf 0 | | 42.3327 |
| | | |
| VHQ 0 bf 1 | | 42.0270 |
| VHQ 1 bf 1 | | 42.2614 |
| VHQ 2 bf 1 | | 42.3151 |
| VHQ 3 bf 1 | | 42.3338 |
| VHQ 4 bf 1 | | 42.4166 |
|-----------------------|---------|---------|
| bf0 default | 41.9028 | 41.8946 |
| bf1 default | 41.8666 | 42.0270 |
| bf2 default | 41.7618 | 41.9992 |
| bf3 default | 41.7480 | 42.0043 |
| bf4 default | 41.7471 | 41.9937 |
|-----------------------|---------|---------|
| altcc-default | 42.1407 | 42.1330 |
| altcc-best | 42.3478 | 42.3417 |
|-----------------------|---------|---------|
| exotic | 42.6715 | 42.9207 |
| exotic pp4 simple | 43.0756 | 43.2172 |
| exotic pp4 walken | - | 43.2652 |
|-----------------------|---------|---------|

Koepi 240603 bframe test:

|---------|---------|---------|---------|---------|
| 1 - 150 | 1 - 190 | 2 - 150 | 2 - 190 | 3 - 150 |
|---------------|---------|---------|---------|---------|---------|
| bf thresh -40 | 41.9154 | 41.9148 | 41.9160 | | |
| bf thresh -20 | 41.9632 | 41.9542 | 41.9632 | | |
| bf thresh 0 | 42.0270 | 42.0494 | 41.9992 | | 42.0043 |
| bf thresh 10 | 42.0545 | 42.0750 | 42.0118 | | |
| bf thresh 20 | 42.0500 | 42.0745 | 41.9598 | | |
| bf thresh 30 | 42.0407 | 42.0689 | 41.9061 | | |
| bf thresh 40 | 42.0286 | 42.0655 | 41.8842 | | |
| bf thresh 60 | 42.0125 | 42.0318 | 41.8165 | | |
| bf thresh 80 | 42.0007 | 42.0225 | 41.7905 | | |
| bf thresh 90 | 41.9973 | 42.0103 | 41.7871 | | |
|---------------|---------|---------|---------|---------|---------|
| bf offset 0 | 41.9844 | | | | |
| bf offset 25 | 41.9887 | | | | |
| bf offset 50 | 42.0181 | | | | |
| bf offset 75 | 42.0270 | 42.0494 | 41.9992 | | 42.0043 |
| bf offset 100 | 42.0115 | | | | |
| bf offset 125 | 42.0104 | | | | |
|---------------|---------|---------|---------|---------|---------|


The defaut reglage for bframe 1/150/75/0 is very good reglage for the bframe for Koepi 24/06/03 and in conclusion for PSNR ... ;-)

[CruNcher]

@ RadicalEd

pp was off but kauehi simple profile still is better then XviD its true in terms of PSNR or a decoder trick but i doubt that :rolleyes:

[Sigmatador]

i think 1/150/75/0 is good (i prefer 1/150/0/10 :D )

It's a good idea to use trellis for PSNR compare but not chroma optimizer (it might increase video quality but not PSNR ^^ )

A question for PSNR nerd: is there a difference between psnr4avi results and avisynth's compare() ones ?


"3 warppoint gmc + vhq is working in their now "thx to syskins work" and gives a little psnr boost." !!!!!! i want it ^^, going to compile the dev-api4 branch

[Sgt_Strider]

Originally posted by Sagittaire
@ Sgt_Strider

640*272 resolution ...
average fps ...

RV9 EHQ
EHQ 80 + MSL 60
5.88 fps

XviD
Ultra hight + VHQ4 + bframe + chroma ME
7.65 fps

DivX Tahanea
Slow + bframe
5.42 fps

WMV9
Complexe + Quality max
5.09 fps

Thx Sagittaire, but you still haven't answer my question. Did you just two pass bitrate encoding for WM9 or 1 pass VBR?

[Sagittaire]

WMV9
Adjustments: Maximum quality + Complex + 2 pass
PSNR: The WMV9 is really very powerful in this test. It is a codec which "macroblock" practically never and this some is the bitrate.

I make edit in principal post

Sorry for my english ... ;-)

[temporance]

Originally posted by CruNcher
pp was off but kauehi simple profile still is better then XviD its true in terms of PSNR or a decoder trick but i doubt that :rolleyes: If you have any doubts, you could try using different decoders - just change the fourCC between XVID and DX50.

I believe Kauehi has configurable postprocessing - could be interesting to see if this improves PSNR. Might need to use DirectShowSource() in AviSynth if the VfW decoder doesn't do PP.

[temporance]

An update on AviSynth PSNR averaging: I rewrote part of AviSynth Compare() to create a more reliable sequence PSNR. It's output to the logfile as "Overall PSNR".

I posted my code on the AviSynth development forum here (http://forum.doom9.org/showthread.php?s=&threadid=57744) in the hope that the AviSynth guys like what I've done and incorporate it into their next release.

[Sgt_Strider]

Originally posted by Sagittaire
WMV9
Adjustments: Maximum quality + Complex + 2 pass
PSNR: The WMV9 is really very powerful in this test. It is a codec which "macroblock" practically never and this some is the bitrate.

I make edit in principal post

Sorry for my english ... ;-)

If you use 2 pass for your encodings, shouldn't you list your bitrate that you use?

[Sagittaire]

@ Sgt_Strider

Video : 570 Kbps 576x240
Video : 715 Kbps 640x272
Video : 950 Kbps 720x288

Infomrmation in principal thread ........ !?

[temporance]

Is that the bitrate that you set in the codec, or the actual bitrate that the codec decided to produce? Were all the codec's filesizes similar?

WMV9 especially is known to exceed the bitrate that you tell it.

[Sagittaire]

The difference between the file is less 1% for each encoding ...

A variation of 1% for the file by no means variation of 1% on the PSNR ...

I arranged myself so that the variation is favorable to XviD ...:devil:

I did not want to have problem with the fans of XviD ...:D

[Sirber]

:D

XviD fans can be dangerous... :eek: Just kidding :D

[Sgt_Strider]

Originally posted by Sagittaire
@ Sgt_Strider

Video : 570 Kbps 576x240
Video : 715 Kbps 640x272
Video : 950 Kbps 720x288

Infomrmation in principal thread ........ !?

lol oops, I miss it :o

[temporance]

@Sagittaire,

I was just trying to replicate some of your results and it seems to me that there are much better settings for at least one of the codecs...

Instead of:DivX Pro Kauehi
Adjustments: Bframe + Slowest + Bitrate Modulation "Slow Movie" + 3 Pass I would recommend:
B frame + slowest + bitrate modulation slider to right (+0.2) + 3 pass + new (slider all the way to right) Psychovisual in passes 2 and 3.

To my eye this gives much better results than your settings (the new PV makes all the difference). PSNR is slightly better, but the video looks much nicer.

[Sagittaire]

the objective in a test PSNR is to obtain the best possible PSNR. But it is true for a better visual quality DivX Labs recommend bframe + Slowest + BM 0.25 + New Psy + 3 Pass. but the PSNR decrease a little.

[bond]

Originally posted by Sagittaire
the objective in a test PSNR is to obtain the best possible PSNR. But it is true for a better visual quality DivX Labs recommend bframe + Slowest + BM 0.25 + New Psy + 3 Pass. but the PSNR decrease a little.only the visual quality counts ;)

and really, divx labs recommend "officially" to use bm 0.25 (interesting i always used 0.2 in my tests)?
where did you read that?

[temporance]

I don't think it makes a lot of difference whether you set 0.2 or 0.25. The main thing is to set a positive value.

I found that new PV can actually increase PSNR. So if PSNR's what you want then it's worth a try.

[bond]

Originally posted by temporance
[B]I don't think it makes a lot of difference whether you set 0.2 or 0.25. The main thing is to set a positive value.hm, perhaps with 0.25 we get the same output as with original 2-pass (which isnt the case with 0.2)?

[temporance]

Originally posted by bond
hm, perhaps with 0.25 we get the same output as with original 2-pass (which isnt the case with 0.2)? Is that a good thing? I mean, do you still find original 2-pass to be better??

[bond]

Originally posted by temporance
Is that a good thing? I mean, do you still find original 2-pass to be better?? yes, because it shows more details in still scenes and i dont have the time to make three passes (and i think that 2 passes should be enough for a good codec to encode a movie normally)

but why do you think n-th pass is better?


[as this is totally off topic perhaps a mod can move the last posts to this thread (http://forum.doom9.org/showthread.php?s=&threadid=52163&perpage=20&pagenumber=2) :D ]

[temporance]

see you on that thread :)

[Sagittaire]

Update principal post ...

~~ Visual Test ~~

Video Encode 576*240 570 Kbps

http://jfl1974.free.fr/Video/XviD-Koepi.rar
http://jfl1974.free.fr/Video/RV9.rar
http://jfl1974.free.fr/Video/WMV9.rar
http://jfl1974.free.fr/Video/Manihi.rar
http://jfl1974.free.fr/Video/DivX3-ffvfw.rar

Download and compare ... no comment is necessary ...

[midiguy]

The WMV9 clip looks best to me. The RV9 clip looks less detailed... the background appears very washed out.

[Sagittaire]

Update with Manihi Codec: the codec is more better than Kauehi codec. For me Manihi Codec is the best isocompliant-MPEG4 codec now ... sorry for the XviD fan ... ;)

[Nic]

Thanks for the supplying the clips and all your work Sagittaire :)
The Kauehi.rar link doesn't work for me ? Might be a problem my end, but I get a 404 when I click on it.

[Sagittaire]

Update with DivX SBC for the fans: with 570 Kbps and 40% compressibility the encodage with this codec is'nt bad ... DivX SBC is'nt dead with ffvfw ... ;-)

[LordIntruder]

I don't think it makes a lot of difference whether you set 0.2 or 0.25. The main thing is to set a positive value.

You mean you always encode at 0.25 or at least with positive values in the bitrate modulation? I wonder why, this option is here to let the user to force the codec just a bit to increase bitrate in high or slow motion.

If you have a high motion movie it is recommended to use positive values and on the contrary if you have a slow motion movie using negative values.

However the human eye is much more receptive to quality in slow motion as in high motion everything moves so fast you can't perceive a loss in details where you can in scenes that don't move or just a bit.

So it's not easy to determine the better choice to get the better rip except making several full encodes and compare but... Personnaly I chose to support slow motion in most of my encodes movies (usually -0.1) and results are very nice on both slow and high motion.

[Teegedeck]

Great test! Thanks for your work!

info:
altcc-best = "altcc-h-150-100-50"

Does that mean alt.cc with the 'high curve aggression'-setting?

[Animaniac]

I'd like to the same test with 1000 kbps and 1200 kbps bitrates... :o

[Teegedeck]

Where could we download that trailer you used as a source?

[Sagittaire]

It's an trailer of Harry Potter's DVD ... 75 Mo

[Teegedeck]

I just wondered because the highest resolution trailer I've found in the net has a 480x260-resolution. :o Is it a trailer from the DVD?

[Sagittaire]

it's a trailer from HP II DVD 720*576 high quality with no entrelacing ...

[h9903209]

Thx for your testing. Is that possible to include the speed of encoding? How long does each one take to finish the encode?

[Sagittaire]

Update first post ...

XviD devapi4 test
PSNR, SSIM, VQM test and graph
DivX3 SBC ffvfw with Post-Precessing test

RV9 EHQ ... winner

Here my conclusions on these test:

The RV9 is the best codec has weak resolution and weak bitrate far in front of the others.

DivX Manihi and XviD Devapi4 are codecs which are almost equivalent.

WMV9 is the most powerful codec in high resolution and high bitrate.

DivX SBC is still a very good codec and it is still very powerful in encodings with a compressibility superior to 50%.

[SeeMoreDigital]

Sagittaire

How about doing some more PSNR tests, using the same 'video only' input file, but this time generating the same output file size.

And all without cropping and resizing?

I think your results will be quite a bit different!

Cheers

[CruNcher]

@Sagittaire

it's ridicilous, you still do this XviD Ultra High vs DivX Slowest compare thats flawed you need 3x more with Slowest Me in DivX then with Ultra High in XviD, wich is like Standard in DivX and you only win for this 3x slowness 0.1 dB, how you can say this makes DivX the most iso complaint Mpeg4 Codec, sorry can't follow you here. And use compareyv12 instead of converttoyuy2

[Sagittaire]

It's a quality test ... not a speed test ...

But is true the Manihi is slowest than XviD in slow mode but the standard mode is very good too (download DivX5 for demo).

The best codec is not Manihi but RV9 EHQ ...

[kastro68]

@Sagittaire

Thanks for the tests. You have obviously spent a great deal of time and effort.

I've been using RV9 for low bitrate anime encoding [3.5hrs into 700megs] ever since I found out it was free. I'm just shocked that it gives higher psnr at higher bitrates too.



@animaniac
I'd like to the same test with 1000 kbps and 1200 kbps bitrates...

You can use linear interpolation for these.[Assumes that the variance of the white noise is constant and normally distributed]

At 1000kbps RV9 and wmv9 should be on par. From the graphs, it looks like WMV9 would dominate at higher bitrates.

edit:typo

[vinouz]

I may seem an idiot, but what means SSIM & VQM ? What are they based on ?

[kilg0r3]

1. @vinouz
you can google can you?
http://www.its.bldrdoc.gov/n3/video/Default.htm
http://www.cns.nyu.edu/~zwang/files/research/ssim/

2.
@sagitaire
could you add the fps or encoding time? And, it would be nie to see if qpel adds something to xvids performance.

3.
what method did the ct magazine use to meseure the codec quality?

[Lefungus]

[Edited]: I need to think before posting

[Sagittaire]

PSNR: high is better quality max = 105
SSIM: high is better quality max = 1
VQM: small is better quality max = 0

Qpel and GMC decrease PSNR and quality test ...

[SeeMoreDigital]

Although this thread has been both informative and useful, and long may it continue to be. For me, it's confirmed that PSNR etc tests are not the diffinative way (or answer) to measuring how good an encode/codec is!

Reading the entire thread, reveals refreshing opinions regarding the codecs used and how we all have different methods and views about quality.

I'm happy that we are not all 'Mini Me' clones with regard to our opinions.

[Sagittaire]

for the eyes tests ... ;-)

~~ Visual Test ~~

Video Encode 576*240 570 Kbps

http://jfl1974.free.fr/Video/XviD-570.rar
http://jfl1974.free.fr/Video/RV9-570.rar
http://jfl1974.free.fr/Video/WMV9-570.rar
http://jfl1974.free.fr/Video/Manihi-570.rar
http://jfl1974.free.fr/Video/DivX5-570.rar
http://jfl1974.free.fr/Video/DivX3-570.rar


Download and compare ...

