QuEnc .71 VS HC .20

[gizzin]

The source used was Gargoyles Season 2 Disc 1 R1. My system is a amd am2 4200+ x2 with 1 gig of ram. It was interlaced so I didn't even bother using CCE for this. :). For HC I used the best profile with the default matrix. As for Quenc I used high quality, 2pass with VBR, and trellis quantization. Obviously since I have 2 cores I had multiple processors enabled.

Overall Bitrate : 2,961/2,369Kbs
HC encoding time Phase II ENCODING completed in 161 minutes.
QuEnc encoding time Phase II ENCODING completed in 1264
minutes.
Hc file size 4,683,005,952 bytes
QuEnc file size 4,624,261,120 bytes

Now onto the quality part. I watched both of these back to back. For HC the noise around the edges is quite noticeable, not very bad but its there. I noticed very little pixelation, there was some but it was very minor, I would say not even noticeable. QuEnc little different story here the noise around edges slim to none comparable to the original and thats a low avg bitrate that was used. But it seemed like it just moved it to the flat areas because those are nice and blocky.

Remember this is just my opinion. Well as for speed you can see the encoding times. I don't care about encoding times more so quality of the video. Saying what's better in this one is tough decision there both good. But if I had too it would probably be HC. It was more accurate and 1/10th the encoding time.

[dragongodz]

the latest version of QuEnc is 0.72.

1/10th the encoding time.
try it without trellis next time. its been said over and over and over again, trellis is completely not optimised and a great speed killer for a very small(if at all) benefit.

i am very tempted to just remove it since people use it and then wonder why it takes so long.

hank315 has been doing some very detailed speed tests. hopefully he will post them when he has finished and others can try to emulate the types of tests he has done with different cpus.

[techreactor]

try it without trellis next time. its been said over and over and over again, trellis is completely not optimised and a great speed killer for a very small(if at all) benefit.

i am very tempted to just remove it since people use it and then wonder why it takes so long.


I think you should remove it immediately, it brings a bad name to your encoder especially when it adds no/little value to the quality part.

[BR7]

@dragongodz
I love the trellis feature and I for one hope you keep it available for those of us who like it :) I see no reason to remove it especially since you have the option to enable or disable the feature.And I definitely don't think it brings a bad name to the encoder

[hank315]

hank315 has been doing some very detailed speed tests. hopefully he will post them when he has finished and others can try to emulate the types of tests he has done with different cpus.
Here it is.


Hardware
--------
MB Asus P5B de luxe
CPU Intel core 2 duo E6600, overclocked, running at 3.16 GHz
Memory 2*1 GB Corsair CM2X1024-8500C5, 5-5-5-15
OS XP Pro


Movie
-----
Matrix Revolutions, PAL, movie only (194544 frames), running time 2:09:42
Using DVD-RB 1.22 Pro (multiple encoder processes) and Avisynth 2.5.6.0


Encoders used
------------- time size

CCE basic
- 2 pass 34 min 4.32 GB

Procoder 2
- High Speed 39 min 4.32 GB
- High Quality 40 min 4.32 GB
- Highest Quality 49 min 4.32 GB
- Mastering Quality 55 min 4.32 GB

QuEnc072
- 1 pass VBR, no HQ 16 min 4.28 GB
- 1 pass VBR, HQ 22 min 4.28 GB
- 1 pass VBR, HQ/OPV 22 min 3.49 GB
- 2 pass VBR, no HQ 31 min 4.32 GB
- 2 pass VBR, HQ 46 min 4.32 GB
- 2 pass VBR, HQ, trellis 91 min 4.30 GB
- 2 pass VBR, HQ, extreme 143 min 4.30 GB
- 2 pass VBR, HQ, extreme, trellis 194 min 4.26 GB

HC020
- 2 pass fast 30 min 4.32 GB
- 2 pass normal 31 min 4.32 GB
- 2 pass best 45 min 4.32 GB
- 1 pass fast/OPV 20 min 4.23 GB
- 1 pass normal/OPV 23 min 4.21 GB
- 1 pass best/OPV 32 min 4.20 GB

AutoQmatEnc
- 4 pass without Qmatopt 60 min 4.32 GB
- 4 pass with Qmatopt 86 min 4.31 GB

Rejig (transcoder)
- default 4 min 4.26 GB


Times shown are encoding times only.

