How can I unleash Xvid quality for 7 CDs?

[dihelson]

Hello, friends,
Perhaps my question is a bit crazy, but I'd like to know the best parameters for using Xvid to convert a DVD to the maximum quality disregarding the number of discs used.
May be even equal to the DVD size, just for curiosity, say 7 or 8 CDs...

My actual settings don't go more than 3 CDs, on most films, they are here:

================================================
6 ultra high
MPEG
XVID
1-mode decision
300
1
Use chroma motion
Quarterpel
2
150
100
0
DX-50 B-VOP compatibility

Quantization:
2, 31, 2, 31
DEBUG: Chroma optimizer

On the 2 passes...

==================================================

NOTE:
It would be interesting if we could raise the size of the file, but could improve the quality also. raising only the size of the file don't have any sense from making them on 2 CDs only.

thanks in advance,
Dihelson Mendonca

[mf]

Originally posted by dihelson
It would be interesting if we could raise the size of the file, but could improve the quality also. raising only the size of the file don't have any sense from making them on 2 CDs only.
Exactly my field of expertise :D. Try this in AVISynth:

Import("SharpTools-v0.3.avs (http://mf.creations.nl/avs/functions/SharpTools-v0.3.avs)")
SharpResize(1280, 544)

You may want to add some noise to make your big resolution look more natural. Oh and btw, beware for your poor CPU, cause decoding this will require beefy specs. It is the only way to get XviD to encode above-DVD quality though. Oh and use MPEG-Custom with the HVS Best matrix. It'll increase your size more and give more quality.

[Acaila]

1) Use 1-pass mode at quantizer 1, 2-pass makes no sense in this case
2) Disable B-frames
3) Don't resize your video, cropping the black away shouldn't be a problem -> i.o.w use anamorphic encoding.
4) Set maximum I-frame interval to 1, this forces all keyframes
5) If you feel like it use a custom matrix consisting of only 1's.

Voila, a maximum quality video. But the size will be huge (LARGER than the DVD original) and it probably won't be playable by your computer either :D. To keep the filesize somewhat acceptable only use #1-3 and use the MPEG matrix (since #4 and #5 only raise the filesize with an invisible increase in quality).

[Tommy Carrot]

If you want maximum quality, and the filesize doesn't matter, simply use 1-pass quantizer based encoding mode with quantizer 2. Of course set bframes and qpel off.

No need for 2 pass for this.

EDIT: quantizer 1 is overkill. Don't listen to Acaila! :D

[symonjfox]

It's quite interesting topic. I think that Xvid can be improved on high bitrates. We all use it for 1-2 CD rips, so we pratically test it everyday on the same settings.

It would be interesting to test Xvid with uncompressed material and try to get the best from Xvid, in all settings.
I also think that it will be more difficult to see quality improvement from 8000 kbs to 9000 kbs in Xvid (while you clearly see differences from 800 to 1000 kbs in most cases).

For example it would be fine if we could use Xvid to create MPEG4 DVDs, but we need a perfect source (a DVD is not enough).

I don't know how explain ... It's like we all want to decrease the bitrate, giving good quality, but quality freaks (like me) still want to get the best from their work. It's like MP3 @ 192 kbs ... it's high bitrate, but quality is very very good, but we all want AAC+ @ 64 (that sounds good, but MP3 192 is still better and has a good compression ratio). I hope to be clear.

[Selur]

At higher datarates more accurate custom matrices help. (search the forum)

Cu Selur

Ps.: Here the custom settings I use atm for high bitrate captures:
IntraFrame-Matrix
_8_10_11_12_13_15_16_16
10_10_11_13_14_15_17_18
11_11_13_14_15_17_18_21
12_13_14_15_17_18_21_22
13_14_15_17_19_20_22_26
15_15_17_19_20_22_26_28
16_16_19_20_22_24_27_64
16_19_20_22_24_26_64_128

InterFrame-Matrix
_8__9_10_11_11_13_13_16
_9__9_11_11_13_14_16_16
10_11_12_13_14_16_16_21
10_12_13_14_15_17_21_27
12_13_14_15_17_20_27_32
12_14_15_17_20_27_32_34
15_15_18_19_27_32_34_72
15_18_19_27_32_34_72_144

[trodas]

Speaking about quality - okay, i set the MPEG-Custom, but where to set the "HVS Best matrix"...??? :rolleyes:
Pls, help.