[SeeMoreDigital]

Thanks Sagittaire,

I downloaded and unpacked the following video only 'visual test' clips: -

RMV9-570kbps (8,939KB or 8.729MB)
Manihi-570kbps (9,008KB or 8.796MB)
XviD-570kbps (9,160KB or 8.945MB)
WMV9-570kbps (9'170KB or 8.955MB)

And this confirms exactly what I've been saying for many months now. Which is when you take the same 'video only' input file and encode at the same bitrate with different codecs, you obtain totally different file sizes!

OK the file size difference may be small (the clip is only 128 seconds long) but if you increase the bitrate setting of the other codecs to match the largest file size (which is always WMV9) then the 'visual' and 'PSNR etc' results will be different again.

This then means, that when you take this anomaly into consideration, the Real Media encode is even better!

[Sagittaire]

a difference in size of 1% does not mean at all profit of 1% of PSNR for example ... make the test ...

and to visually make the difference between an encoding with 565 Kbps or 575 Kbps is impossible ...

The test are right and the small difference in size of encoding do not affect results ...

[SeeMoreDigital]

Yes, like I said the file size differences are small.

However, if you encoded the entire Harry Potter 2 DVD (which is 9261 seconds long), you would have a 16.32MB difference between the RMV9 and WMV9 file sizes. Which is 2.6%!

And when encoding at low bitrates it's important to utilize every last MB!

RMV9 still comes out on top.... Karl will be pleased!

Cheers

[Lefungus]

Originally posted by SeeMoreDigital

However, if you encoded the entire Harry Potter 2 DVD (which is 9261 seconds long), you would have a 16.32MB difference between the RMV9 and WMV9 file sizes. Which is 2.6%!


Why ?
It may as well still have a 1% difference, like the trailers.

[Sagittaire]

2% size difference it's 0.1 dB PSNR difference for my test (with my PSNR graphic) ...

The test are right and the small difference in size of encoding do not affect results and conclusion ...

[temporance]

Originally posted by Sagittaire
Update first post ...

XviD devapi4 test
PSNR, SSIM, VQM test and graph
DivX3 SBC ffvfw with Post-Precessing test

RV9 EHQ ... winner

Sagittaire,
If you're using "HVS-aware" metrics, then IMHO you really should try enabling various codec's psychovisual enhancements. E.g. Manihi's new psychovisual enabled (fast or slow) and xvid's luma masking.

PSNR is just no good for picking up the real improvements that these tools bring.

Edit: I don't know how you find the time to do all these tests and post them here: well done!!! All I manage to do is try a few things, tests, etc and get a general impression of things!

[SeeMoreDigital]

Lefungus

Originally posted by Lefungus
Why ?
It may as well still have a 1% difference, like the trailers.

Please get out your calculator. The difference in file size between the RMV9 encoded trailer and the WMV9 encoded trailer is nearer 2.6%, not 1%.

I like your sig name!

[Sagittaire]

A new test with the trailer "two towers". Desactivate the PP of the WMV9 results as well visual as in the test are much better. The PP of the WMV9 is especially has my effective opinion for bitrates very very weak ...

With this trailer the WMV9 and RV9 EHQ detache largely of the others codecs. The RV9 being most powerful A weak bitrate and the WMV9 for the tops bitrates ...

Kaukuru slightly shows superior in the tests but also visually compared to XviD. But the extreme slowness of DivX for a weak variation on the level of quality gives the advantage to XviD for Iso-MPEG4 codecs ...

Download 715 Kbps trailers what corresponds has a bitrate for a movie of 2H00 on a cdr of 80 min. I awaits your impressions, comments and insults impatiently ... ;)

[SeeMoreDigital]

Sagittaire
... Download 715 Kbps trailers what corresponds has a bitrate for a movie of 2H00 on a cdr of 80 min...
For newbie coders who require a little more info. Here is a brief list of the 'Bitrate per min per 700MB CD-R'

For 60mins A/V per CD-R with audio at 64kbps. Set the video to 1556kbps
For 60mins A/V per CD-R with audio at 96kbps. Set the video to 1524kbps
For 90mins A/V per CD-R with audio at 64kbps. Set the video to 1016kbps
For 90mins A/V per CD-R with audio at 96kbps. Set the video to 984kbps
For 120mins A/V per CD-R with audio at 64kbps. Set the video to 746kbps
For 120mins A/V per CD-R with audio at 96kbps. Set the video to 714kbps
For 150mins A/V per CD-R with audio at 64kbps. Set the video to 584kbps
For 150mins A/V per CD-R with audio at 96kbps. Set the video to 552kbps

*Mp3 audio at 64kbps does not actually sound that good. 96kbps is generally accepted as being the absolute minimum.

The longest film I have encoded to fit onto a 700MB CD-R is Star Wars 2 at 137mins (in full anamorphic 720x576).

Love it or hate it - DivX still looks better than VHS or S-VHS tape - even at low bitrates!

With regard to DivX Kaukuru it is DivX's finest codec so far. However, this version has even more user settings than previous versions, which in my opinion will confuse many users. And as such cause much argument regarding 'what setting is the best etc to use'!

I've generated many low bitrate tests using all manor of Kaukuru's settings. But at the end of the day I would dump the lot and just set up the codec to write/read MV files - when you do this the encoder 'goes into warp drive'!

[Sagittaire]

Update WMV9 Harry Potter II test with PP desactived ...

[Sgt_Strider]

Hey Saggitaire, I'm wondering have you tried going even higher for the bitrates? I'm talkin about at least 1200 and I'm wondering if you're going to give that a thought?

[temporance]

Kaukuru slightly shows superior in the tests but also visually compared to XviD. But the extreme slowness of DivX for a weak variation on the level of quality gives the advantage to XviD for Iso-MPEG4 codecs ... Just wondering, Saggitaire, why you don't try the codec's psychovisual tools now that you are measuring VQM and SSIM? I for one would be very interested in the results! :)

[superdump]

dev-api-3's lumimasking isn't very good and I would not recommend using it. If you insist on using it only use it during the second pass. dev-api-4's adaptive quantisation is currently very limited and also not very good. I would not recommend using that either. Lefungus is working on something which could be implemented which would make adaptive quantisation much more effective.

Personally I think you should use the settings which produce the best picture quality for a given size. I don't care if you use psychovisual stuff to do that or not.

[Sagittaire]

To make tests has 1200 Kbps is not useful with traditional pixel 1:1. All the codecs are powerful with this bitrate and even DivX3 SBC is very good with this bitrate. MPEG2 VBR starts has to be good in 480*480 if the source is rather compressible ...

On the other hand that could be amusing to make tests into anamorphic pixel 16:11 to 1200 Kbps ...

[CruNcher]

@ Sagittaire

why you dont use Global Motion Compensation for XviD in your theatrical trailer comparesion ?, it would get better result for this kind of content with it. What your testing here btw is the hardest situation for a codec and one that apparently is only given for this kind of trailers and sure RV9 looks nicer in this High Motion Situation @ low bitrates but RV9 stays allways like that because of the inloop filtering (soft washy effect) even in Low Motion Scenes and even @ higher bitrates so you are excepting the lose of detail in every kind of situation where in Mpeg4 you only lose it in High Motion. In a theatrical trailer you notice this lack of detail in this High Motion scenes because their are no Low Motion ones that follow so your eyes adapt to this situation and it looks ugly over time, imo wmv9 is the best solution for this kind of content but the most content people encode here has most the times the same amount off High Motion Low Motion scenes so it should compensate it really well.

[Sirber]

Do you call artefacts (squares, muskito noise, etc) details? I've seen more enjoyable 1 CD movie in RV9/RA which were more detailled, more textured, bigger resolution then it's 1 CD MPEG4 equivalent...

[superdump]

Sirber: Noone considers artifacts to be details. XviD people prefer to have more details and some blocking than a slightly smoothed image and no blocking. RV9 people prefer the opposite and each opinion is perfectly reasonable. I'm an XviD person, but I think RV9 is a very powerful codec and performs better than XviD in a number of cases. For example, the trailers that Sagittaire has tested the codecs on is a tough case for most codecs. As CruNcher says, high motion and a low bitrate do not bode well with MPEG-4 codecs but they can ususally compensate with intelligent rate control and bit allocation when there are low motion scenes within the video also.

I personally think that WMV9 is between XviD and RV9 so if you are looking for something with slightly more detail than RV9 but less blocking than XviD then WMV9 would probably be a wise choice. I don't however think that (note, this opinion is based on PP being disabled) WMV9 maintains more detail than XviD but would be delighted if you could show me some pictures to prove otherwise.

RV9 also seems to be a long way ahead in cartoon encoding, but I will possibly report my findings in this area later.

What I will say is this, I'm glad we have a competitive field with a number of very good codecs. It makes for interesting conversation.

[Sigmatador]

"Do you call artefacts (squares, muskito noise, etc) details?" don't be so annoying :devil:

[Mei]

Personally I think that DivX/XviD are the best because they are not propriatery and that we can use it in MPEG-4 Stream File (.MP4).
Also, buy Terminator 2 Extreme DVD Edition, and see what the Windows Media Video 9 codec can do !
With a resolution of HDTV 1080p Anamorphic (1920x1080 but 1440x816 effective) you can see a real pure image quality... Okay a ~6,5Mbps Bitrate... But it show that WMV9 is really good in high bitrate... Not the case of other MPEG-4 codec...

[Sirber]

"Do you call artefacts (squares, muskito noise, etc) details?" don't be so annoying :devil::D Just bringing some gazoline on the fire... ;)

[midiguy]

Originally posted by superdump
I don't however think that (note, this opinion is based on PP being disabled) WMV9 maintains more detail than XviD but would be delighted if you could show me some pictures to prove otherwise.

I've done quite a bit of testing with both, and have noticed something with wmv9. It can be just as detailed (or sometimes even more detailed) than XviD if the video is encoded very well (if it is very compressible it helps) or if you use a high bitrate. The thing is, once you get to lower bitrates where artifacts (blocking, noise, etc.) are introduced, the two codecs start to behave differently. WMV9 tries to smooth out the artifacts (which is done internally I believe), while XviD won't mind giving you a few blocks here and there to preserve the details. But yeah, for higher bitrate stuff, wmv9 and XviD behave quite similarly (for HDTV resolution material, or very high bitrates, I think WMV9 is the more powerful codec). For lower bitrates (or sources that are very uncompressible), it is really just a matter of preference.

[Sigmatador]

i'm not confident that the wmv9 can't be beaten by the rv9 at HDTV encode, no ?

[SeeMoreDigital]

Originally posted by Sigmatador
i'm not confident that the wmv9 can't be beaten by the rv9 at HDTV encode, no ?

I'm quite confident that WMV9 can be beaten by RV9!

I recently carried out some 'video only' trancode tests using the WMV9 high-def (1440x816) Terminator trailer as a source. And I'm happy to say that the 1440x816 RV9 encode looked every bit as good as WMV9 source. And used far less processing power on replay!

My 1440x816 DivX encode was a different story. Although the image looked very good it was far too choppy to be an enjoyable watch.

Cheers

[superdump]

Originally posted by midiguy
I've done quite a bit of testing with both, and have noticed something with wmv9. It can be just as detailed (or sometimes even more detailed) than XviD if the video is encoded very well (if it is very compressible it helps) or if you use a high bitrate. The thing is, once you get to lower bitrates where artifacts (blocking, noise, etc.) are introduced, the two codecs start to behave differently. WMV9 tries to smooth out the artifacts (which is done internally I believe), while XviD won't mind giving you a few blocks here and there to preserve the details. But yeah, for higher bitrate stuff, wmv9 and XviD behave quite similarly (for HDTV resolution material, or very high bitrates, I think WMV9 is the more powerful codec). For lower bitrates (or sources that are very uncompressible), it is really just a matter of preference.

Well said midiguy. I think you've nailed it. :) WMV9 scales incredibly well on high resolution and very high bitrate stuff that few PCs can play back. Lol. I have seen some of these HDTV demo encodes from the microsoft site and they look _awesome_.

Thanks for your post midiguy, you've added a bit of clarity to this very interesting but slightly muddled discussion. :)

[Sigmatador]

Originally posted by SeeMoreDigital
I'm quite confident that WMV9 can be beaten by RV9!

I recently carried out some 'video only' trancode tests using the WMV9 high-def (1440x816) Terminator trailer as a source. And I'm happy to say that the 1440x816 RV9 encode looked every bit as good as WMV9 source. And used far less processing power on replay!

My 1440x816 DivX encode was a different story. Although the image looked very good it was far too choppy to be an enjoyable watch.

Cheers


For HDTV, these are two good codec, but RV9 still have better technologies. But, with some few PSNR/SSIM test, with a high compressibility (above 60%) the WMV9 have better results (but not a huge gap) with very high compressibility (above 80%), there's a more noticiable difference (and the xvid have better results than RV9 too, but less than WMV9)

The problem is, such compressibility are very difficult to reach with HDTV resolution, so the RV9 can be a good solution for these encoding.

A word about the wmv9 trailer, yes they look pretty good, but it's only thanks to the high resolution. Look the video with the real resolution (no undersizing, use panscan if you don't have a 22" screen :D ) you can see som very ugly blocks and gradients that a good old 720*yyy mpeg2 doesn't have. So don't forget what we have learned with mpeg-4 codec, a high reolsution is stupid without a high bitrate ^^ (that's why i think the rv9 can beat wmv9 from the same source (not the already wmv9'ed ^^) for HDTV )

[temporance]

Originally posted by superdump
WMV9 scales incredibly well on high resolution and very high bitrate stuff that few PCs can play back. Lol. I have seen some of these HDTV demo encodes from the microsoft site and they look _awesome_.HDTV is nothing special, just a (much) higher resolution that we're used to. Everything else scales (including the bitrate). The only thing that perhaps doesn't scale linearly is motion estimation - a motion that covers 20 pixels/frame at SDTV is going to cover something like 50 pixels/frame at HDTV. So motion estimation is going to be very important at HDTV res.

There are two main reasons why WM9 looks good at HDTV... First that Microsoft have $$$ and connections within the movie industry and so can access super high-quality uncompressed, low frame rate (I saw one demo that had slomo surfing at approx 12fps), noise-free sources for their encodes. So they pick and chose content that makes very nice looking compressed video. Second, they are the first to have a decoder that can smooth out processor load. One reason why xvid HDTV doesn't look so good is that you'd need a Pentuim 4GHz because the decoder CPU load is not well managed by the xvid DS decoder.

So, apart from these factors, there is no reason why we should not think HDTV will be any different from any other video source. More pixels needs more bits, more codec CPU, more RAM bandwidth, better graphics card. But these factors affect all codecs the same - things just scale.