[jdobbs]

Great info, Hank. Thanks.

[Rippraff]

Wow, incredible speeds, Hank. :eek:
I should put some money aside and get rid off my old XP-M I guess. :rolleyes:

:goodpost: thanks.

Cu Rippraff

[techreactor]

Good job hanks, I believe you didnt test for AutoQmatEnc with and without QmatOP enabled.

Great effort!!!

[techreactor]

@dragongodz
I love the trellis feature and I for one hope you keep it available for those of us who like it :) I see no reason to remove it especially since you have the option to enable or disable the feature.And I definitely don't think it brings a bad name to the encoder


@dragongodz

I agree and disagree partly to this. If its an option , it might be left like that, with a remark in your help file about the increased timelines.

But on the other hand if it does not add any value/little value to the encode (as you say), no point keeping it as an option too.

[gizzin]

To be honest I didn't think that encoding time is right, I remember doing encodes awhile ago, It took awhile but no that long it was like a day heh. Theres probably something wrong with my computer. Im not trying to make your encoder look bad. And anything to increase quality would/should be left in. Your mpeg2 encoder is of high quality theres no doubt about that. If you told me you put 3 more options that increased quality but slow downed the encoder drastically I'd still use it. It should always be quality not speed. Sorry for any inconvience.

[sockeye]

@dragongodz

I agree and disagree partly to this. If its an option , it might be left like that, with a remark in your help file about the increased timelines.

But on the other hand if it does not add any value/little value to the encode (as you say), no point keeping it as an option too.
It is an option that is very clearly marked (very slow)
in the settings menu.
Some users love it. I vote leave it alone. :)

[dragongodz]

I believe you didnt test for AutoQmatEnc with and without QmatOP enabled.
hank315 tested the 2 free encoders that come with DVD-RB and also 2 commercial ones with MANY options taking MANY hours. compare that to other tests and see the difference and be thankful.

To be honest I didn't think that encoding time is right, I remember doing encodes awhile ago, It took awhile but no that long
yes it did seem rather longer than i would have expected even with trellis. still my suggestions would stand from my previous posts.
actually i should also say that interlaced encoding with QuEnc should be slower than its progressive encoding. the reason being it uses a different "interlaced" ME etc.
Sorry for any inconvience.
no inconvenience, its just your test was too limited to paint a true picture of the options available. :)

about hank315's tests. it would be nice if others with different cpus could try to do more complete tests like his. that is with multiple settings etc. no i do not expect people to have the time to test everything nor do i expect them to have both the commercial encoders. however people can download the trial version of CCE for example to test and HC and QuEnc are free and included.
for testing the main ones i would suggest of note for HC and QuEnc are

QuEnc072
- 1 pass VBR, no HQ
- 1 pass VBR, HQ
- 2 pass VBR, no HQ
- 2 pass VBR, HQ

HC020
- 2 pass normal
- 2 pass best
- 1 pass normal/OPV
- 1 pass best/OPV

these being a good minimum of options to test IMHO. with these you can get a good idea of speeds achievable. though 1 pass methods are better suited for higher bitrate encodes of course.

if it does not add any value/little value to the encode
extreme type of settings are for what they suggest. that is extremly hard footage or very low bitrates where you need to try and squeeze every trick in the book to gain quality. they do cost.
here is an example where the extreme(not trellis) setting made a noticable difference.
http://forum.doom9.org/showthread.php?t=110992

so i wont remove trellis ... for now. :devil:
any further discussion for trellis should be taken to QuEncs thread of course.