[symonjfox]

Originally posted by Selur
At higher datarates more accurate custom matrices help. (search the forum)
I know about matrices, I just wanted to say that IMHO Xvid is well tuned for low and mid bitrates, while is not for high bitrates. Maybe it's the MPEG4 standard that have this troubles, who knows?

I tried MPEG matrix, but sometimes it's too blocky in my tests. I tried other matrices but, to tell the truth, I couldn't notice what effectly was the best one for that source (for example H263 or HVS good).
As I repeat, it's easy to see the difference between 800 and 1000 kbs, it's very very difficult to see it between high bitrates.

[Zarxrax]

I just wanted to chime in here and mention something. I was recently trying to encode an anime, and even at quant 2 it would get blocky. I tried encoding it in Divx 5 and it turned out great though. I examined the divx5 file and saw that divx 5 was using quantizer 1.
I think xvid could really be improved if quantizer 1 encoding was optimized somehow. It currently produces output thats way larger than quant 2. There needs to be some sort of middle ground between these two, or something. Quantizer 1 is never used because its so huge, although quant 2 doesnt always give best quality. If there was something between 1 and 2, that didn't make it too much larger than quant 2 (maybe 50% larger at best) and still provided the visual quality of 1, then I think it would be a great improvement for xvid.

[Selur]

@trodas: "HVS Best matrix" is just a name of a spicific mpeg custom matrice floating around in the board. ;)

@Zarxrax: ever tried mfoon (http://forum.doom9.org/showthread.php?s=&threadid=52066) ?
Instead of using quant 1 you could use a custom matrix,...

@symonjfox: "it's easy to see the difference between 800 and 1000 kbs, it's very very difficult to see it between high bitrates." depends on the source,...

Cu Selur

Ps.: btw is the topic-starter 'dihelson' still reading this ? *gig*

[Zarxrax]

@Selur: yea I've tried mftoon, but I didn't need it for this. I had a very very nice quality dvd.
I thought about using a custom matrix, but I know I read somewhere on here before that custom matrices don't comform to mpeg-4 standard, and I don't want to create a file that might not be supported in the future.

[Selur]

@Zarxrax: Afaik there's no problem with custom matrices and the mpeg-4 standard, they should be standard conform. (maybe you mixed up custom matrix and modulated quantizer?)

MPEG matrix (either default or not default) is a part of mpeg-4 advanced simple profile. It should be decoded by every decoder which decodes this profile.
by syskin (http://forum.doom9.org/showthread.php?s=&threadid=50673)

Cu Selur

[dihelson]

See, this topic is becoming very useful and interesting.

The reason why I posted this simple question was exactly what some colleagues talked about: CAPTURES.
I recently put my hands on some nice, RARE LD (Laserdiscs) of shows, Jazz, etc, and I'd like to capture with some nice quality. I often use HUFFYUV for this. The problem is.
What's next ??
I don't know how to reduce this tremendous big files (some 50Gb) to a thing that could FIT on a DVD, on a DVD quality basis, so I thought about pushing XVID to a DVD quality independently from the size.
Perhaps if I could convert the source material to MPEG2 would be nice (I can't find an MPEG2 Codec, do you know one ??) (could someone explain better ??)
But, I will try some hints told here in order to see how these improved.
Perhaps these hints could for instance solve the problem I have sometimes: When I convert some kind od DVD to Xvid for optimal quality, say, 3 CDs, the file is less than 1200Mb, and the quality is not so good, could be better, since I'm pushing for it.
Thanks for the answers, and please, continue to discuss this. I'm learning a bit by reading all of you.