[Sirber]

I guess we can forget H264 HDTV :D. A 320x200 clip barely play on my 2000+...

[Sagittaire]

And Final Test with entrelaced cartoon "Corto Maltes"

http://jfl1974.free.fr/images/PSNR3.gif

~~ Visual Test ~~

Video Encode 576*288 570 Kbps

http://jfl1974.free.fr/Video/XviD-570.rar
http://jfl1974.free.fr/Video/RV9-570.rar
http://jfl1974.free.fr/Video/WMV9-570.rar
http://jfl1974.free.fr/Video/Kaukura-570.rar
http://jfl1974.free.fr/Video/DivX3-570.rar


RV9 and WMV9 ars cleary best for this source type. Download and compare.

Conclusion: "XV9" are the best

[Sirber]

Seems WMV9 is quite good at high bitrate. Also, it must be a shock for XviD fans, to get beatten by DivX :D

[Sagittaire]

It's a test with XviD Devapi4 which is much better than XviD Devapi3 of koepi ...

The WMV9 (with PP=<2) is thus really better than XviD Koepi 24/06/03 ...

The WMV9 could appear worse because of the PP which is really too strong with defaut setting (PP=4) ...

DivX Kaukuru and XviD are really very similar for quality ...

Download and compare ...

But RV9 and WMV9 qaulity for low bitrate with anime are unbeatable especially for animate ...

[temporance]

@Saggitaire,

Could you also published the uncompressed source in uncompressed form (say, huffy)? I'd like to try some different settings on various codecs,

Cheers,

[SeeMoreDigital]

Originally posted by Sigmatador
... A word about the wmv9 trailer, yes they look pretty good, but it's only thanks to the high resolution. Look the video with the real resolution (no undersizing, use panscan if you don't have a 22" screen :D ) you can see som very ugly blocks and gradients that a good old 720*yyy mpeg2 doesn't have. So don't forget what we have learned with mpeg-4 codec, a high reolsution is stupid without a high bitrate ^^ (that's why i think the rv9 can beat wmv9 from the same source (not the already wmv9'ed ^^) for HDTV )

I agree with you regarding your 'high resolution' comment. However, I also think that there's very little point for a distributor to supply vast quantities of 'standard resolution' images via the HD platform!

[temporance]

Originally posted by Sirber
Seems WMV9 is quite good at high bitrate. Also, it must be a shock for XviD fans, to get beatten by DivX :D My experiments with VQM and JND (using natural video not anime) indicate the exact opposite of this: that WMV9 is worse than xvid and DivX at high bitrates. I get a "crossover" - WMV9 is better at lower bitrates and worse at higher bitrates. At the point of the crossover, all three codecs are producing pretty mediocre results anyway!

[SeeMoreDigital]

Originally posted by temporance
My experiments with VQM and JND (using natural video not anime) indicate the exact opposite of this: that WMV9 is worse than xvid and DivX at high bitrates. I get a "crossover" - WMV9 is better at lower bitrates and worse at higher bitrates. At the point of the crossover, all three codecs are producing pretty mediocre results anyway!
I have to agree also.

In my experience, as soon as you begin to generate 720x576 DivX encodes above 950kpbs the 'visual' image looks alot better than WMV9. However, when you get to around 1500-1600kbps, the images look about the same!

[Sagittaire]

WMV9 exemple
"Terminator 2 Extreme DVD Clip" (1080p Anamorphic, 6.8 Mbps, 1:42 min, 91 MB)
BPF = 6800000/(1920 x 1280 x 24) = 0.12
O.12 for very low compressibility source thus it's high resolution but not high relative bitrate ...

MPEG2 exemple
For a MPEG2 source with same bitrate ...
BPF = 6800000/(720 x 576 x 25) = 0.65

With 720*576 and 950 kbps the RV9 EHQ and WMV9 are the best ...
BPF = 950000/(720 x 576 x 25) = 0.09
It's a very small BPF and DivX or XviD can't be the best ...

[Sigmatador]

@Sirber
currently there is no good implementation of the h.264 standard (which is not finalized yet btw). I think rv9 specs are quite closed to the h.264 ones (but i can't affirm it ^^) so let a chance to h.264 (but real networks has a big advance (dunno if it's mean something in english ^^ ))

for my personnal test with rv9 i can't see the différence between a 700mo and a 1400mo ¤_¤ rv9 compress too much, it's the most powerfull for low bitrate, quite good for medium, but too destructive for my eyes at high bitrate (and its 2passvbr is not very good because of the MSL >_<). but for low compressibility, rv9 is unbeatable ^^ (so usefull for akira ^^ )


another good point from Mei, xvid/divx are quite standard ^^ and will sure have hardware decoding everywhere in the futur, i doubt for rv9/wmv9. (i really don't care, i only use my computer ^^, but i'm not alone ;D )


ps: HD become bigger, internet connection too, let's return to the good old mpeg2 8mbit/s :D

[Gaia]

for my personnal test with rv9 i can't see the différence between a 700mo and a 1400mo ¤_¤ rv9 compress too much, it's the most powerfull for low bitrate, quite good for medium, but too destructriv for my eyes at high bitrate

I totally agree with you on this one. I once encoded very high detailed short clip of Matrix with RV9(EHQ enabled) and XviD. For XviD i used 900kbps and 1200 for RV9. XviD clip looked a lot better than RV9. RV9 clip looked far too washed out for me. RV9 is very good codec for low and midium bitrate but it's not my choice for backing up DVD's. Like a lot of people have said it's matter of taste...

Speed(if some codec is just slightly better than others) and editing capabilities are important too.

[karl_lillevold]

Originally posted by Gaia
RV9 clip looked far too washed out for me
At 1200 kbps RV9 should not look washed out, and I think you must have run into some kind of problem, either with your encoding, or playback. Do you still have your samples so I can take a look?

Co-incidentally I recently encoded a 1200 kbps 704(->860) x 364 ten minute demo with 5.1 sound that was played back on a huge projection screen in a mini-theater prior to a showing of a regular DVD in the same theater. The RV9-EHQ encode looked more detailed than the DVD, with every grain and detail showing.

[Sagittaire]

for the top bitrates (with BPF > 0.25) whole codecs are worth and the differences are weak ...

What makes the power of a codec it is encoding capacity in the weak BPF ...

Conclusion: RV9 EHQ winner

[SeeMoreDigital]

Originally posted by Sagittaire
WMV9 exemple
"Terminator 2 Extreme DVD Clip" (1080p Anamorphic, 6.8 Mbps, 1:42 min, 91 MB)
BPF = 6800000/(1920 x 1280 x 24) = 0.12
O.12 for very low compressibility source thus it's high resolution but not high relative bitrate ...

This is not correct. The actual pixel frame size of the T2 clip is 1440x816. So 6,800,000 /(1440x816x24 = 28,200,960) = 0.24

Cheers

[Sigmatador]

Originally posted by Sagittaire
for the top bitrates (with BPF > 0.25) whole codecs are worth and the differences are weak ...

What makes the power of a codec it is encoding capacity in the weak BPF ...

Conclusion: RV9 EHQ winner well why all people doesn't store music with ogg@64 :devil: maybe who use mpc@200 are very stupid :devil:
wmv9,ra8,ogg etc... are decently good at 64kbps against mpc (which is horrible ^^), but there's currently no codec who can beat mpc quality 4 (approximately 200-250kbps) (conclusions obtained with a spectrum compare). Well, audio codec is not the point, but it's just to say (illustrate^^), that the winner at extreme low bitrate isn't the winner at all !!


@Karl
i agree with Gaia, do a xvid fixq2 encode (i suppose you know how to config the codec ^^), don't care about the filesize (it may produce something between 1000 and 2000 kbps) and encode the same clip with the same filesize in rv9 (you can higher the resolution if you want, i did it) and the rv9 seems to my eyes more washed like Gaia said.

[Acaila]

well why all people doesn't store music with ogg@64 Because some of us actually like it when that music sounds close to the original... Storing music is for me all about getting transparency, which for me would be MPC @ lvl 5 (around 170 kbps) with my sound equipment.
Speech on the other hand is an entirely different story.

Well, audio codec is not the point, but it's just to say (illustrate^^), that the winner at extreme low bitrate isn't the winner at all !!
I agree, which rules out RV9 or WMV9 as a winner for me, because I would never make a DVD backup at 570 kbps, I want my backups to closely resemble the original.

[Sagittaire]

the RV9 and the WMV9 are better codecs than the Iso-MPEG4 for 570~950 Kbps in all my test and for all types of sources ...

My objective is to put between 90 min and 150 min of video on cdr of 80 min what corresponds to the bitrates selected ...

If you make your ripp with XviD in quant2 this test does not concern you however you will obtain little better quality with the WMV9 with same bitrate ...

I show you only the tests most serious possible that I make and the videos that i encode ... sorry for the fans of XviD ... but my conclusion is RV9 is winner ... that all ... :D

Make test ... :D

[Zhnujm]

@Sagittaire:
I surely dont question your results and the work you put into it, but imho a test should be made with a whole movie, not with some short clips.
Have you already tried encoding a whole movie and look if the results are the same ?

[Gaia]

Fine...

I want best possible quality with original ac3 audio track. I do 2-3 cd rips and i don't care about file sizes because i encode just for my own personal use. I don't release or spread my encodes.

So for me best codec is XviD. Sure RV9 is a winner if you want to put 150 min on one cdr.

[temporance]

@Sagittaire,

We all have our own opinions of which codec is best for us and, until someone polls the opinions of a room full of double-blind viewers, we must agree to differ.

Having said that, I'm not sure that anyone disagrees that RV9 is good when you want few artefacts at low bpp ratios. Do we at least all agree on that?

[Sagittaire]

For high bitrate on 2 or 3 cd the 720*xxx anamorphique encoding is the best quality solution and RV9 EHQ and WMV9 are better than iso-MPEG4 codec (it's my opignon confirmed by PSNR, VQM and SSIM ...)

For iso-MPEG4 encoding DivX Kaukura is little best than XviD Devapi4 and thus verry better than XviD Koepi 24/06/03 (it's my opignon confirmed by PSNR, VQM and SSIM ...)

I post demonstration with 1500 kbps 720x432 RV9 EHQ anamorphique encoding if you want ...

[Sigmatador]

@acailia
"well why all people doesn't store music with ogg@64 " -->it was ironic ^^ as you said, i just want to point that, many people like quality, and it imply a real bitrate ^^.

@Sagittaire
well, with these bitrate, you've right, wmv9/rv9 are better tha mpeg-4 iso codec, but some people save their encode on 2cd (and now 3 or 4 encodes on one dvdr, and like me, on harddisk ^^)
i've just said, these bitrate aren't the "universal bitrate that decide which codec wins ^^" (and as a xvid fan, i know how good is the rv9 codec ^^)
and don't forget the iso standard problem, hardware decoding will become more imortant in te futur. WMV9, thx to the m$ evil pawa :D, have more chance to impose his "standard".

@Zhnujm
yes, trailer or clip are very difficult to encode, more than a whole movie, but for a movie like LOTR... 3 hours, the rv9 will win, it really has more advanced technologies, mpeg4-iso codec can't beat it at this bitrate, we can wait for h.264 but... real has advance, when there is currently no good h.264 implentation.

@nobody aimed... well why not karl ^^
Rv9 is surely the most advanced codec, but it has a real problem... its f*cking MSL :D, it really needs a real 2pass curve compression, to achieve the most quality ^^.
but i did a re-re-test this afternoon, for high bitrate, rv9 compress too much, can't see the différence with the (same bitrate)/2, and even with a 1400mbps, it looks too washed, specially in very high motion.:D

[karl_lillevold]

I completely agree with temporance in that we must agree to differ. I am simply following up because I want to understand what goes wrong, and to try to improve the playback experience with RV9.

So I followed Sigmatador's suggestion, made a 2min30sec example at 704x304 with Neutral Bicubic resizing. High quality source, no pre-filters.

Video=Crop(Video,10,58,704,364)
Video=BicubicResize(Video,704,304,0,0.5)


Then I used the latest 'unstable' XviD from Koepi (XviD-24062003-1.exe). Set it to fixed QP=2, 1-pass, 1 B frame, Ultra-High, H.263 quantization (maybe I should have used MPEG? let me know, will try again), VHQ mode 4, 10 second key frame interval. I ended up with a 35 MB encode at 1997 kbps. PSNR = 45.18.

After this I used AutoRV9 set to target filesize 35 MB, EHQ Extra High (85), 2-pass etc, all my recommended options. I got about the same filesize (35.3 MB). PSNR = 45.80.

Even though PSNR is a measurement of how well the video matches the original, it may not say anything about this "washed out" problem. So, finally, I went ahead and stacked both video vertically, like this:

a = Avisource("test1.avi")
a = subtitle(a, "XviD")
b = DirectShowSource("test1.rmvb" , fps=23.976)
b = subtitle(b, "RV9-EHQ")
video = StackVertical(a,b)

Player for this AVS script: Media Player Classic, MPEG-4 decoder: latest ffdshow, no PP.
I both played through end to end, and stopped at key points with a lot of texture, face close-ups etc. Both videos looked great. However, in all cases, the RV9-EHQ had a little sharper appearance with slightly more details on my 1600x1200 Dell LCD screen.

My video card is nVidia GeForce 4. If you play your RV9 encodes with RealOne or RealPlayer may make a difference. I am not sure about the quality of older RealPlayers' renderers. Also, if you have an ATI card, RealOne may choose YV12 output color space instead of YUY2. Maybe this affects the appearance of sharpness. There have been known to be bugs in video cards' implementation of YV12. It is also possible small amounts of mosquito noise gives a sense of sharpness. I am not sure, just trying to improve video quality, and understand this "washed out" problem that some are reporting. Sigmatador, Gaia, feel free to PM me your system specs and how you played back RV9. Maybe there is something in common.

@Sigmatador: 2-pass VBR improvements in the works. No ETA yet.

[Sigmatador]

Originally posted by karl_lillevold
@Sigmatador: 2-pass VBR improvements in the works. No ETA yet. [/B]

Good news :cool:

[Crabba]

I have done quite a lot of codec comparisons lately, and until about a week ago I was convinced that WMV9 was the best choice for me. I always aim for the highest quality @ high res. When doing my first comparison encoding really tough DVD material checking random frames throughout the whole clip (50 min episodes @ 700mb each), I came to the conclusion that WMV9 had the best overall quality and with the same or better detail than XviD, although XviD was pretty close. DivX5.05 was not even close... RV9 was never an option for me since I always use the original AC3 soundtrack for my encodes! And AFAIK that's not possible with RV9.