[Fishman0919]

I did my own little test with diff encoders,
it 2014 frames from Casino_Royale airport scene.
The Org m2v demuxed with DGIndex 1.49 Beta 16 is
5.858 Mbps size 60,248 KB. I reencoded for this test at 2500k and a max of 7500k.

LoadPlugin("c:\test\dgmpgdec\DGDecode.dll")
mpeg2source("c:\test\test.D2V",idct=2)
ConvertToY##2()

GOP=12 (Auto for HC)
idct=9
Matrix=Mpeg2 standard (except AQE in 3 encodings)
Windows XP
AMD 64 3400+, 1g mem

I tested both CCE SP 2.70.02.12 and CCE SP2 1.00.00.13 but just put the results from CCE SP because all results were the same... Time, Size and Qual


Time Size PSNR Score
--------------------------------------------------------------------------------------------------------------
CCE Basic 2.70 | :59 sec| 25,588 KB | 45.2893
--------------------------------------------------------------------------------------------------------------
CCE SP 2.70 | | |
2-pass AQM | 1:00 | 25,611 KB | 45.3050
--------------------------------------------------------------------------------------------------------------
2-pass no AQM | :59 | 25,609 KB | 45.3133
--------------------------------------------------------------------------------------------------------------
3-pass AQM | 1:29 | 25,668 KB | 45.3217
--------------------------------------------------------------------------------------------------------------
3-pass no AQM | 1:29 | 25,687 KB | 45.3597
--------------------------------------------------------------------------------------------------------------
4-pass AQM | 1:59 | 25,676 KB | 45.3287
--------------------------------------------------------------------------------------------------------------
4-pass no AQM | 1:59 | 25,659 KB | 45.3535
--------------------------------------------------------------------------------------------------------------
10-pass no AQM | 4:59 | 25,669 KB | 45.3553
--------------------------------------------------------------------------------------------------------------
100pass no AQM | 50:46 | 25,669 KB | 45.3553
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
OPV Q-40 | :30 | 25,357 KB | 45.4022
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
HC .20 |
Best 2-pass | 1:47 | 25,627 KB | 45.7679
--------------------------------------------------------------------------------------------------------------
Normal 2-pass | 1:15 | 25,639 KB | 45.5196
--------------------------------------------------------------------------------------------------------------
Fast 2-pass | 1:11 | 25,632 KB | 45.3756
--------------------------------------------------------------------------------------------------------------
QuEnc .72 | | |
Nothing 2-pass | 1:00 | 25,028 KB | 45.3575
--------------------------------------------------------------------------------------------------------------
HQ | 1:29 | 24,190 KB | 45.3911
--------------------------------------------------------------------------------------------------------------
HQ Trellis | 2:32 | 24,271 KB | 45.1942
--------------------------------------------------------------------------------------------------------------
HQ Trellis | | |
Extreme | 4:58 | 24,154 KB | 45.2267
--------------------------------------------------------------------------------------------------------------
HQ Extreme | 3:47 | 24,800 KB | 45.4234
--------------------------------------------------------------------------------------------------------------
AQE | | |
No QMatOp | 1:26 | 25,692 KB | 45.6325
--------------------------------------------------------------------------------------------------------------
Full QMatOp | 3:14 | 25,629 KB | 44.0711
--------------------------------------------------------------------------------------------------------------
Mod QMatOp | | |
max_intra=95 | | |
max_inter=85 | 3:13 | 25,630 KB | 44.2795
--------------------------------------------------------------------------------------------------------------
Mod QMatOp | | |
max_intra=85 | | |
max_inter=45 | 3:13 | 25,628 KB | 44.5490
--------------------------------------------------------------------------------------------------------------
Procoder 2 | | |
Master Qual | 1:57 | 25,636 KB | 45.6491
--------------------------------------------------------------------------------------------------------------
Highest Qual | 1:50 | 25,635 KB | 45.6382
--------------------------------------------------------------------------------------------------------------
High Qual | 1:31 | 25,643 KB | 45.1911
--------------------------------------------------------------------------------------------------------------
Speed Qual | 1:29 | 25,641 KB | 44.8248
--------------------------------------------------------------------------------------------------------------
TMPGEnc 4.0 | | |
Highest with | | |
error correction | 3:26 | 25,573 KB | 44.6268
--------------------------------------------------------------------------------------------------------------
Standard | 3:19 | 25,577 KB | 44.4551
--------------------------------------------------------------------------------------------------------------
MainConcept | | |
Default | 1:36 | 25,649 KB | 43.4772
--------------------------------------------------------------------------------------------------------------