Dihelson Mendonca

[Selur]

Instead of Huffyuv I would take one of the lossless codecs that supports Yv12.

Avi=>Mpeg works fine with TmpgEnc and with the doom9 guides it should be no problem to da an avi to dvd conversion. ;)

"and please, continue to discuss this"
we probably will ;)

Cu Selur

[mikeson]

@Zarxrax:
I was recently trying to encode an anime, and even at quant 2 it would get blocky.
And what matrix did you use? AFAIK DivX5 uses only H.263 (less blocks, less details), in XviD you can choose or create your own one.
I think there must be something else, because at quant 2 (are we talking about 1-pass quant 2 mode here, aren't we?) picture in XviD almost never gets blocky. Did you use Qpel, B-frames, VHQ, Chroma motion? What is your AviSynth script?

BTW quant 1 is overkill IMHO.

[Zarxrax]

@Selur: Thanks, I'll be sure to try custom matrices in the future then. I must have just misread something at some point.

@mikeson: I used the H.263 matrix. The aviysynth script just contained deen and temporal cleaner, aside from the standard stuff like decomb. Even with no advanced settings on xvid, the image came out the same. The difference between the xvid and divx encodes was simply that divx wasn't afraid to use quant 1 (it's even enabled by default). I still really think the xvid guys ought to look into this, because there's obviously a reason divx would allow quant 1 support if it was never needed, as everyone claims.

[Selur]

Another thing about quantizer numbers in Xvid and DivX:

If you don't use the same matrix you can't compare the quantizer numbers.

e.g. quantizer 2 with mpeg is something like six or seven (if I remember correctly) with 'my' matrice

Since in general the h.263/mpeg quantizer 1 used in Xivd turns out to be 'unpredictable' large it doesn't seem to be a bad idea to disable it for 'inexperienced' users. (If one needs more 'exact' quantizers change the quantizer matrix.)

Cu Selur

[mikeson]

@Zarxrax: Please remeber to hit Load defaults button in XviD config between tests when using custom matrices.

BTW it would be nice if you'd provide us some samples...

[Soulhunter]

@mf
Import("SharpTools-v0.3.avs")
SharpResize(1280, 544)

Yap ! You script is simply WOOOW ! ;)

The Matrix with res. of 1280x512 @4300kbps + AC3 = WOOOW !!!

But let my PC 2days+ running is maybe a bit tooooo much... :D :D :D

@symonjfox
I think that XviD can be improved on high bitrates. We all use it for 1-2 CD rips, so we pratically test it everyday on the same settings.

I think XviD is already tuned for high bitrates...

Ive done some small tests !

Heres a quote from my test...
Small DivX/XviD test...

DivX 5.05 -VS- XviD Koepis-24.06.03

Gladiator chapter (about 00:06:30)

720x576 pix. / 25fps

2000kbps and 3000kbps encode

DivX / Orig. 2Pass / Default
XviD / MPEG Matrix / Default

Result:

The 2000kbps DivX file was less blocky than the XviD file, but also a bit less detailed...

I would say at 2000kbps XviD and DivX are same good... Depends on personal favor !!!

DivX = less blocks / less details
XviD = more blocks / more details


But at 3000kbps the things changed...

Because at 3000kbps BOTH files have no "real visible" blocks, but XviD looks still more detailed...

DivX = no blocks / less details
XviD = no blocks / more details

So, I think this shows that XviD's pros really begin to start at higher bitartes...
And this is even more true for Q1 or Q2 encodes...

Heres a thread (http://forum.doom9.org/showthread.php?s=&threadid=59794) where Ive posted about XviD Q2 encodes...