But later when checking my WMV9 encodes more closely I've noticed some really annoying 'bugs' that WMV9 has. The most annoying one to me is that it sometimes inserts really bad-looking smoothed out keyframes, especially when the bitrate is not high enough. Most of the time quality and detail is back in just a couple of frames, but it still really annoys me to have these ugly keyframes in there.

Another sneaky little thing WMV9 does is that it sometimes for no apparent reason applies some smoothing in specific areas of the image! This can also be quite tricky to spot since you really need to do frame-by-frame comparisons with the avs-source to find this but, like the keyframe problem this also annoys me since I can't turn this damn smoothing off in any way and I can't be confident after an encode that I have a great quality encode with full detail intact!

So, I did another comparison with the latest Koepi 0624 build and the Nic 0716 build encoded a couple of times using different settings. All encodes with highest possible quality settings. I used Qpel but not GMC since it's not recommended, bframes with 2,150,100 and chroma motion etc. When comparing I was relieved to see that XviD does not have any of the keyframe or smoothing problems that WMV9 has, and quality was also a lot better than last time I compared with XviD. Only some scener were a bit better looking with WMV9 and a few scenes had minor blockyness but on the whole I must say I prefered the Koepi 0624 build over WMV9.

Since there has been a lot of talk about DivX Kaukura I decided to include that in the test as well. But it didn't take long to realize that Kaukura still has a LONG way to go to reach either WMV9 or XviD!! A LOT of blockyness problems and I didn't have to check very long to realize that it was not an option for me!

I also did some PSNR tests between the different codecs, I haven't done one for DivX Kaukura yet since it's really not an option for me anyway.

All encoded files were basically the same size, except for the DivX Kaukura encode which got oversized by more than one MB.

WMV9:
Minimum Average Maximum
Mean Absolute Deviation: 0.0145 1.3855 3.7924
Mean Deviation: -0.2991 +0.0104 +0.3366
PSNR: 32.9762 42.6142 66.4800

XviD Koepi:

Minimum Average Maximum
Mean Absolute Deviation: 0.0025 1.4469 3.5209
Mean Deviation: -0.5177 +0.0018 +0.3659
PSNR: 33.3547 42.0230 74.1121

XviD Nic:

Minimum Average Maximum
Mean Absolute Deviation: 0.0062 1.4815 3.7009
Mean Deviation: -0.5184 +0.0046 +0.4191
PSNR: 32.7629 41.8253 70.0340

As you can see WMV9 got the best PSNR average score, but like I said before, I still prefer the XviD image and I wouldn't be surprised if the reason WMV9 got better PSNR is because of the smoothing.

Too bad the new XviD API builds are not available for testing yet, since they seem to be even better...

[unplugged]

One thing that I see bad in this comparison [Edit: referring to main thread] is that clips that have been offered to download for direct (eye) comparison are the same that was targetted for top PSNR results.

I'm an XviD user and my common range of use is 750..3000 kbps.
Well, XviD in my direct (by eye) comparisons has always produced a better linearity (and better correspondence) with QPel enabled than without, plus, with difficult encodes (low BPF) that generally demand high quantizers like 4..12 the MPEG matrix has quite better image than H.263 (used in test).

And also for low noise sources like most DVDs, when H.263 2nd-pass requires an average quant just like 3, I go for MPEG matrix redoing 1st-pass. (in this situation generally MPEG m. give the same quantizer average for 2nd-pass, but with better detail preservation, especially in mid-low motion).

My current XviD settings (20030624) are:
Ultra-high(6), H.263, VHQ4(1 for speed), Chroma-ME, QPel, B-Frame 1(or 2)-150-75-0, Chroma opt.
As said I tend to choose MPEG matrix for tight encodes, that mostly ask high quantizer.
I disable QPel encoding only with constant high-motion stuff like sport, motor sports (F1 Grand Prixes :D) or particular sources.

[karl_lillevold]

RV9 was never an option for me since I always use the original AC3 soundtrack for my encodes! And AFAIK that's not possible with RV9.
@Crabba: Yes it is now. With Matroska you can mux RV9 with AC3. In fact, I just tried with my sample from above:

mkvmerge -o test2.mkv test1.rmvb "test1 AC3 T01 3_2ch 448Kbps DELAY -130ms.ac3"

Works like a charm. Incrased file size from 35 to 45 MB, and MPC loads up ac3filter to play the AC3 track. You can of course also use RealAudio 5.1 (http://forum.doom9.org/showthread.php?s=&threadid=56684) for true 5.1 channel playback, but only if you can connect 6 ch out from your PC. It will not work with AC3 out to amp, which is what I am guessing you are using AC3 for... ;)

[Sagittaire]

Very interessing discution ... i like that ... ;-)

@ Crabba and all

Here encoding in very high quality ...

Video: RV9 EHQ 720*432 anamorphique 1024*432 1370 Kbps
Audio: RA surround 132 Kbps for rmvb
Audio: AC3 stereo 192 Kbps for mkv

http://jfl1974.free.fr/Video/RV9-RA8.rar
http://jfl1974.free.fr/Video/RV9-AC3.rar

No comment is necessary for detail, for blocking, for ringing. The quality is extremely high ... difficult for make the difference between source and encoding ...

PS: upload is not finish ( up with 16 ko/s ... ;) )

[Sigmatador]

first i just want to say i also agree with temporance :o

@Crabba
i confirm rv9+ac3+ssa works like a charm ^^

@unplugged
for xvid encoded movie, i'm a fan of the HVS-good matrix. for anime h.263 is very good (sometimes i use CG Matrix ^^) and with current koepi's build, h.263 matrix allows you to use thetrellis quantization, which seems to be a very good features too ^^.

@Sagittaire
well, a not too deaf discussion about codec compare, it's very rare :D

[Joe Fenton]

karl_lillevold,

About your test with XviD, did you set the fourCC to XVID? If you did, remember that the playback postprocessing options must be set in a program that pulls up the codec configuration dialog, like VitualDub or GordianKnot. There you can set the decoder options. Currently, there is only luminance and chromiance deblocking, no deringing or other optimizations that perhap the RV9 decoder was performing. You could try setting the fourCC to DX50 so that it used the DivX decoder for full postprocessing.

[Sagittaire]

PP increase PSNR only for very low bitrate else PP decrease PSNR. I think have to choose the best setting for XviD and RV9 EHQ are better for PSNR test ...

[SeeMoreDigital]

Hello again,

>Temporance
I agree also. RV9 does appear to be the better codec at the moment. At most, if not all bitrate encoding speeds and pixel frame sizes. So well done to Karl and the RealMedia Video team.
So, when will be seeing deals with hardware manufacturers? As I would love to be able to spin 'RM9' CD-R's (or DVD R's) in a stand alone machine etc!

>Crabba
I'm sorry you've just found out that generating 50min/700MB WMV9 encodes are not as good as Mpeg4 (XviD). You just happen to have used WMV9 in that '950kbps - 1450kbps' grey area!
I wrote about this problem some months ago, so it's a great shame you missed it. If I remember correctly, I also informed MingCL (of M$) but no response has been forwarded as yet!

Maybe I should generate some more 'side by side' encode tests and mail them to him. To jog his memory!

>All
As most of you know, I like my Mpeg4. But you "have to admit defeat when you know you've been beat!" And RealMedia certainly cuts the mustard.

Now, just changing the subject a bit. Mpeg1 has an .mpg file extension/container. Mpeg2 has an .mpg file extension/container. So, is there any reason why Mpeg4 can't have an .mpg file extension/container too?

Just asking!

[superdump]

Can I just say that I have _never_ been able to get RV9 to complete a PSNR test on my computer. I have tried copying Sagittaire's scripts but I cannot get it to work. It compares all but the last three frames then hangs.

I'm using MPC, RealSplitter and the following script:
name= "rmvb.rmvb"

source=avisource("vble.avi")
test=directshowsource(name,fps=25).converttoyv12()

compareyv12(test,source,"", name+".psnr.log",false)

I have also tried with ".deleteframe(0)" on the source as I noticed that in Sagittaire's tests.

I have tried many combinations (of players, scripts, source files, etc), including using the standard compare but still it hangs 3 frames before the end so I don't actually get an output.

Could somebody offer some assistance?

[Sagittaire]

# --> Video Opening <--

source=AviSource("D:\Mes dossiers\B.A\Les deux tours\Encodage-715.avs")
source=DeleteFrame(source,0)
source=ConvertToYUY2(source)

video=DirectShowSource("D:\Mes dossiers\B.A\Les deux tours\RV9-950.rmvb",fps=25)
video=ConvertToYUY2(video)

# --> PSNR analysis <--
Compare(video,source,"","PSNR-RV9-950.txt")



Always:
-First source frame delete for comparing with RV9.
-End frames not decoded by DShow

Sometime:
-Frame desynchro between source and video

But there are a PNSR fonction integred in producer if PNSR with avs script is not possible.


<codecProperties type="bag">
<encoderComplexity type="uint">85</encoderComplexity>
<calcPSNR type="bool">true</calcPSNR>
<customPacketSize type="uint">16000</customPacketSize>
</codecProperties>

[karl_lillevold]

@superdump: In order to avoid a hang at the end, try my suggestion in in this post (http://forum.doom9.org/showthread.php?s=&postid=349102&highlight=0.01#post349102). I had to delete the first two frames and the last frame from the original, and the first and the last frame from the DirectShowSource video (RMVB). When I did this, I was able to verify that the Avisynth method with a RV9 clip as DirectShowSource calculated a PSNR number only differing by 0.01 dB from the number output by producer's calcPSNR function explained in referenced post).

I also seem to recall when I did experience a hang (saving RV9 as YV12 in VirtualDub) when not deleting the last frame. However, VirtualDub would exit gracefully aftering hitting the Abort button producing a "good" AVI file, while VirtualDubMod would hang until killed with an invalid AVI as a result.

So I did another XviD encode, with QPel, chroma MC, and MPEG quantization. PSNR dropped from 45.18 to 44.63, filesize up 1.5 MB, but the video quality was sharper, textures more defined, close to RV9-EHQ (PSNR=45.80), sometimes even sharper, but at the cost of slightly more chroma noise, moving mosquito noise, and just a few visible block edges. I am not really sure which I prefer of the two XviD encodes, both looked great. Probably this last one, due to increased sharpness, but I think it would depend on viewing environment.

Re PP: when loading an XviD in an Avisynth script (StackVertical comparison, or PSNR), I believe the native XVID decoder DLL is used, with no PP, which is what you want for the sharpest appearance and highest PSNR. Yes, I prefer no PP when playing MPEG-4 encodes, so my ffdshow is also set up with no PP :)

Regarding perceived RV9 smoothness, relative to MPEG-4: Ramirez told me about a nice option in ffdshow... Set it up to handle raw formats (YV12), then enable Sharpen->asharp and experiment with the sharpness params playing RV9 in MPC with its DirectShow handler (RealMediaSpliter). Verify that it loads ffdshow to handle output format YV12. This adds any amount of sharpness and crispness you like, from just a "touch" to an extreme sharpening effect. Since there is little blockiness in RV9, the sharpening appears to behave quite well.

[superdump]

OK, thanks for the responses. When I create a clip I use "trim(startframe,-numberofframesinclip)". So if I were to use -1500 in the trim, I expect this to be the resulting compareyv12 script when considering Karl's recommendation to delete the first two frames and last frame from the source and the first and last frames from the directshowsource:

name= "rmvb.rmvb"

source=avisource("vble.avi").deleteframe(0).deleteframe(1).deleteframe(1499)
test=directshowsource(name,fps=25).deleteframe(0).deleteframe(1499)

compareyv12(test,source,"", name+".psnr.log",false)

I assume this is correct (if a little long-winded).

It however, does not work. Now it resizes the player window but does not appear to do anything. (I left it for 10 minutes and it was using 99% cpu time.) I will try Sagittaire's method in a bit.

Does the internal PSNR calculation of producer.exe produce an overall PSNR or an average PSNR?

[karl_lillevold]

here is is a script that worked on my system (Avisynth 2.5.2, opened and "played" with VirtualDub):

Video1=AviSource("original.avi")
Video1=DeleteFrame(Video1,0)
Video1=DeleteFrame(Video1,0)
Video1=Trim(Video1,0,684)
Video2=DirectShowSource("compressed.rmvb", fps=23.976)
Video2=DeleteFrame(Video2,0)
Video2=Trim(Video2,0,684)
Compare(Video1,Video2,"Y","compare.log")


Producer PSNR is average PSNR, and it also lists individual frame PSNRs.

[unplugged]

Originally posted by karl_lillevold
So I did another XviD encode, with QPel, chroma MC, and MPEG quantization. PSNR dropped from 45.18 to 44.63, filesize up 1.5 MB, but the video quality was sharper, textures more defined, close to RV9-EHQ (PSNR=45.80), sometimes even sharper, but at the cost of slightly more chroma noise, moving mosquito noise, and just a few visible block edges. I am not really sure which I prefer of the two XviD encodes, both looked great. Probably this last one, due to increased sharpness, but I think it would depend on viewing environment.
VHQ4 and B-Frames 2-150-75-0 too?
I have tested it and VHQ4 works very well with MPEG matrix! Fewer blocks! :)
If you are curious of comparisons/results, the always strong point of xvid is motion encoding, take care of those detailed frames (those haven't much flat surfaces) that has progressive motion (especially full screen motion) or complex motion (water...), the detail preserved frame-by-frame and its reflection with the original is stunning.
Here respect to DivX 5 (also Kaukura) xvid is in another world, and I see difficulty any other encoder to reach that incredible fidelity during *massive motion* (I'm NOT referring to unpredictable or fast motion, for that every codec trick is valid... :D ).
Recently I have tested 40s sequence of movie "Nid de guêpes" at anamorphic res. 704x432 (original 2.35:1 without black bars) for 900kbps results. The clip is quite crisp, totally noise free and has still, low-motion and a long and progressive camera panning part.
Xvid was set at 900kbps, Ultra-high(6), MPEG matrix, VHQ4, Chroma-ME, QPel, B-Frame 2-150-75-0, alt curve compression (alt-cc) set with medium curve and standard values.
WMV9 set to 900000bps and max quality (slow) have reached 1800kbps (!) average bitrate after encoding, so I have immediately discarded it from my comparison :D (hate such young software behaviour! crap!!).
DivX 5 Kaukura was set with BF and QPel enabled, perf. set to slow, psy to slow, I-Frame distance to 1000 (as I set XviD) and performed with 2 pass using new algo.
The image quality is very similar between xvid and divx in static or low motion frames, certain particulars with complex parts (contrasts) looks like original in divx video whereas xvid doesn't take enough care of detail, but overall the xvid images have much less distorsions than divx.
During real motion or camera panning here comes the historical limits of divx 3, 4, 5, 5.0.2, 5.0.3, 5.0.5 and every version known :), my point is that its motion encoding engine seems most tricky than effective and making progressive comparisons frame-by-frame, xvid here really does the difference and shines even at high quantizers (think 5-6-7) jointly with B-Frames!! The image distorsion is minimal (and the transients looks solid and greatly motion-coded), also because there is a light loss but distributed over all details.
Originally posted by karl_lillevold
Re PP: when loading an XviD in an Avisynth script (StackVertical comparison, or PSNR), I believe the native XVID decoder DLL is used, with no PP, which is what you want for the sharpest appearance and highest PSNR. Yes, I prefer no PP when playing MPEG-4 encodes, so my ffdshow is also set up with no PP :)

If you use AVISource() and similars then VFW will be used, so xvid.dll kickin (no PP)
If you use DirectShowSource() .AX components kickin (xvid.ax), maybe with PP, it depends by plugin setting of course.