100 pass test still in progess....DONE


Edit: Added OS, CPU, notes and fixed spelling


like dragongodz said (post below)... do not take metric scores as absolutes... just thought I would add it seeing that I had all this .m2v files from the test.

[Rippraff]

Nice :)

What do the last values mean?
e.g. CCE Basic 2.70 | :59 sec | 25,588 KB | 45.2893

Cu Rippraff

[Fishman0919]

Nice :)

What do the last values mean?
e.g. CCE Basic 2.70 | :59 sec | 25,588 KB | 45.2893

Cu Rippraff

Sorry, I just edited it.... PSNR (http://en.wikipedia.org/wiki/Peak_signal-to-noise_ratio) quality score.... higher is better

[Rippraff]

I was first thinking of reduction percentage. :p

I'm really thankful that this test shows that PSNR score won't get better with more than 3 CCE passes! :)

Cu Rippraff

[dragongodz]

Fishman0919 - thanks for the tests. can you also include what cpu you are using aswell ?

i must say i never expected to see anyone insane enough to use 100 passes. good thing it was just a test on a short clip. imagine on something like a 2 hour plus movie. :eek:

PSNR quality score.... higher is better
and the standard warning applies. do not take metric scores as absolutes. they do not measure what a person will find more pleasing. infact since everyones perception is different thats practically impossible.

[smok3]

so, can you see any difference between worse and best encode?

[techreactor]

Sorry, I just edited it.... PSNR (http://en.wikipedia.org/wiki/Peak_signal-to-noise_ratio) quality score.... higher is better

And how did you get them, using what tool ???

[Fishman0919]

i must say i never expected to see anyone insane enough to use 100 passes. good thing it was just a test on a short clip. imagine on something like a 2 hour plus movie. :eek:

I did Casino_Royale movie only for the fun of it with DVD-RB 1.23.1 CCE SP2 100 passes.... it took over 2 days to finish.... man was I disappointed :mad: The test copies I made with CCE Basic, HC.20 and QuEnc looked the same.;)

[Fishman0919]

so, can you see any difference between worse and best encode?

The worst score was MainConcept (43.4772) and best was HC .20 (45.7679) and to me there is a nite and day difference.... But I could also see a nite and day difference between AQE Full QMatOp (44.0711) and MainConcept (43.4772). To me the only encodings that looked bad were MainConcept and TMPGEnc 4.0 but someone else may like what is produced by those encoders for this scene.

[Rippraff]

And not to forget that bitrate is pretty low and HC is known to do an excellent job with low bitrates.
Would be interesting how it looks like with an average say 3500 or 4000.

Cu Rippraff

[Fishman0919]

And not to forget that bitrate is pretty low and HC is known to do an excellent job with low bitrates.
Would be interesting how it looks like with an average say 3500 or 4000.

Cu Rippraff


I started the test at 3750k but all result were very similar to each other so I dropped the bitrate down.... maybe tonite I'll finish that test and post my findings.

[hank315]

I did Casino_Royale movie only for the fun of it with DVD-RB 1.23.1 CCE SP2 100 passes.... it took over 2 days to finish.... man was I disappointed The test copies I made with CCE Basic, HC.20 and QuEnc looked the same.
Cool, more than 2 days computer power down the drain :D :D
But it shows CCE doesn't get any better after 3 passes.
Thanks for running the tests.

AutoQmatEnc encodes are added to the list here (http://forum.doom9.org/showthread.php?p=970400#post970400)
Also Rejig was tested, a transcoder, well that's really fast...