IMHO, Q1 is for sure overkill... :D

Bye

[Zarxrax]

I don't think people are understanding what im saying.
I encoded on both xvid and divx using the same quantizer matrix, H.263.
BOTH xvid and divx produced blocky output if i fixed them to quantizer 2. However both produced perfect output at quantizer 1. The difference between them is, divx allows quantizer 1 by default, while xvid people just say "no, you will never need quantizer 1, because you can't tell the difference".

BTW, im not trying to bash xvid or anything, its the only codec I ever encode to. I'm just saying I seriously think there is possiblity for improving the maximum quality by looking into quantizer 1 some more.

I would be glad to post sample, but I don't have the dvd here with me at the moment, so thats not possible. I could have it again in a week or 2 though...

[symonjfox]

Originally posted by Soulhunter
[BI think XviD is already tuned for high bitrates...
Ive done some small tests !

Heres a quote from my test...[/B] Maybe, it's a while I don't use Xvid. To tell the truth I bought a DVD-RW so I switch to MPEG2 and I'm satisfied.

I told you this because in my past tests (this summer) I tried to Xvid encode some DVB stuff, but also with MPEG quantization and Q2 is still somewhere blocky or not good as I expected (Q2 MPEG means that Xvid output is bigger than original M2V file and has lower quality). DVB stuff is not like DVD quality, I know, but I expected more (as I saw that Xvid @ mid bitrates and normal resolutions is GREAT).

Maybe should be nice if someone could test UNCOMPRESSED studio material and report us how did xvid work?

[Tommy Carrot]

Maybe it's time to try rududu codec.:) With quality=0 and node cost=1, the image is really artifact-free (no blocks, no edge-noise, etc.). The image is a bit softer than the source but still excellent. Not to mention it gives lower bitrate than xvid at Q2.

[SeeMoreDigital]

If your source is an PAL DVD, personally I would not even crop the mattes away!

If you wan't to have a bit of fun. Try encoding your PAL DVD's to either DivX or XviD at 1024x576 pixels, a true 16:9 frame size. Instead of 720x576 pixels, an anamorphic frame size.

I've found this technique works very well with low bitrate encodes. There's no reason why it should not work for 'very' high bitrate encodes too!

Cheers

[mf]

Originally posted by SeeMoreDigital
If you wan't to have a bit of fun. Try encoding your PAL DVD's to either DivX or XviD at 1024x576 pixels, a true 16:9 frame size. Instead of 720x576 pixels, an anamorphic frame size.
Do you ever use SharpResize for that? :D

[Joe Fenton]

Very interesting thread. It also serves a purpose. I know people with film scanners. Rich people often have their own theater at home with a collection of 35mm films from Hollywood. Suppose they wanted to view their legally purchased films at a super-hi quality on their computer with $50,000 plasma large-screen display. They'd take the film to someone and say "convert this to something high-q I can play on my PC." I know someone in L.A. who must deal with requests like this all the time. Rich people want better than what the "have-nots" have (DVD) or there's little point in being rich. So you use your film scanner and a spanking new blu-ray drive from Japan and make an encode that is several times the size of a DVD, and hopefully several times the quality.

Now given a situation as described, what settings in xvid accomplish our goal? I've never played around with anything more than 3000 to 4000 1-pass with the goal of re-recoding it later at 1000 2-pass.

[SeeMoreDigital]

Originally posted by mf
Do you ever use SharpResize for that? :D I've tried all the DivX settings when encoding at low bitrates. However 'sharp resize' does appear to look the best!

As I said I've only ever tried doing this myself using low bitrates (below 800kbps). So I'll have to give it a go at above say, 4000kbps!

I was intending to make further tests some time ago, however I did not bother after I discovered that my Xcard could not take the extra encoded pixels... shame, but the PC likes it just great!

Strange that this method only works 'perfectly' when encoding from PAL DVD or DV.avi sources. Oh well!

Cheers

[mf]

Originally posted by SeeMoreDigital
I've tried all the DivX settings when encoding at low bitrates. However 'sharp resize' does appear to look the best!
I meant AVISynth (http://www.avisynth.org) SharpResize() (http://mf.creations.nl/avs/functions/SharpTools-v0.3.avs).