[Sagittaire]

WMV9 set to 900000bps and max quality (slow) have reached 1800kbps (!) average bitrate after encoding, so I have immediately discarded it from my comparison

......... ?!!!

My test is making with XviD Devapi4. The XviD Devapi4 is verry better than XviD Koepi 24/06/03 with visual(blocking, ringing), PSNR, VQM and SSIM ...

My setting Ultra High, H263, VHQ4, Treillis, Uncontraining, bframe 1/150/75/0. MPEG Matrix are very little influance and decrease PSNR for exemple ...

But WMV9 is better than XviD Devapi4 and very better than Devapi3 and with PP or with not PP ...

[unplugged]

Sure I'll try WMV9
but, it doesn't respect the bitrate, every time I tried it doubles, inflates the output so...
must I try kind ballistic solutions like type 450,000 bps into codec properties hoping to obtain my video at 900,000 bps ???

For example, how you target 1CD rip with WMV9 with that inflated results, do you try bitrates until you reach your size? :devil:
MPEG Matrix are very little influance and decrease PSNR for exemple ...
Once more I read "...decrease PSNR...", now these comparison tools are really becoming the way to measure codecs?
The way also to choose their setting? (MPEG yes or no because of the PSNR?)

2nd, you have not used QPel for your xvid clips, the combination of B-Frames and QPel disabled has always given me a bad impact on the result, loss of overall detail, a kind of wash effect (things that just an xvid user dislike), does't matter what happen to PSNR.
IMHO for natural images and movies xvid without Qpel has no chances to give its best for image preservation (PSNR, VQM, SSIM apart), and in particular way with B-Frames.

For example in my experience:
- MPEG matrix starting from quantizer 3 do the best quantization for EYES almost always
- You can compare MPEG quant 3 with quant 2.5 in H.263 as detail preservation
- At quant 3 or more MPEG compresses same or better than H.263 (unless input is noisy)
- At higher quantizer (4,5,6...) detail fades in linear way with MPEG, instead with H.263 shows a random and hard detail cutting (threshold sensible), quite ugly for EYES

But the PSNR stuff doesn't see most part of this, sure, for example this is one of the reasons because audio codecs quality tests are made directly by humans.

[Sagittaire]

If you want the best visual quality with quant 2 use HVS matrice for exemple ... but WMV9 and RV9 EHQ are better in visual, PSNR, VQM and SSIM test for all bitrate and all XviD seeting ...

I make all the possible setting with all version (Koepi, Unimaniac, Devapi4, ffvfw 3 pass ...) and my XviD setting are the best for me (it's my opignon) ...

My test have to make in 570~950 Kbps bitrate interval and in this interval use QPel or GMC isn't good for quality ...

The advantage of iso-MPEG4 codec is this compatibility with hardware dec (Kiss ...) but not if you actived Qpel or GMC MPEG4 ...

[Peters]

At least with Divx 5.xx, GMC works (since 2.6.4 firmware) on Kiss players.

from doom9 Codec shoot-out 2003
http://www.doom9.org/codecs-103-1.htm

"RV9 and WMV9 have the most visible tendency to smooth out details. On the other end of the spectrum, SBC and XviD tend to show the largest amount of details. DivX5 is situation in the middle of those two factions."

speaking of RV9 "the foreground is usually rather detailed, whereas there aren't many details in the background."

And 'His' conclusion

"Personally I prefer codecs that retain more details than RV9 but that's just a personal preference. If "no blocks" is all you care about RV9 might just be your thing. "

I agree with Doom9, i see the same defaults in your RV9 EHQ trailers
(faces with less details, background washed)

[Sagittaire]

In this test codec is old and with no better setting:

- XviD Koepi devapi3 ...
- WMV9 beta version with PP=4, PP for WMV9 is a smmother and 4 is the maxi setting ... !
- RV9 without EHQ and SML optimisation ...
- DivX 5.05 ...
- DivX SBC with NanDub ...

It's no representative test of actual situation. Here the actual situation:

- XviD Devapi4 ...
- WMV9 final version with PP=0 ...
- RV9 EHQ=85 and SML=60 ...
- DivX Kaukura ...
- DivX3 SBC with ffvfw ...

My test is a very representative test of veritable power for actual codecs and my conclusion: the best codecs are in order RV9, WMV9, Kaukura, XviD and DivX3 ...

"RV9 and WMV9 have the most visible tendency to smooth out details. On the other end of the spectrum, SBC and XviD tend to show the largest amount of details. DivX5 is situation in the middle of those two factions."

Here demo who change your position ... ?

http://jfl1974.free.fr/Video/RV9-RA8.rar

[Sirber]

Doom9's comparison didn't use EHQ for RV9, so you can't use it now to talk about RV9 :p

[Peters]

Originally posted by Sirber
Doom9's comparison didn't use EHQ for RV9, so you can't use it now to talk about RV9 :p

I know it! but his conclusion in regard of the smoothing out of details with RV9 don't change ,for me, with RV9 EHQ

(maybe Doom9 will update his test when he got time)

Have you seen the faces in Sagittaire HP 2 trailer? idem with The two towers.

@Sagittaire

Yes this demo is brillant like your RV9 EHQ encodings and i repeat what i said in an other post it's , even for me, the most pleasant codec for the eyes (globally).

The only thing where i disagree is when you say there a much details with RV9 EHQ . For me some areas are really washed saving bits for the rest of the frame.

[superdump]

Originally posted by Sagittaire
My test is a very representative test of veritable power for actual codecs and my conclusion: the best codecs are in order RV9, WMV9, Kaukura, XviD and DivX3 ...

Why don't you try on some full films instead of trailers? I know this would take an age but the results would stand up better than tests on trailers which have lots of fades and are very action packed. I'd rather wait a week and have results from full film encodes than wait a day and have results from trailers personally.

[CruNcher]

@Sagittaire and Sirber

i have a clip here made with the preview version of RV9 and its still compatible to the now used RV9, so my conclusion since then nothing important changed in RV9 only Rate Controll did, so the visual impression of the used compression algortihm is still the same and Doom9s visual impression therfore is fully valid.

[karl_lillevold]

Originally posted by CruNcher
i have a clip here made with the preview version of RV9 and its still compatible to the now used RV9, so my conclusion since then nothing important changed in RV9 only Rate Controll did
I am afraid something must have gone wrong during your RV9-EHQ encode... For instance in Producer M5 there is a problem enabling EHQ using the most common method, unless you get a fixed DLL.

The difference between RV9 and RV9-EHQ is in how the encoder spends more time to choose more optimal encoding modes, better motion representation, and larger packet sizes. The resulting improvement is significant, should be visible in almost all cases, and has nothing to do with rate control.

[Peters]

Originally posted by karl_lillevold
The difference between RV9 and RV9-EHQ is in how the encoder spends more time to choose more optimal encoding modes, better motion representation, and larger packet sizes. The resulting improvement is significant, should be visible in almost all cases, and has nothing to do with rate control.

Questions:

The MSL max is 60 s? what's happen when there is a long action scene (ex 3mn of a car pursuit)

Is really the max bitrate useful?
Maybe i'm not lucky but i have never seen large fluctuation of the bitrate

Thank's

[Sigmatador]

Originally posted by karl_lillevold
I am afraid something must have gone wrong during your RV9-EHQ encode... For instance in Producer M5 there is a problem enabling EHQ using the most common method, unless you get a fixed DLL.

The difference between RV9 and RV9-EHQ is in how the encoder spends more time to choose more optimal encoding modes, better motion representation, and larger packet sizes. The resulting improvement is significant, should be visible in almost all cases, and has nothing to do with rate control.

agree ^^ i'm not a RV9 nerd like some people here, but the last time i tested rv9, before the EHQ feature, was with the preview version.
and there's a more than noticiable difference ^^.

[CruNcher]

@karl_lillevol

im sure i made no mistake i was really carefull and used your dll with your version info and HPG

http://www.mufflastig.com/CruNcher/XviD/rv9-softwashy/

as you can see rv9 does extreme blurr high motion scenes compression is better but the lose of detail even in high motion is for many people recognizable and as you can se the GMC frame looks more detailed ok more blocks but i can live with that

[Dark-Cracker]

@cruncher
i also find rv9 blur to much the hight motion scene i have added in my software an option to add grain only on hight motion and i am interested to test it on your sample .perhaps could u provide me a link on a sample of this material ?

Bye.

[unplugged]

Does Windows Media Video 9 VCM encoder (VFW) performs just like Windows Media Encoder 9, anyone has reported variations about efficiency?
Do you set the PP for WMV9 from Windows Media Player application?
Thanks (couldn't find this explained in thread).

[69link]

@Cruncher,

about RV9-EHQ being soft at highmotion scenes.

Isnt that a good thing? I mean, for me its better that the few bits available is spent on scenes where its more noticeable, than on scenes where you cant appreciate the quality anyway.

[Sagittaire]

@ Cruncher

It's more interessant if you post sample. RV9 EHQ is not better for each frame but for large majority of frame ...

[temporance]

You can see from the sizes of the png's that detail has been lost in the rv9 frame:

XviDdevapi4-sprite.png 20-Aug-2003 22:03 244k
original.png 20-Aug-2003 23:02 414k
rv9-ehq.png 20-Aug-2003 22:03 208k

Isnt that a good thing? I mean, for me its better that the few bits available is spent on scenes where its more noticeable, than on scenes where you cant appreciate the quality anyway.It's only a good thing if it results in an perception of overall quality improvement. If the only difference we can notice is the washed-out rv9 frames, then it's a bad thing. How does the rest of the movie look?

[karl_lillevold]

@Peters: very accurate observation. MSL=60 is not always useful, since the current VBR scheme in general stays a little too close to the average and thus rarely fills this startup buffer, in fact it rarely even filled the old 20 second buffer. However, this parameter will be more useful. We are experimenting with a new and improved, perhaps adjustable, 2-pass VBR scheme..

@all: so since the codec stays reasonably close to average bitrate, high motion scenes sometimes appear too soft, and if you manage to freeze the video just at the right spot, this will be apparent. Personally, I prefer such behavior, since I am really bothered by blockiness even during high motion, and the softness is not that noticable during regular "non-comparing" video playback for entertainment purposes. That said, it is likely that the 2-pass VBR improvements we are considering, will (i) allocate more bits for high motion, and/or (ii) be adjustable, since preferences vary. Fortunately there are plenty of codecs to choose between, and like has been mentioned before, we have to agree to differ. From this useful discussion I will take back the advice about high motion scenes preferences and rate control. thanks.

@CruNcher: that's a good example of what I just described, but there's also an old saying ... "one freeze frame does not make a whole video codec comparison" ;)

@temporance: kind of interesting datapoint.. details and blocks do not compress as well with png as a smoother image..

[Sagittaire]

If you want i cant find frame in my encoding with catastophic quality for XviD Devapi4 ... it's not difficult ... but not representative ... it's perhaps beta bframe for RV9 and a iframe for the XviD ...

Post a sample please and one will be able to discuss ...

[Sagittaire]

Little PSNR test for detail mesurement. resize influance ...

Source with lanczos

source=Mpeg2Source("D:\Mes dossiers\B.A\Les deux tours\azerty.d2v")
source=Crop(source,24,88,-24,-72)
source=LanczosResize(source,640,272)
return source


Test with bilinear

source=Mpeg2Source("D:\Mes dossiers\B.A\Les deux tours\azerty.d2v")
source=Crop(source,24,88,-24,-72)
source=BilinearResize(source,640,272)
return source


PSNR test

# --> Video Opening <--

source=AviSource("D:\Mes dossiers\B.A\Les deux tours\source.avs")
source=ConvertToYV12(source)

video=AviSource("D:\Mes dossiers\B.A\Les deux tours\test.avs")
video=ConvertToYV12(video)

# --> PSNR analysis <--
CompareYV12(video,source,"","test-Bilinear.txt")


Sharp bicubic: 57.2557 dB
Neutral bicubic: 54.6731 dB
Soft bicubic: 48.9181 dB
Bilinear: 47.2373 dB

[superdump]

Karl and Sagittaire: Thanks for your help. It's working now. I'd messed up the deleting frames because I forgot to renumber them after I deleted each frame.

This is my working script for an n framed clip:
source=avisource("original.avi").deleteframe(0).deleteframe(n-2)
test=directshowsource("test.rmvb",fps=25).deleteframe(n-1)
compareyv12(test,source,"","rv9psnr.log",false)

Of course that's for 25fps clips.

Sagittaire: Could you include Lanczos and Simple too?

[Sirber]

Will there be a way to have sound with DirectShow source?

[temporance]

Originally posted by Sagittaire
Little PSNR test for detail mesurement.

Sharp bicubic: 57.2557 dB
Neutral bicubic: 54.6731 dB
Soft bicubic: 48.9181 dB
Bilinear: 47.2373 dB Sagittaire,

I'm not sure what this little test proves, apart from that the sharp bicubic FIR kernel (looks like a curvey valley, taller hill, valley) is the closest match to the Lanczos FIR kernel (looks very similar but can have extra, much smaller hills and valleys on either side). The bilinear kernel (just one, tall, triangular shaped hill) is, not surprisingly, the most different. These facts could be shown mathematically, in the spatial or frequency domain, without ever using an image. Incidentally, of the filters you've used Lanczos is the best approximation to the theoretically ideal, "brick-wall" frequency response resampling.

It doesn't prove that a sharper image gets a higher PSNR - it is the image with smallest sum of squared differences that gets the highest PSNR (and this is skewed somewhat if you're using Average, not Overall PSNR).

Just a little scientific peer-review ;)

Having said that, you've given me an idea. How about trying encoding at different resolutions at the same bitrate to see if we can improve PSNR that way? E.g., say encoding 400k @ 640x480 gives PSNR of 38dB. If I resize to 320x240 and encode again at 400k, decode and resize back up to 640x480, will I get a better, or worse PSNR?