[Rippraff]

Also Rejig was tested, a transcoder, well that's really fast...
I think I'ld need more time to copy a dvd from one drive to another... :D

Cu Rippraff

[Fishman0919]

Rejig (transcoder)
- default 4 min 4.26 GB


lol

[hank315]

I couldn't believe it myself, the whole thing,PREPARE/TRANSCODE/REBUILD, was done in 12 min :)

[18:40:48] One Click encoding activated...
-----------------
[18:40:48] Phase I, PREPARATION started.
- DVD-RB v1.22
- AVISYNTH 2.5.6.0
- ReJig transcoder selected.
- "Movie Only" mode is enabled.
- Source: RIP
- VTS_01: 2,955,451 sectors.
-- Scanning and writing .D2V & .AVS files
-- Video demultiplexing...
-- Processed 194,544 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 82.8%
- Overall Bitrate : 4,245Kbs
- Space for Video : 4,032,244KB
- HIGH/LOW/TYPICAL Bitrates: 5,184/4,021/4,245 Kbs
[18:43:10] Phase I, PREPARATION completed in 3 minutes.
[18:43:10] Phase II ENCODING started
- Reducing size of VTS_01 segment 0
- Reducing size of VTS_01 segment 1
- Reducing size of VTS_01 segment 2
- Reducing size of VTS_01 segment 3
- Reducing size of VTS_01 segment 4
- Reducing size of VTS_01 segment 5
- Reducing size of VTS_01 segment 6
- Reducing size of VTS_01 segment 7
- Reducing size of VTS_01 segment 8
- Reducing size of VTS_01 segment 9
- Reducing size of VTS_01 segment 10
- Reducing size of VTS_01 segment 11
- Reducing size of VTS_01 segment 12
- Reducing size of VTS_01 segment 13
- Reducing size of VTS_01 segment 14
- Reducing size of VTS_01 segment 15
- Reducing size of VTS_01 segment 16
- Reducing size of VTS_01 segment 17
- Reducing size of VTS_01 segment 18
- Reducing size of VTS_01 segment 19
- Reducing size of VTS_01 segment 20
- Reducing size of VTS_01 segment 21
- Reducing size of VTS_01 segment 22
- Reducing size of VTS_01 segment 23
- Reducing size of VTS_01 segment 24
- Reducing size of VTS_01 segment 25
- Reducing size of VTS_01 segment 26
- Reducing size of VTS_01 segment 27
- Reducing size of VTS_01 segment 28
- Reducing size of VTS_01 segment 29
- Reducing size of VTS_01 segment 30
- Reducing size of VTS_01 segment 31
- Reducing size of VTS_01 segment 32
- Reducing size of VTS_01 segment 33
[18:46:59] Phase II ENCODING completed in 3 minutes.
[18:46:59] Phase III, REBUILD started.
- Processing VTS_01
- Reading/processing TMAP table...
- Rebuilding seg 0 VOBID 2 CELLID 1
- Rebuilding seg 1 VOBID 2 CELLID 2
- Rebuilding seg 2 VOBID 2 CELLID 3
- Rebuilding seg 3 VOBID 2 CELLID 4
- Rebuilding seg 4 VOBID 2 CELLID 5
- Rebuilding seg 5 VOBID 2 CELLID 6
- Rebuilding seg 6 VOBID 2 CELLID 7
- Rebuilding seg 7 VOBID 2 CELLID 8
- Rebuilding seg 8 VOBID 2 CELLID 9
- Rebuilding seg 9 VOBID 2 CELLID 10
- Rebuilding seg 10 VOBID 2 CELLID 11
- Rebuilding seg 11 VOBID 2 CELLID 12
- Rebuilding seg 12 VOBID 2 CELLID 13
- Rebuilding seg 13 VOBID 2 CELLID 14
- Updating NAVPACKS for VOBID_02
- Rebuilding seg 14 VOBID 3 CELLID 1
- Rebuilding seg 15 VOBID 3 CELLID 2
- Rebuilding seg 16 VOBID 3 CELLID 3
- Rebuilding seg 17 VOBID 3 CELLID 4
- Rebuilding seg 18 VOBID 3 CELLID 5
- Rebuilding seg 19 VOBID 3 CELLID 6
- Rebuilding seg 20 VOBID 3 CELLID 7
- Rebuilding seg 21 VOBID 3 CELLID 8
- Rebuilding seg 22 VOBID 3 CELLID 9
- Rebuilding seg 23 VOBID 3 CELLID 10
- Rebuilding seg 24 VOBID 3 CELLID 11
- Rebuilding seg 25 VOBID 3 CELLID 12
- Rebuilding seg 26 VOBID 3 CELLID 13
- Rebuilding seg 27 VOBID 3 CELLID 14
- Rebuilding seg 28 VOBID 3 CELLID 15
- Rebuilding seg 29 VOBID 3 CELLID 16
- Rebuilding seg 30 VOBID 3 CELLID 17
- Rebuilding seg 31 VOBID 3 CELLID 18
- Rebuilding seg 32 VOBID 3 CELLID 19
- Rebuilding seg 33 VOBID 3 CELLID 20
- Updating NAVPACKS for VOBID_03
- Updated VTS_C_ADT.
- Updated VTS_VOBU_ADMAP.
- Updated IFO: VTS_01_0.IFO
- Correcting VTS Sectors...
[18:52:30] Phase III, REBUILD completed in 6 minutes.