[SeeMoreDigital]

Originally posted by mf
I meant AVISynth (http://www.avisynth.org) SharpResize() (http://mf.creations.nl/avs/functions/SharpTools-v0.3.avs). Thanks mf.

I have to admit I have not used scripts for quite a while to generate my encodes. I'm far too lazy I know!

However, I think I should give AVISynth a go again. Are you able to write a script that could take an PAL DVD 720x576 source and resize it to 1024x576?

Cheers

[Soulhunter]

@SeeMoreDigital
Are you able to write a script that could take an PAL DVD 720x576 source and resize it to 1024x576?
Uhm...???
SharpResize(1024, 576) :D

@mf
Think there are some other possibilities that would be cool to add this script...

Maybe additional low intense "Unfilter()" ???

Or "AddGrain()" ! Cause processing before downsizing would make the noise get a bit interpolated with the source...
Think this looks more natural than adding it afterwards... !?!

Bye

[SeeMoreDigital]

Ok, ok. I asked for that!

Told you I had not used scripts for a while!

Cheers

[gotaserena]

This thread is very similar to one posted by the same user a few months back, so I will repeat myself (with a twist).

XviD is not optimised for high-bitrates. It's simple as that. It was created to do DVD -> CD(s) back-up so it is not the right tool for the job. If you want you can do either in other to convert captures:

1) Do like myself and use a MPEG-2 capture card. With editing software you will be able to burn that DVD in about 30min after pressing "stop" in the VCR/Laserdisc. There's no sense resizing VHS or LDs above 720 anyway. After a few hours of research into MPEG-2's bitrates, GOPs and the like you will be able to optimise quality. Alternatives: MPEG-4 capture device (can't vouch on them), or Huffyuv + CCE (or TMPEnc if you don't have that kind of money ;)). The latter one allows you more flexibility in the end result but it is also a steep learning curve (AVISynth as well).

2) Go ahead and use XviD's amazing flexibility to work out a solution for yourself. This was actually what I did for Formula-1 races captures. I tried to squeeze 4 hours of video into a DVD, and wanted full D1 resolution (NTSC), and had a pretty rough time optmising the setup. At high bitrates (or high quality or low quantisers) you want to be as conservative with the codec as possible: MPEG or H.263 quantization (custom matrices usually make use of the fact that you don't notice every little detail in every frame to smear out high spatial frequencies. The effect is quite noticeable at high bitrates, though) No B-frames, No global motion compensation, No quaterpel, No lumi masking, No Chroma Optimiser, VHQ mode turned off or in "1". If you use ffdshow for decoding you will also want to watch for post-processing options not to "overdo" corrections. Even then you may need more than 2500 kbps to actually render a 640x480 digital capture block-free/mosquito-free. Expect even more for analogic captures -- and please use that S-Cable. Processing times will sky rocket as you try and find the best setups for VagueDenoiser(), Convolution3D() or Deen() and occasionally CNR2().

And even 2) have its caveats... No matter how hard you try at KernelDeint(), TomsMoComp(), SmoothDeinterlace() or even your preferred bob AVISynth plug-in, you will face the fact that you will lose quality and that XviD does not work well with interlaced sources. The best way to keep quality is of course not to de-interlace the source, in which case you will be better off with MPEG-2.

There, I said it once more...

[Selur]

custom matrices usually make use of the fact that you don't notice every little detail in every frame to smear out high spatial frequencies. The effect is quite noticeable at high bitrates, though
This really depends on the custom matrix you use, you could also write one that preserve far more details than the standard mpeg/h.263 matrix.