[Sigmatador]

mouhahahahahah, a resizing algo compare by using PSNR measure between each other LOL the most stupid PSNR test i've ever seen. *ROFL*

@temporance
Yeah you've right, the only thing this prove: sharp bicubic is the closest algo to the lanczos one (Wow I would have never imagine that :D )

Having said that, you've given me an idea. How about trying encoding at different resolutions at the same bitrate to see if we can improve PSNR that way? E.g., say encoding 400k @ 640x480 gives PSNR of 38dB. If I resize to 320x240 and encode again at 400k, decode and resize back up to 640x480, will I get a better, or worse PSNR?A more better idea ^^

the problem is, the 320*240 encode resized in 640*480 will be smoother than the 640*480 encode. And the PSNR algo doesn't really like smooth. But with a HVS algo like SSIM, i wonder if that would not be better? (for sure, when a 640*480 xvid encode looks too blocky, the 576 or 512 ones look better for human eyes)

[temporance]

I did some quick tests (DivX Kaukura, default settings), using AviSynth's Lanczos resize in YUY2 colorspace:

Results (native resolution is 640x304)
width height bitrate passes Overall PSNR
1024 480 400 3 37.7668
640 304 400 3 39.6329
576 272 400 3 39.7113
480 240 400 3 39.7837
400 208 400 3 39.6520
320 160 400 3 38.9221
So, reducing resolution before encoding at low bitrate, then expanding image after decoding can produce a PSNR improvement. Yes, PSNR doesn't like the softening of resizing, but it obviously hates the blocking and ringing produced at full res even more.

I'm sure bigger improvements are possible. I should think this is very content dependent (my clip was fairly average action clip from a movie and even at the native resolution, 400kbit/s isnt't a particularly low bitrate for this clip with 3-pass Kaukura). Got to do some real work now, anyone got some time to explore this further?

Btw, I have suspected that WMV9 uses this "pre-encode-shrink, post-decode-expand" to get better results at low bitrates. Perhaps we'll never know. If you're listening Karl, perhaps you can tell us if RV9 does something like this?????

[SeeMoreDigital]

Hey temporance

See what happens to your PSNR results when you do this. With the same source, encode at 1024x576 (or crop/resize to 768x436 if it's a 2.35:1 image)!

Cheers

[Gaia]

This thread is getting funny... PSNR tests with different resizing methods... What that should prove? Some people are getting really obsessed.

What display device you use to judge these test?
Monitor,LCD screen,HDTV,Video Projector,tv(old/new)? Your encodes look quite different in HDTV than in some old tv or 17" monitor.

Some people like more sharper, crispier detailed images some like more washed out, block free.

Also full movie test are better than just trailers or short clips. Try with different sources, noisier, high detailed etc. Good clean high guality dvd's are very easy to encode with every codec. Try some really ugly noisy source.

One thing i found out is that RV9 is bad with high motion nature scenes. Grass, trees, forest etc. Looks really washed out.

One person can't judge and say this codec is the best. You would have to organize some kind of blind test. Like those audio tests. Use different display devices etc.

I would be also nice too see some really high insane bitrate tests...

[temporance]

@SeeMoreDigital: I've edited my post to include a datapoint at resoultion 1024x480. As expected, PSNR is worse (blocks are smaller though ;) ).

@Gaia: Yes, it's insane to use PSNR (or any other metric) to compare resize methods, and it's accidental that this thread has drifted onto that topic. However, we have discovered that we can improve the max PSNR of a codec by letting it operate at a lower resolution than the video we're testing. i.e.

[reduce resolution] -> [encoder] -> [decoder] -> [increase resolution to original]

[SeeMoreDigital]

Gaia, everyone likes to experiment in their own way!

One thing i found found is that RV9 is bad with high motion nature scenes. Grass, trees, forest etc. Looks really washed out.I have to agree with you on this. For me I still prefer to look at a WMV9 720x576 image (.wmv or .avi) when encoding at a bitrate of 900kbps or below. The whole image just seems to be more engaging!

One person can't judge and say this codec is the best. You would have to organize some kind of blind test. Like those audio tests. Use different display devices etc.Unless we all agree on encoding the same test file(s) using the same encoder at the same settings and are able to use the same display devices... Thias will never happen. Shame!

I would be also nice too see some really high insane bitrate tests... Yep, this would be fun. The HD WMV9 version of T2 extreme is over 6000kbps (at 1440x816) But high bitrate encodes would run alot better at a true 1.77:1 'square pixel' PAL frame size of 1024x576... Don't know if the 1.77:1 'square pixel' NTSC frame size of 854x480 would work!

[temporance]

Originally posted by SeeMoreDigital
I have to agree with you on this. For me I still prefer to look at a WMV9 720x576 image (.wmv or .avi) when encoding at a bitrate of 900kbps or below. The whole image just seems to be more engaging!IMHO, the DivX betas with slow modes, new PV and new PP work very well at these bitrates - they produce more detail and no more artifacts than WMV9.
But high bitrate encodes would run alot better at a true 1.77:1 'square pixel' PAL frame size of 1024x576... Don't know if the 1.77:1 'square pixel' NTSC frame size of 854x480 would work! Disagree. There is no point going to a higher horizontal resolution than 720 when encoding for SDTV, for two reasons: 1, SDTV source only ever had 720 horizontal pixels when originated on professional kit, and 2. When you output to a TV, image is resampled to 13.5MHz = 720 visible pixels / line. Perfectly square pixels don't really help a codec.

[SeeMoreDigital]

Originally posted by temporance
@SeeMoreDigital: I've edited my post to include a datapoint at resoultion 1024x480. As expected, PSNR is worse (blocks are smaller though Sorry temporance 1024x480 will not work properly. If your PAL source image is 1.77:1 you can resize this to 1024x576.
If your PAL source image is 1.85:1 you can crop and resize to 992x554.
If your PAL source image is 2.35:1 you can crop and resize to 768x436.

I normally just leave them at 1024x576. On a PC they look great. Like you say smaller blocks!

But these frame sizes don't work on my Xcard. Bummer!

[SeeMoreDigital]

Originally posted by temporance
... Disagree. There is no point going to a higher horizontal resolution than 720 when encoding for SDTV, for two reasons: 1, SDTV source only ever had 720 horizontal pixels when originated on professional kit, and 2. When you output to a TV, image is resampled to 13.5MHz = 720 visible pixels / line. Perfectly square pixels don't really help a codec. Yes, encoding 720 horizontal pixels to 1024 shoul'nt make much improvement. But as you've said yourself, it does produce smaller blocks which makes a high speed scene look much better!

It's a strange one for sure!

Another one of these... 'Why bees should not be able to fly, but they can'. Conundrums!

Cheers

EDIT
In my opinion a PAL DVD should never have used a pixel frame size of 720x576. It should have used a true 16:9 (1.77:1) frame size 1024x576 to begin with. Just as HD broadcasts use a true 1.77:1 pixel frame size. Anamorphic re-framing is a good idea but a poor substitute for the true frame size.

A 'stand alone' DVD player could quite easily have coped with the electronics required to make such an idea possible. That said it could have also coped with 'film speed' 24fps DVDs!

Anyway it looks like this will arrive with Blu-Ray!

[Sagittaire]

Little PSNR test: Blur influance ...

Source

source=Mpeg2Source("D:\Mes dossiers\B.A\Les deux tours\azerty.d2v")
source=Crop(source,24,88,-24,-72)
source=LanczosResize(source,640,272)
return source


Test with blur 0.3

source=Mpeg2Source("D:\Mes dossiers\B.A\Les deux tours\azerty.d2v")
source=Crop(source,24,88,-24,-72)
source=LanczosResize(source,640,272)
source=Blur(source,0.3,0.3)
return source


PSNR test

# --> Video Opening <--

source=AviSource("D:\Mes dossiers\B.A\Les deux tours\source.avs")
source=ConvertToYV12(source)

video=AviSource("D:\Mes dossiers\B.A\Les deux tours\test.avs")
video=ConvertToYV12(video)

# --> PSNR analysis <--
CompareYV12(video,source,"","test-Bilinear.txt")


blur 0.1: 54.5353 dB
blur 0.2: 50.1381 dB
blur 0.3: 47.4559 dB
blur 1.0: 39.6771 dB

[temporance]

Originally posted by SeeMoreDigital
Sorry temporance 1024x480 will not work properly. If your PAL source image is 1.77:1 you can resize this to 1024x576.I'm not sure I know what you're saying. I started with a progressive clip 640x304, just wanted to do some PSNR tests encoding at different resolutions, maintaining approximate aspect ratio.

In my opinion a PAL DVD should never have used a pixel frame size of 720x576. It should have used a true 16:9 (1.77:1) frame size 1024x576 to begin with. Just as HD broadcasts use a true 1.77:1 pixel frame size. Anamorphic re-framing is a good idea but a poor substitute for the true frame size.

A 'stand alone' DVD player could quite easily have coped with the electronics required to make such an idea possible. That said it could have also coped with 'film speed' 24fps DVDs!The resolutions used for SDTV (and interlace) are the result of a long history of developments in the TV industry. Using a non-standard video format would not have been easy.

[temporance]

Originally posted by Sagittaire
Little PSNR test: Blur influance ...

blur 0.1: 54.5353 dB
blur 0.2: 50.1381 dB
blur 0.3: 47.4559 dB
blur 1.0: 39.6771 dB Yes, nice experiment, expected results. What's the conclusion?

[Sagittaire]

It's just test for observing blur influence on PSNR ... no conclusion ...

[CruNcher]

@Sagittaire

i dont have the space nor the traffic to do this right now but i will cut out samples from both if i know how to cut .rmvb so if you help me here i can provide smaller clips :) both trailers are 20 mb so i would need 40 mb for both + traffic if someone would sponser this no problem but i doubt somebody will.

[Sagittaire]

If I were to make a conclusion it is that in general test PSNR, SSIM and VQM consolidate my visual impressions ...

[SeeMoreDigital]

I'm not sure I know what you're saying. I started with a progressive clip 640x304, just wanted to do some PSNR tests encoding at different resolutions, maintaining approximate aspect ratio. If you encode from an 720x576 anamorphic source (instead of an 640x304 source) you may 'see' what I mean!

It's too bad there aren't many DVD's that occupy all of the 720x576 'anamorphic' image pixel frame size (ie 1.77:1). Most are ofcourse 1.85:1 or 2.35:1 which have far less active pixels.

But please stay with it and give it a go, as I would like another opinion.

I've generated a few full frame 1024x576 1.77:1 using DivX and WMV9 (VCM) encodes at 674kbps and 950kbps. And I was quite surprised 'visually' at what happened!

Cheers

[Sagittaire]

I think pixel format for 720*576 isn't 1.77:1 but 1.45:1 (16/11 in fact) ...

1.77:1 is video format for 16/9 but not pixel format for 720*576 anamorphic ...

You can make anamorphic wmv9 encoding with WMEncodeur in wmv format ...

[SeeMoreDigital]

Originally posted by Sagittaire
I think pixel format for 720*576 isn't 1.77:1 but 1.45:1 (16/11 in fact) ...
1.77:1 is video format for 16/9 but not pixel format for 720*576 anamorphic ... True. 720x576 is actually 1.25:1. I've ammended my previous post accordingly.

[Sagittaire]

Here anamorphic 720*432 demo in the most powefull format (it's my opignon ...)

http://jfl1974.free.fr/Video/RV9-RA8.rar
http://jfl1974.free.fr/Video/WMV9-WMA9.rar

[Sagittaire]

@ CruNcher and all

http://jfl1974.free.fr/Video/RV9-RA8.rar
http://jfl1974.free.fr/Video/XviD-MP3.rar

I made test with the quicktime source for matrix reloaded with this avs script ...


LoadVFAPIPlugin("C:\Program Files\AviSynth2\plugins\QTReader.vfp", "QTReader")

Clip1=QTReader("D:\Mes dossiers\Mes fichiers reçus\trailer_final_1000_dl.mov")
Clip1=FlipVertical(Clip1)
clip=ConvertToYV12(Clip1)
Clip1=SelectEvery(Clip1,25,1)
Clip1=LanczosResize(Clip1,640,344)
Return Clip1



XviD Devapi4
640*344 950 kbps H263, Ultra high, VHQ4, bframe 1/1.50/0.75
MP3 VBR 96 Kbps (ultra high quality encoding)

RV9
640*344 950 kbps EHQ 85, MSL 60
RA surround 96 Kbps

download and compare: RV9 EHQ winner

[Sirber]

RV9 is really the winner. Q is 2 times better IMHO in RV9 for this clip.

[Teegedeck]

ehe. I like reencoding quicktime-trailers myself, but it's hardly a good idea to reencode an MPEG4-encoding with an MPEG4-codec now, is it?

Judging from your SW2-encodes only, RV9 still can't denounce it's streaming-origin; WMV9 beats it very easily with the level of details. Sadly WMV9 doesn't play smoothly for me with WMP 6.4.

MP3 VBR 96 Kbps (ultra high quality encoding) :)

[Sagittaire]

ultra high quality setting for encoding with MusicMatch Converter ... sorry but Quality isn't bad for MP3 96 Kbps ... ?

RV9 EHQ Winner and in order WMV9, Kaukura and final XviD ... allways same result for me ...

Note: real bitrate is 939 kbps for RV9 and 960 for XviD but rmvb must have a strong header ... this probleme don't exist with mkv container ... :-)

[CruNcher]

@Sagittaire

640*344 look @ it closely again dont you see something ?

[CruNcher]

@ all
here is a sample of a lowmotion followed by highmotion clip @ 550 kbps
it was hard to hit (almost) the same 550 kbps with Rv9 but useing this was the best result after 6 encodes


<audience xmlns="http://ns.real.com/tools/audience.2.0">
<avgBitrate type="uint">480000</avgBitrate>
<maxBitrate type="uint">481000</maxBitrate>
<streams>
<stream xsi:type="videoStream">
<pluginName type="string">rn-videocodec-realvideo</pluginName>
<codecName type="string">rv9</codecName>
<codecProperties type="bag">
<encoderComplexity type="uint">80</encoderComplexity>
<noisyEdgeFilter type="bool">false</noisyEdgeFilter>
<calcPSNR type="bool">true</calcPSNR>
<customPacketSize type="uint">16000</customPacketSize>
</codecProperties>
<encodingType type="string">vbrBitrate</encodingType>
<quality type="uint">100</quality>
<maxStartupLatency type="double">60</maxStartupLatency>
<maxFrameRate type="double">25</maxFrameRate>
<maxKeyFrameInterval type="double">10</maxKeyFrameInterval>
<enableLossProtection type="bool">false</enableLossProtection>
</stream>
<stream xsi:type="audiostream">
<pluginName type="string">rn-audiocodec-realaudio</pluginName>
<codecName type="string">cook</codecName>
<codecFlavor type="uint">28</codecFlavor>
<streamContext type="bag">
<presentationType type="string">audio-video</presentationType>
<audioMode type="string">music</audioMode>
</streamContext>
</stream>
</streams>
</audience>


<encoderComplexity type="uint">80</encoderComplexity> is 80 as suggested by karl_lilevold and used his .dll for producer as gui HPG was used

http://www.mufflastig.com/CruNcher/XviD/phagain/

argh sorry i ran out of diskspace so please download the xvid first i will let it up for severall hours and then i finish the rv9 one sorry :(

[Sagittaire]

The profil is very strange ...
<avgBitrate type="uint">480000</avgBitrate>
<maxBitrate type="uint">481000</maxBitrate>

and
<quality type="uint">100</quality>

setting are not compatible: it's an encoding with bitrate or quality. For bitrate encoding ton must have avgBitrate x 2.1 = maxBitrate. Use AutoRV9 or RMFactory for better encoding RV9 ...