Done.
[18:52:30] PREPARE/ENCODE/REBUILD completed in 12 min.

[Fishman0919]

I couldn't believe it myself, the whole thing,PREPARE/TRANSCODE/REBUILD, was done in 12 min

For those who need their movie quick fix.

[Fishman0919]

Added CCE SP OPV to test LINK (http://forum.doom9.org/showthread.php?p=973580#post973580)

[techreactor]

Apart from the speed tests on transcoders(pls include shrink too now), it would be good to know/see some screenshots of hard to encode scenes with both DVD shrink and Rejig. I wonder who can produce better results amongst these two.

[dragongodz]

hard to encode scenes with both DVD shrink and Rejig
if you use the AEC setting in dvdshrink it should look better than Rejig for tough scenes when compressing by a large amount. AEC means Adaptive Error Concealment(or Correction) so its trying to hide/reduce the imperfections from the transcode process. of course that also takes much longer.

i hope anyone that has posted comments like "i tried QuEnc ages ago and it wasnt very good" both here and on other forums takes note of not only the speed results but comments. maybe then they will think again before crapping on others work which they couldnt do themselves. :angry:

[techreactor]

Is this thread reviving old memories (http://forum.doom9.org/showthread.php?t=98633)

[jdobbs]

I think it's interesting how these PSNR comparisons come out every year or so, we look at them in awe, then we discuss them, then we slam them, then we all agree that it's only what you view that really counts, we then say a few words about how video reduction modeling has to be based upon visual perception and numeric models always fail to take that into consideration, then we all move on.

Here's the bottom line. I have, on many occasions, taken two videos and compared them visually only to find that the one with the better PSNR doesn't look as good. It may on another... but the point is that I can't seem to make a good connection between PSNR and quality except at a very macro level (large differences).

A good example is a transcode that may not touch the I frames (many transcodes don't). The fact that the I-Frame is identical to the original artificially raises the PSNR values -- but the differences between the I, P, and B frames hurt the playback visually. Inconsistency is a bad thing in playback.

Anyway... I guess we can again boldly go where we've gone before. I know we will. :)

@Fishman0919

No offense intended toward you. I know you do good testing because I've seen it before. Just wanted to make the PSNR comment before all the noobs got their panties in a bunch about this "new method" of comparison.

[Fishman0919]

@Fishman0919

No offense intended toward you. I know you do good testing because I've seen it before. Just wanted to make the PSNR comment before all the noobs got their panties in a bunch about this "new method" of comparison.