Cu Selur

[gotaserena]

I guess I was sloppy -- in the name of concisiveness. Yes, what you said is true, in fact you can do a lot with quantization matrices. But, as usual, there is a trade-off: at higher frequencies, larger values mean less ringing and more compressibility but also less detail, and vice-versa. Of course you can try for yourself and reduce all the numbers to keep the detail, but it will also reduce compressibility. So you compromise -- increase the numbers at higher frequencies carefully to keep detail -- and arrive at a solution, the MPEG and H.263 cases being the standard. If you work a more intelligent solution (like andreas78's or HVSs), you make use of the fact that you quoted. The message is: either you choose to increase the numbers in the quantization matrix or to increase the bitrate.

If you choose to increase the bitrate, you will quickly arrive to a point where XviD (or any MPEG-4 for that matter) makes little sense over, say, MPEG-2. My experience is that before long you are running to DVD-like bitrates when dealing with captures, which was the situation where MPEG-2 was optmised for.

[Selur]

You are right, that sometimes mpeg2 might be better suited and that smaller numbers in the quant matrices mean more details ans so less compressibility, but isn't that what the topic was about?

How can I unleash Xvid quality for 7 CDs?

For me this sounds like (at least the topic starter) was/is searching for a way to produce Xvid files with better quality, not careing so much about size. :)

Cu Selur

[gotaserena]

Well, my point is that isn't really practical to ask this of XviD. It was, after all, designed with pretty restrictive size limits. It sounds like asking "can I move my books to another city with a Smart (http://www.missyeartly.com/id199.htm)? -- Sure, but a van will do it faster and easier". Of course one can still claim: "right, but I want to use a Smart (http://www.missyeartly.com/id199.htm)". This does surely make for an interesting question, but I fear it will have little practical value.

[symonjfox]

Originally posted by gotaserena
And even 2) have its caveats... No matter how hard you try at KernelDeint(), TomsMoComp(), SmoothDeinterlace() or even your preferred bob AVISynth plug-in, you will face the fact that you will lose quality and that XviD does not work well with interlaced sources.
Good news, Syskin said that new Dev4 Xvid fixed some interlacing bugs.

I agree with you that keeping interlaced material gives higher quality than deinterlacing.

[SeeMoreDigital]

Blimey

OK gotaserena. I agree, XviD was not specificly designed for use at high bitrates. However it is based on Mpeg4 which can be optimised for high bitrates. If you take a look at the DivX or Sorenson version of the Mpeg4 codec it will allow you to encode at around 9000kbps or more!

I dont agree with your view that generating encodes above 720 is nonsensical idea. And who said anything about resizing VHS or LaserDisc sources anyway?

But staying on the subject for a while, if you've got a treasured VHS tape or LaserDisc that you wan't to digitize, it makes perfect sense to me to capture your digitized first generation encode at the highest resolution and bitrate your equipment will allow.

So on that basis, I would highly recommend to anyone, wanting to create such a first generation encode, to capture any true 4:3 source at 1440x1080 and any true 16:9 source at 1920x1080. And stay far away from anamorphic frame sizes!

Better to downsize/resize your second generation copy

And don't forget that the 1920x1080 image pixel frame size has been adopted as a high definition standard across many parts of the world now. Which is an industry first in television broadcasting!

Oh, and if you are serious about capturing from VHS tape and/or LaserDisc sources, don't forget to remove or disable your equipments RF modulator first. It reduces quite a bit of electro-magnetic noise!

Capturing direct to high bitrate Mpeg2 produces great results. I know this, because I have an ADS USB 2.0 capture device and it's great. However, I find working with high bitrate Mpeg1 much easier - I'm expecting that remark to spark more debate!!!

Cheers

[Mug Funky]

hmm. something that could be added to sharpresize might be a tomsmocomp based thing.

i've been using "turnleft().tomsmocomp(1,-1,0).turnright().tomsmocomp(1,-1,0)" to upsize old VCD encodes to a nice DVD framesize with some success. and it's faster... all it does is double the size, and there's a bit of a problem with a green line/dot in the bottom left though.

maybe tom could make a simple resizer based on tomsmocomp's -1 mode?