<?xml version="1.0" encoding="UTF-8"?>
<audience xmlns="http://ns.real.com/tools/audience.2.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://ns.real.com/tools/audience.2.0 http://ns.real.com/tools/audience.2.0.xsd">
<avgBitrate type="uint">550000</avgBitrate>
<maxBitrate type="uint">1200000</maxBitrate>
<streams>
<stream xsi:type="videoStream">
<pluginName type="string">rn-videocodec-realvideo</pluginName>
<codecName type="string">rv9</codecName>
<codecProperties type="bag">
<encoderComplexity type="uint">85</encoderComplexity>
<customPacketSize type="uint">16000</customPacketSize>
</codecProperties>
<encodingType type="string">vbrBitrate</encodingType>
<maxStartupLatency type="double">60</maxStartupLatency>
<maxFrameRate type="double">30</maxFrameRate>
<maxKeyFrameInterval type="double">10</maxKeyFrameInterval>
<enableLossProtection type="bool">false</enableLossProtection>
</stream>
</streams>
</audience>

[karl_lillevold]

this setting:

<avgBitrate type="uint">480000</avgBitrate>
<maxBitrate type="uint">481000</maxBitrate>

pretty much disables VBR, effectively making it CBR. No wonder the high action would look a little smooth. Please set maxBitrate to something like 2.1X avgBitrate. Most tools do this automatically.

EDIT: Adjust maxBitrate in HPG's audience generator. It is default to 4X average bitrate, but I recommend 2.1X or so. Either is fine, and should not make a huge difference, but it's important that maxBitrate is at least 2X avgBitrate, to not limit the VBR operation.

[CruNcher]

@ karl_lilevold

i did it now with the script sagittaire posted but still im alot off from the XviD size
i used encoding complexity 80 btw not 85

8,30 MB (8.710.412 bytes) RV9
7,73 MB (8.112.498 bytes) XviD

[temporance]

@Sagittaire, I like your Matrix compare, but maybe the bitrate's a bit on the high side. I also did a DivX Kaukura encode and, IMHO, that looks better than RV9:

1st pass: -bvn1 950 -psy 0 -log "c:\divx.log" -w -b -pq 5
2nd pass: -bvnn 950 -mod 0.25 -psy 2 -log "c:\divx.log" -w -b -pq 192

However, I think your RV9 encode came in undersized.. so that might be why it looks a bit soft(?)

[Sagittaire]

@ Cruncher

The real bitrate of your XviD Sampler is 536 Kbps not 550 kbps ...

[SeeMoreDigital]

This is begining to get interesting now!

I think many of you will agree that DivX begins to superseed WMV9 at around 950kbps. However, my own tests show that WMV9 begins to overtake DivX again at around 1500kbps. But once you get to above 3000kbps I really don't think there's much difference!

It would be most helpful to learn if RV9 has 'dead zones' where other codecs begin to look better!

I recently generated some very high bitrate (8000kbps) 720x576 anamorphic tests using Xvid, RM9, DivX & WMV9. And they all looked the same to me! I used Contact Chapter 33 as my source.

However when you try and encode using larger frame sizes such as 1280x720 or 1440x816 things get wierd. As only WMV9 and RV9 can cope.

Cheers

[Sagittaire]

In the very high bitrates all the codecs are worth visually and even the MPEG2 is very powerful ...

but me what interests to me is the zone 570~950 Kbps and in this zone RV9 EHQ is the best for a normal encoding ...

In the zone 1000-1500 Kbps the WMV9 is little better for me but all the codecs are very good visually and even DivX3. It's thus useless to try to make comparative in this zone ...

With 3000 Kbps and more we can make MPEG2 720*576 with a very good quality then why use another codec?

The use of advanced codec (MPEG4 for example) in the very high bitrate is justified only if one uses very strong resolution (1024*704 or 1280*880 anamorphic for example)

For me what makes the power of a codec it's its encoding capacity in the weak BPF bit/(fps*pix) and in this zone RV9 EHQ is the best. More BPF is better and more the codecs are equivalent ...

[CruNcher]

@ Sagittaire

yes my fault again XviD also didnt hold the 550 kbps as i wanted it

but with 536 kbps

7,90 MB (8.291.299 bytes) RV9 is still a little of :P

[SeeMoreDigital]

Thanks for the reply Sagittaire.

I totally agree with you with regard to your interests. As I've always been interested in generating low bitrate encodes!

I've been very impressed with the new DivX beta's, as they really do seem to be able to offer much better quality 'low bitrate encodes' over DivXPro 5.0.5.

Unfortunately I think the Mp3 audio (.avi container) will let DivX down in the end, especially now, as RMV9 and WMV9 (.wmv) offer much better quality sounding audio at low bitrates (say 64kbps) and even 6Ch surround sound at 128kbps!

Also, yes, why bother encoding above 5000kbps with XviD, DivX RM9 and WMV9 when Mpeg2/AC3 DVD does a good job. Personnaly I think we should be experimenting with high-def type encodes at this bitrate. It's just a shame there is'nt very much in the way of 'source material' to experiment on. Well not in Europe anyway!

Cheers

[CruNcher]

@ SeeMoreDigital

the bitrate saving is sure not in WMA or Real Audio its in AAC+

http://www.codingtechnologies.com/products/aacPlus.htm

sound test have shown that its far more advanced then Vorbis, WMA or RA in low bitrate conditions :)

btw

7,83 MB (8.219.912 bytes) RV9 im nearing the target :) i think 2 encodes more and i will hit almost the same size :)

[SeeMoreDigital]

CruNcher

What you say may be right. I'll follow your link and give it a read!

But I as I have an Xcard I have to generate Mpeg4 (DivX) encodes. I've currently been testing Sorenson Squeeze 3.5 and creating 2pass VBR Mpeg4/AAC files which also play on the Xcard.

But as yet AAC+ does not appear to work on the Xcard. And it also can't be placed within an .avi container!

[CruNcher]

@ All

ok im finished i didn't hit the size 100% RV9 has some bits more :P

7,75 MB (8.136.704 bytes) RV9
7,73 MB (8.112.498 bytes) XviD

the used script was the one from Sagittaire modified avg bitrate to hit the correct size


<?xml version="1.0" encoding="UTF-8"?>
<audience xmlns="http://ns.real.com/tools/audience.2.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://ns.real.com/tools/audience.2.0 http://ns.real.com/tools/audience.2.0.xsd">
<avgBitrate type="uint">515000</avgBitrate>
<maxBitrate type="uint">1200000</maxBitrate>
<streams>
<stream xsi:type="videoStream">
<pluginName type="string">rn-videocodec-realvideo</pluginName>
<codecName type="string">rv9</codecName>
<codecProperties type="bag">
<encoderComplexity type="uint">80</encoderComplexity>
<customPacketSize type="uint">16000</customPacketSize>
</codecProperties>
<encodingType type="string">vbrBitrate</encodingType>
<maxStartupLatency type="double">60</maxStartupLatency>
<maxFrameRate type="double">30</maxFrameRate>
<maxKeyFrameInterval type="double">10</maxKeyFrameInterval>
<enableLossProtection type="bool">false</enableLossProtection>
</stream>
</streams>
</audience>


the AVS used


low=MPEG2Source("pearlharbor.d2v",idct=5).crop(2,74,716,428).LumaFilter().UnDot().trim(17500,19000).LanczosResize(640,272)
high=MPEG2Source("pearlharbor.d2v",idct=5).crop(2,74,716,428).LumaFilter().UnDot().trim(7540,8980).LanczosResize(640,272)
low+high


ok its up same addres

http://www.mufflastig.com/CruNcher/XviD/phagain/

now you can compare my result is clear for me :)

especialy look in the faces not only in the static scene also in the shaky cockpit scene

[SeeMoreDigital]

CruNcher

Your Pearl Harbour encode is most impressive.

Has it been encoded using RV9 EHQ or just RV9?

And can you confirm the source please. Was it PAL or NTSC as it looks a little speeded up?

Cheers

[CruNcher]

@SeeMoreDigital

its the PAL Version :)
the sample up on the site is RV9 EHQ as you can see from the script that was used.

And this is no Sample from a Full Encode its a special test clip i did to gurantee real conditions and not some fuzzy theatrical trailer compare like Sagittaire like to do them

and you should also watch the XviD one if you not got it pm me :)

[superdump]

From viewing the two clips, RV9 shows motion on the white background during the slow section and also appears to over compress the skin and make it look weirdly semi-flat in some areas rather than shapely as the faces should be. To its credit the line of the eye-test lady's cheek against her hair is absolutely immaculate and the detail preserved isn't too bad.

dev-api-4 exhibits what we call "block soup" on the sky during the high motion section and there are quite visible blocks elsewhere, but there is also more detail.

Both are good, neither are excellent. :) Personal preference once again.

[SeeMoreDigital]

Both are good, neither are excellent. Personal preference once again.

Well I think when you take into consideration that CruNcher's RV9-EHQ clip was encoded at an average bitrate of just 545kbps. It's excellent!

Well done

[Sigmatador]

RV9-EHQ is more plaisant for my eyes, sure ^^. MPEG-4 ASP can't really do anything against rv9 at this bitrate.

well, this is the expected results, but i'm sure PSNR-Man (aka Sagittaire) wants to do another (again) codec comparaison at low bitrate which allow him to say is jingle: "perroquet mode enabled: RV9-EHQ is the winner"

[Sagittaire]

Not i am a little to tire and i don't know any more what to make to convince ...

One tells me XviD is better with the trailer of Matrix II ... I makes the encoding and I prove that RV9 EHQ is much better ...

Then one says to me with Pearl Harbour Xvid is much better ... but it's a very slow scene and thus very compressible thus all the codecs are powerful ...

Me I use trailer where see all the types of scenes (slow, fast, sinks, luminous...) and which are thus representative of films that we encode ...

[Sirber]

Originally posted by Sigmatador
well, this is the expected results, but i'm sure PSNR-Man (aka Sagittaire) wants to do another (again) codec comparaison at low bitrate which allow him to say is jingle: "perroquet mode enabled: RV9-EHQ is the winner" IIRC the same thing was done for XviD vs DivX... :p

[CruNcher]

@ Sagittaire
you can surely call my name you dont have to use "one" for that, but i didn't wanted to offend you in any way or you beliveings everyone has his own personal preferences as for me i like it sharp first reason for that is the filttering of TVS i have an old LG one and if i watch a RV9 clip on it i can see that the blurry gets blurrier, becuase of that a XviD looks like a RV9 on TV, so i don't want to loose anymore detail. I allways trie to preserve as much details as possible and RV9 can't guarantee that as you can see on the Skins of the Actors and this scene as i said it before is a real condition scene low motion is followed by high motion if you see 2 low motion scenes here Sagittaire sorry to say you must be blind. Read what Superdump wrote it's by far the most professional analysis of what i did and what i saw when this all started. And i hope tonight we can end it ;)

RV9
+ dont blocks
+ dont looks as washy as wavelets
+ clear lines
- stil looses details

XviD (MPEG4)
+ preserves details
- blocks (most times only really visible in high motion situations)
- lines are crapy

and i dont say RV9-EHQ clear winner its not at least not for me a mix up of them both would be a Clear Winner for me ;)

btw i uploaded the XviD sequence again for the people that missed it :)

[Peters]

Originally posted by Sagittaire

Me I use trailer where see all the types of scenes (slow, fast, sinks, luminous...) and which are thus representative of films that we encode ...

It's noy very fair and you know that...
Because of his streaming origine, RV9 is almost a CBR codec (it's funny to see maxbitrate to 2x averagebitrate, i think it's useless)

So the power of RV9 works (also) with a short clip but the situation is very different for Xvid(or Divx) who are really VBR and who needs more than 2 mns to work

So a trailer isn't really representative of a whole movie encoding ...

For this reasons, Doom9 made his codecs comparisons whith full encodings

[Sagittaire]

@ Cruncher

Sorry for "one" but i use an autotomatic traductor for french -> English sometimes ...

@ Peter

Peal Harbor has some scenes with very high motion (japones attack and boat explosion ...) and with this scenes and with this bitrate i'am curious to see result with XviD and RV9 with full encoding ...

[Peters]

Originally posted by Sagittaire

@ Peter

Peal Harbor has some scenes with very high motion (japones attack and boat explosion ...) and with this scenes and with this bitrate i'am curious to see result with XviD and RV9 with full encoding ... [/B]

Understand me, even if you're right (probably) in your conclusion, i was discussing of the method, a trailer is not correct for making comparisons.

In fact, even if i found less details in RV9, this codec is very impressive. I dream of what we can expect if it will be a true VBR!

And it's a shame there are not( and never i'm afraid) RV9 Hardware players

[superdump]

Originally posted by Sagittaire
i'am curious to see result with XviD and RV9 with full encoding ...

Then do it! Maybe then the differences between the codecs will become clearer and we won't have an excuse to criticise your results.

I wholely agree with Peters. RV9 is a very good codec as are XviD and WMV9. However, you cannot judge the absolute performance of the codecs solely on one encode, let alone a theatrical trailer. It does however give a fair idea and I understand that it isn't feasible or legal to upload full film encodes.

[Sirber]

Originally posted by Peters
And it's a shame there are not( and never i'm afraid) RV9 Hardware players IIRC, Karl said that there were some people working of RV9 hardware players...

[karl_lillevold]

@CruNcher: thanks for creating such a good comparison of some of the trade-offs that are being made, and how these affect video quality. First of all, this is a tough sequence for VBR, and like I have said a few times now, there is room for improvement in the current RV9 implementation. Then, with such a long sequence of high action at a low bitrate, on one side you have the blockiness/details of MPEG-4, on the other side the smoothing of RV9. Try to enable enough PP in MPEG-4 to remove all the blocks, and check the detail level then. I don't want to start an argument over what's more preferable, blocks/details or no artifacts/less details, but will simply accept that preferences vary, just like viewing conditions. Just like you mentioned, my 5 yr old TV is so much more forgiving than my LCD computer screen, but blocks still tend to show up. Ringing, however, not so easily. So an adjustable amount of post-processing is a nice property. As you all know by now, this is not available in RV9, and some will know why. Maybe we can come up with an adjustable post-enhancer, or maybe it will have to be a bitstream change for the next RV version.