None takin' ;)

Just had all the .m2v file with some time to kill.

and the standard warning applies. do not take metric scores as absolutes. they do not measure what a person will find more pleasing. infact since everyones perception is different thats practically impossible.

... the test was just food for thought.

[dragongodz]

the test was just food for thought.
and such an extensive test is highly appreciated. it was not meant towards you, but as jdobbs said a warning to newbs looking at it and thinking psnr scores were going to tell them what they would think looked better. :)

[Fishman0919]

and such an extensive test is highly appreciated. it was not meant towards you, but as jdobbs said a warning to newbs looking at it and thinking psnr scores were going to tell them what they would think looked better. :)

Oh np :) .... again none takin'.... I should have noted that in my first post with the test... I have since edited that post.

[techreactor]

The answer lies in MSU Perceptual Video Quality tool (http://compression.ru/video/quality_measure/perceptual_video_quality_tool_en.html), as compared to the MSU Video Quality Measurement Tool (http://compression.ru/video/quality_measure/video_measurement_tool_en.html) which only uses certain metrics/ calculations like PSNR for final judgement in 5 min, the MSU Perceptual Video Quality tool needs much more evaluation and patience/people for the final judgement and takes into consideration the human visual factors also.

[chickenmonger]

In an effort to get a wider range of CPU types and times, I did some similar testing on my own computer.

Hardware
--------
MB SiS M848A
CPU AMD Athlon XP 2200+ (1800 Mhz) (MMX, 3DNow, SSE1)
Memory 768 MB (512 MB PC2700, 256 MB PC2100)
OS XP Pro


Movie
-----
Buffy the Vampire Slayer Season 1, Disc 1 NTSC
Frame Count: 276375
Running Time: 3 hrs, 7 min

DVD-RB 0.98.2 Freeware Edition

Encoders used
-------------

HC020
- 2 pass fast 297 minutes
- 2 pass normal 311 minutes
- 2 pass best 416 minutes

QuENC
- 1 pass normal 114 minutes
- 2 pass normal 238 minutes
- 1 pass HQ 153 minutes
- 2 pass HQ 422 minutes
- 1 pass HQ, trellis 261 minutes

AutoQMatEnc
- 2 pass, QMATOP 735 minutes


That 2 pass HQ time seems off. I'll try it again later.

[Rumbah]

For the German speaking people:

You can download a test DVD at http://technik.movie2digital.de/board.php?boardid=207&sid=386eceff453e342381f6744749b93ff9 and participate in a mpeg2 blind test (and a mp4 one).

[techreactor]

i hope anyone that has posted comments like "i tried QuEnc ages ago and it wasnt very good" both here and on other forums takes note of not only the speed results but comments. maybe then they will think again before crapping on others work which they couldnt do themselves. :angry:

Well let me comment on Quenc on my first usage (test), I found that its as fast as any other encoder (HC/AQM) in its speed and quality is also decent on 2 pass HQ (but can be tweaked a bit more for low bitrates).

I didnt use Quenc earlier, so I cannot comment on how it was earlier ;)

[techreactor]

based on guidance given by Hank, I studied about DCtune (Used in a variant by SAPSTAR in QmatOP), in short DCtune is about perceptual image compression as seen best by Human eyes, for beginner on DCtune (like me) this article (http://scien.stanford.edu/class/psych221/projects/98/dctune/yuke/index.htm)helped.

[sockeye]

For the German speaking people:

You can download a test DVD at http://technik.movie2digital.de/board.php?boardid=207&sid=386eceff453e342381f6744749b93ff9 and participate in a mpeg2 blind test (and a mp4 one).

I like your approach to a controlled test where everyone uses the same files for comparison.
Hopefully someone can setup a test bed like this (beyond my capabilities) for English speakers.