[trodas]

SeeMoreDigital - at resolutions about 1440x1080 and 1920x1080 is proven that no codec can match WM9 M$ one, so, stay away from Xvid for such big qualities ;)

Xvid looks like being optimal for bitrates about 8000-1600 (typical DVD rip), and performs worse for all other extreme situation. Like you can see on very high bitrates or, into another tread there, where i complaining about the fact, that Xvid codec get saturated very fast into 240x180 15fps 3min movie, so i can't get more decent quality from it, even offering very big bitrate (10 000) and more size - both pases are at bit exactly same big and Xvid seems definitively worser that orginal Sorenson3 encode.
Ofcourse things might be better for Xvid, if it can work with orginal material, but even there i fear that developers have to do something with it... :rolleyes:
The quality drop is too big, and the saturation is a problem there.

And same for the high bitrates. So, for this, use mpeg2 and loooots of passes, to get most of what you can, when come to resolutions bellow 1000 pixels wide.

Over this, use WM9.

Nuff said :D

[Tommy Carrot]

Originally posted by trodas
SeeMoreDigital - at resolutions about 1440x1080 and 1920x1080 is proven that no codec can match WM9 M$ one, so, stay away from Xvid for such big qualities ;)



Sorry, it's not proven. The common mistake is using h.263 quantizer for high quality encoding, while mpeg quantizer is much more suited for that. Mpeg Q2 gives excellent result, the image degradation is hardly noticable, while h.263 quantizer gives obvious loss of detail, at larger bitrate usually.

[trodas]

Well, not? I never saw anything nicer that WM9, when come to BIG resolutions ;)

Xvid simply suxx, when you set very big bitrate and use big resolution... :rolleyes:

And i using Mpeg quant too :D

[temporance]

> Well, not? I never saw anything nicer that WM9, when come to BIG resolutions

I can't agree with this statement. In theory, so long as bitrate remains proportional to (width * height), each codec's performance should be unchanged by resolutions. My experience backs up this theory.

There are two reasons why you might consider WMV9 better at big resolutions:

1. WMV9 has better decoding performance and rendering abilities. Remember, M$ has access to the best programmers and can make sure this is done well. This is nothing to do with the compression algorithm used so xvid can catch up.

2. WMV9 tends to cope better with "sparse" images - and a lot of HD content is like this (probabaly will improve as HD eqipment gets better)

[mikeson]

@trodas:
Xvid simply suxx, when you set very big bitrate and use big resolution...
You shouldn't talk like this, especially when you didn't show any proves. Maybe it is only because you are not able to set AviSynth/XviD right.

[mf]

Originally posted by mikeson
@trodas:
Originally posted by trodas
Well, not? I never saw anything nicer that WM9, when come to BIG resolutions ;)

Xvid simply suxx, when you set very big bitrate and use big resolution... :rolleyes:

And i using Mpeg quant too :D
You shouldn't talk like this, especially when you didn't show any proves. Maybe it is only because you are not able to set AviSynth/XviD right.
[2493] sux: A fairly modern word meaning approximately "I have an exaggerated opinion about something I don't really know anything about"

:D Gotta love DDB.

[chilledoutuk]

I am currently converting a Superbit PAL dvd to divx with a res of 1024 x 416 (would be 576 with black bars)

The quality of these superbit DVD's is impresive.

I highly recomend them

I was especially impressed with divx's directshow filter's effciency when playing back a 1024x416 sample it was only using 20% cpu.

Ill post how the encode comes out tommorow and maybe a sample clip.

Oh does anyone know how these hi res MPEG4's playback on hardware standalone players?

laters

[Blue_MiSfit]

what the hell is a superbit DVD anyway??

[mikeson]

@Blue_MiSfit:

Search instead of cursing. Read here (http://forum.doom9.org/showthread.php?s=&threadid=53729&highlight=superbit+dvd).

[chilledoutuk]

hey just tested a 1024x416 xvid i created in vdubmod on the xbox media player and it played fine without any dropped frames which is cool.