@all: Thanks again for all the useful feedback. Considering the amount of resources we have in the video codec team at RealNetworks for comparisons and experiments, doom9 is our most resourceful input forum for suggestions.

@Peters: there is nothing specifically CBR about the RV9 codec. The rate control lives outside the codec, and the current VBR implementation is not bad, it's just not as tuned and tweakable as the VBR control included with other codecs.

Take a look at a plot of the RV9 bitrate over time (http://www.lillevold.com/files/png/bitrate.png) for the Pearl Harbor extract. This was obtained by DC's RMVB analyzer tool (http://forum.doom9.org/showthread.php?s=&threadid=55392). When I have some more time I will google for a similar tool for AVI files..

Re: RV9 in h/w players. Still a little ways off, but being worked on:
http://www.realnetworks.com/company/press/releases/2002/helix_client.html
for instance by ALi Corporation, a division of Acer, and a manufacturer for decoder chips in DVD players.

[SeeMoreDigital]

I've said it many times before and I'll say it again.

Why don't we agree on a few trailers and/or chapters from the same DVD's and use them for test purposes only.

If we can all agree on some clips I would be happy to rip them onto CD-R and mail them to everybody. As long as it's not many dozens!

Suggestions please!

[Peters]

Originally posted by karl_lillevold
@Peters: there is nothing specifically CBR about the RV9 codec. The rate control lives outside the codec, and the current VBR implementation is not bad, it's just not as tuned and tweakable as the VBR control included with other codecs.



Thank's for these precisions, i have to admit i was wrong on the CBR nature of RV9.

The perspective of RV9 h/w players is exciting...

[deXtoRious]

I have a lot of spare time during the next few days, so I was thinking of creating a new video test. The question is: do you guys want it? If it will be of use to someone, I will of course make it.

BTW: Could anyone tell me why doesn't Sorenson MPEG4 codec play on my computer?

[deXtoRious]

I've only got four more free days left, so I'm gonna ask this the last time. Is anybody interested in a new video test (including RV9, xvid, divx5 (new beta), rududu, sorenson3, etc.) or should I not waste my time on making one? :confused:

[superdump]

dextorious: Feel free. Set up your test in the best way you think possible and we'll pass comment when we see the results. :)

[SeeMoreDigital]

BTW: Could anyone tell me why doesn't Sorenson MPEG4 codec play on my computer? Hi deXtoRious,

In order to obtain a response. You will have to provide a little more info than the above statement!

Cheers

[deXtoRious]

Thanks for the answer :)
Expect to see it here after 1-2 days.

[deXtoRious]

I've got everything ready to encode, just one last question before I start: could anyone please suggest a decent method to measure the speed of the encode without sitting next to the computer with a chronometer?

[superdump]

Well if you're using vdub or vdubmod and you put all your jobs in the job list you can see the times it took to encode accurate to the nearest minute.

[Soulhunter]

And when You not use VDub...
You can watch the Date/Time of the resulting Log/Movie Files...

Bye

[deXtoRious]

All right, thanx. Gonna do the test today and post results either this evening or tommorow morning.

[deXtoRious]

I did some encoding and some PSNR tests and I've got another question. In the end of the log file there is the Average PSNR and the Overall PSNR, so which should I take into account?

It looks like this:

Minimum Average Maximum
Mean Absolute Deviation: 0.0095 5.3357 68.3030
Mean Deviation: -66.0349 +0.0084 +68.2864
PSNR: 7.0340 30.9158 68.3631
Overall PSNR: 22.9140

[Assault]

It's recommended to use the overall PSNR. The difference to the average PSNR is that it treats the whole video as one frame.
But are you sure that everything worked find when you encoded your clips because an overall psnr of 22.9 is very low.

Assault

[deXtoRious]

Well, that's what surprised me... The video looks quite well visually. It's a WMV9 encode at 550 kbps. If it helps, the SSIM I got was 0.965776.

EDIT: The PSNR script was this:
#PSNR
name = "trailer wmv9 550kbps.avi"
a = avisource("trailer.avi").assumefps(100).converttoyuy2()
b = avisource(name).assumefps(100).converttoyuy2().trim(1,0)
compare(b, a, "", name+".txt", false)

[superdump]

dextorious: If using xvid with b-frames and testing the PSNR, remember to remove the first frame from the xvid clip using either ".deleteframe(0)" or ".trim(1,0)". If you do not remove the null frame your results for xvid will be significantly (and incorrectly) marred.


EDIT!!!: Or conversely, remember to remove it when using other codecs. Retest the wmv9 clip without the ".trim(1,0)".

[deXtoRious]

Well, I just retested it without the trim and the Overall PSNR is now 39.3785! :)
Thank you for the help.
Are there any other things that I should about PSNR/SSIM/VQM tests?

[superdump]

I found RV9 to be a bit tricky. If you have a working test script for it then use it, if not get back to me and I'll dig mine out for you.

Aside from that, just the xvid with b-frames thing really. I can't think of anything else anyway.

[deXtoRious]

Well, it'd be really nice if you could give me yours...:o
And I'm truly sorry about another newbie-like question, but how do I open an rmvb through avisynth???

[Sirber]

DirectShowSource

but you won't have the sound.

[superdump]

To open an rmvb via avisynth you have to use:
DirectShowSource("filename.rmvb",fps=25)

And here is the test script for RV9:

name="test.rmvb"

source=avisource("source.avi").converttoyuy2().deleteframe(0).deleteframe(n-2)
test=directshowsource(name,fps=25).converttoyuy2().deleteframe(n-1)

compare(test,source,"",name+"psnr.txt",false)

Alter the fps accordingly. Also, if you have a clip which is 1500 frames long (counting from 1 to 1500) then let n=1500 and fill in the two 'deleteframe()'s.

I believe that is correct. I'm just about to run a quick test so I'll tell you if I find anything wrong with it.

EDIT: Sirber half beat me to it. ;)

[deXtoRious]

Thanks! :)
That should be all. I'll try to include Sorenson 3 MPEG4 Video codec as well, if I'll have problems with it again, I'll write it here.

[Sirber]

Does anyone succeded in installing Sorenson on non-english OS?

[SeeMoreDigital]

Originally posted by Sirber
Does anyone succeded in installing Sorenson on non-english OS? Sirber my friend. It's a real shame (no pun intended) that you've not had sucess installing Sorenson.

Can you confirm which version you are trying to install?

Cheers

[Sirber]

Confirming:

Sorenson Squeeze 3.0: 16-bit app errror when installing video codecs
Sorenson Squeeze 3.5: 16-bit app errror when installing video codecs
Sorenson Video Codecs 3.x: 16-bit app errror

Viva French Operating Systems!!! :D

[SeeMoreDigital]

Originally posted by Sirber
Confirming:

Sorenson Squeeze 3.0: 16-bit app errror when installing video codecs
Sorenson Squeeze 3.5: 16-bit app errror when installing video codecs
Sorenson Video Codecs 3.x: 16-bit app errror

Viva French Operating Systems!!! :D Hey, just a thought. If you think the French OS is the problem. Why don't you create another boot partition and load the English version of XP?

It's aways a good idea to have at least two boot partitions anyway!

Cheers

[Sirber]

I have 2 partitions, one for OS and one for Datas. I don,t like to have 2 OS and I'm french :)

[SeeMoreDigital]

It was just a thought!

All our work PC's (all 6 of them) we have 4no 8GB approx WinXP Pro 'boot' partitions, on an independant 40GB hard drive.
One boot partition for geneneral/office use.
One boot partition for video editing. And Mpeg1,2 & 4, WMV9 (.avi and wmv) and QuickTime encoding
One boot partition for video editing. And RealMedia encoding.
One boot partition to be used as a spare, should one of the other 'boots' crash or become unstable.

All four of the 'Hire Out' PC's have WinXP Pro installed on them 3no times - and are formatted at least 6 times a year.

At home my 'Media PC' has 3 boot partitions. XP Home x2 and Linux. And my main PC. And my main PC has XP Home x3.

I take the view, that with hard disc space being so cheap and M$ operating systems being so expensive. It's more economical to install the Windows O/S's more than once! And the whole exercise saves time (and money) in the long term!

Cheers

[Sirber]

hum... as a linux addict, I like to have a long and hard uptime :) Also, I don't want to waste my precious disk space on multiple Windows installation :D

[C0mPr355]

can you tell me what trailer you used ?

I would like to run some tests as well. :)

ty.

[deXtoRious]

2 C0mPr355

Whom did you ask?

2 all

Sorry for the delays, I'm currently experienced a lack of time because of the disturbing phenomena called school :(
I will try to finish the test tommorrow, but unfortunately I can't promise anything.

[deXtoRious]

I've finally got some free time, for two days :)
I've done almost all the encodes and tests, but my psnr script (which was kindly given to me by a helpful member of this forum) for rv9 doesn't work.

n=3635
name="trailer rv9 550kbps.rmvb"
source=avisource("trailer.avi").converttoyuy2().deleteframe(0).deleteframe(n-2)
test=directshowsource(name,fps=25).converttoyuy2().deleteframe(n-1)
compare(test,source,"",name+"psnr.txt",false)

DirectShowSource just shows that RV9 is an unsupported format. Could anyone help me on this?

[C0mPr355]

i meant to direct that question at sirber. Which LOTR trailer did he use.

[karl_lillevold]

Originally posted by deXtoRious
DirectShowSource just shows that RV9 is an unsupported format. Could anyone help me on this?
You need Gabest's RealMediaSplitter:
http://forum.doom9.org/showthread.php?s=&threadid=56940

You can also enable PSNR in RV9 encoder with the calcPSNR codecProperty [d9 search for calcPSNR], to double-check the Avisynth compare numbers.

[deXtoRious]

Thanks!
I thought I was using it, but it turned out I had removed it some time ago. Another stupid mistake from my part :(

[deXtoRious]

When I try to use this script:
n=3635
name="trailer rv9 550kbps.rmvb"
source=avisource("trailer.avi").converttoyuy2().deleteframe(0).deleteframe(n-2)
test=directshowsource(name,fps=25).converttoyuy2().deleteframe(n-1)
with ssim or vqm, they start yelling that they need yv12 instead of yuv2. So I tried to replace the convert with a converttoyv12. However, after that the compare functions just kept yelling as if nothing had happened...:(
So, how do I convert this to YV12?

[deXtoRious]

Well, I hate to beg for help, but I have to, otherwise my test will never be finished :(
PLEASE HELP! :rolleyes:

[superdump]

dextorious:

n=3635
name="trailer rv9 550kbps.rmvb"
source=avisource("trailer.avi").converttoyv12().deleteframe(0).deleteframe(n-2)
test=directshowsource(name,fps=25).converttoyv12().deleteframe(n-1)

That should work just fine. Although, iirc, RV9 uses YV12 anyway so you shouldn't need the conversion for that one and if your source avi is in VBLE or YV12 huffyuv or YV12 FFV1 or some other lossless YV12 format then that shouldn't need it either.

[deXtoRious]

It still doesn't work, because due to RV9's VFR, the clips are of different length :(

[superdump]

Well that's RV9's own damn fault isn't it? :D Maybe it would be reasonable when comparing a VFR and a CFR video stream to compare the CFR stream's frames to the currently decoded image in the VFR stream? I don't know if that would be possible though. Ask the avisynth gurus. :)

[deXtoRious]

The question is simple: how the hell do I get SSIM and VQM out of RV9???
The test in the first page of this thread clearly shows it's possible.

[superdump]

OK, firstly get the following and load the relevant one for each test using "LoadPlugin("c:\blah\blah\blah.dll")"

VQMCalc DLL (http://perso.wanadoo.fr/reservoir/dl/VqmCalc-0.21.rar)

SSIM DLL (http://perso.wanadoo.fr/reservoir/dl/SSIM-0.23.rar)

They should then be used by scripting your source file, or better, using VBLE and making an AVI as your source file. Then script the following to test VQM:
LoadPlugin("C:\blah\vqmcalc.dll")
source=AVIsource("C:\Video\Source.avs")
test=AVIsource("C:\Video\XviD.avi")
return VqmCalc(source,test,"results.log")

And this for SSIM:
LoadPlugin("C:\blah\SSIM.dll")
source=AVIsource("C:\Video\Source.avs")
test=AVIsource("C:\Video\XviD.avi")
return ssim(source,test,"log.csv","score.txt",lumimask=true,scaled=true)

And just change the function name to SSIM and add .Converttoyv12() and the delete frames as necessary.

Also, did you get the latest CVS build of avisynth? (google for 'CVS Avisynth', it should be the first result) As this provides an Overall PSNR value when running a compare. The average PSNR value that is stated can be biased as it just sums the PSNR of each frame and takes and average. The overall PSNR does something like sum all the MSEs (mean-squared-errors) of each frame, takes an average and then puts that into the PSNR formula giving an accurate output for the PSNR.

I hope this helps. :)

EDIT: Updates and changes to functions.

[superdump]

Oh, and the conditions of the VQMCalc plugin is that both dimensions of your video have to be an integer multiple of 8 (mod 8) and, as you know, they have to be in YV12 format for both SSIM and VQMCalc.

EDIT: Errors.

[lilhobo]

for a newbie into thread will means nothing, where can i go to get explanation for the diffrent tests ( encoding methods i guess, dll and such)

so which codec has the best results?

[CQ]

hey
1. after writing the script what do you do with it? run in VDub or what? and... please be nice to me :rolleyes:

2. how do you obtain those graphs? with excel? how? plz, someone, i really need to know. (to do a subj. test its easy but when it comes to obj i`m stuck, its not my fault that i`m almost a director (stil studying) not a video engineer :)


please superdump make some time and write a full (or almost) guide (with writing the avs, how to run the avs :( , how to make graphs from the cvs data....). a stupid, i mean very stupid codec tester guide :)

[superdump]

The method hasn't changed in the last year, apart from possibly using the CompareYV12 dll where appropriate. If you use either CompareYV12 or the internal Compare function you can use the Java XviD Stats Viewer to create some very nice antialiased graphs but I don't know how Sagittaire makes his SSIM graphs, he may have written a program for it. And yes, you do just play the resulting script in Vdub or mplayerc or whatever until it's played all the way through and then close it.

For XviD with b-frames in AVI you have to remove the first frame (unless using packed bitstream). Even then it's sometimes a real pain to get it to work. Thankfully everwicked is working on optimising his VQS objective program so that we can use that.