[Fishman0919]

Here is another test with the movie ERAGON_WS

sorry hank315... copied your format


Hardware
--------
MB Asus K8N - VM
CPU AMD 64 3400+
Memory 2*512 PC-3200 PNY
OS XP Pro


Movie
-----
ERAGON_WS R1, Full Disc org size 6.42 GB (6,895,667,200 bytes)
Using DVD-RB 1.23.1 Pro w/idct 64-bit IEEE-1180 Ref and Avisynth 2.5.6.0


Encoders used
------------- time size

CCE Basic 2.70.02.15
- 2 pass 97 min 4.31 GB
- 2 pass idct 6 78 min 4.31 GB
- 2 pass NaN's idct 6 77 min 4.31 GB

CCE SP 2.70.02.12
- 2 pass 97 min 4.31 GB
- 2 pass idct 6 78 min 4.31 GB
- 2 pass NaN's idct 6 77 min 4.31 GB

CCE SP2 1.00.00.15
- 2 pass 97 min 4.31 GB
- 2 pass idct 6 78 min 4.31 GB
- 2 pass NaN's idct 6 77 min 4.31 GB

CCE SP 2.50.1.00
- 2 pass 96 min 4.31 GB
- 2 pass idct 6 77 min 4.31 GB
- 2 pass NaN's idct 6 76 min 4.31 GB

Procoder 2.04.02.00
- High Speed 135 min 4.31 GB
- Mastering Quality 170 min 4.31 GB

QuEnc 0.72
- 2 pass VBR, no HQ 103 min 4.29 GB
- 2 pass VBR, HQ 140 min 4.29 GB
- 2 pass VBR, HQ idct 6 116 min 4.29 GB

HC 0.20
- 2 pass fast 116 min 4.31 GB
- 2 pass best 169 min 4.31 GB
- 2 pass best idct 6 146 min 4.31 GB

AQE 0.34.0.2
- 2 pass wo/QMatOp 208 min 4.33 GB
- 2 pass w/QMatOp 269 min 4.31 GB

Times shown are encoding times only.


Edit: added new test

[chickenmonger]

@Fishman0919
Wow. Your QuEnc 2 pass no-HQ and HcEnc 2 pass fast times are really close, whereas mine are way different.

Looks like those SSE2 optimizations in HcEnc really make a difference, especially on the "fast" setting.

[Fishman0919]

Added new tests... LINK (http://forum.doom9.org/showthread.php?p=975525#post975525)

[chickenmonger]

And just for insanity's sake, DVD backup benchmarking on some real hardware.

Hardware
--------
CPU Intel Pentium II 266 Mhz
Memory 64 MB
OS XP Pro


Movie
-----
Buffy the Vampire Slayer Season 1, Disc 1 NTSC
Frame Count: 276375
Running Time: 3 hrs, 7 min

DVD-RB 0.98.2 Freeware Edition

Encoders used
-------------

QuENC
- 1 pass normal 966 minutes

[Fishman0919]

And just for insanity's sake, DVD backup benchmarking on some real hardware.

Hardware
--------
CPU Intel Pentium II 266 Mhz
Memory 64 MB
OS XP Pro


Movie
-----
Buffy the Vampire Slayer Season 1, Disc 1 NTSC
Frame Count: 276375
Running Time: 3 hrs, 7 min

DVD-RB 0.98.2 Freeware Edition

Encoders used
-------------

QuENC
- 1 pass normal 966 minutes


LOL.... how must coal did you have to shovel to finish that encoding ;)

[sockeye]

LOL.... how must coal did you have to shovel to finish that encoding ;)

Not much memory left to go around downsteam of XP, on that setup, eh Fishman? :)

[kolak]

I don't know how Procoder 2 and CCE can have almost the same speed.
There is some problem with your setup.
There is no encoder which is faster than CCE. With proper disk array and quad core PC (3.0Ghz) is about 10 times faster than RT.
Procoder 2 with "split encoding for multiple CPU" option used (can degrade quality) is at most 4 RT.

[techreactor]

Procoder can run 2 instance in rebuilder during the encode whereas CCE cannot, thats how its the same speeds.