[HD 1080p24 Challenge] MPEG2, VC-1 and H264 with real uncompressed source

[benwaggoner]

Is that 10 bits per channel?
10-bit luma. v210.

[Golgot13]

Where do I even begin? Microsoft can't forbid *everybody* to talk about something, especially about VC-1 which isn't even owned by Microsoft. Microsoft can sign NDAs (Non-Disclosure Agreements) with companies it licenses its software to, which is a fairly standard practice in the software world when pre-release or beta software is in question.

Yes, I know it's so why I stopped or removed part on all my thread about VC1 (see my post at January 2007).
But I would to show only the quality, because nobody (independant person or company)
could compare truly VC1 at other codec. And my company received a release version
not a pre-release or beta-version.

And I think when a software (implementation of Codec) is compare at other it is better for it:
Sagitaire find some bug on PeP (which influence the quality and HDDVD compliant)....



BTW, the text you keep linking to will be changed before the end of the month, removing some of the more ambiguous references to MPEG-2 and H.264.

Nice, so you remove my prefer VC1 reference link because I like to use it :rolleyes: :eek:.


Though I'm sure you'll probably just keep linking to some archive page after that. :rolleyes:


No, I'm not like you I use it because it is on your server and all people can see it. :p :p

[45tripp]

I think it's eager anticipation for the grain source,
making people a bit twitchy.

With high enough compression, metrics become rather unreliable, yes?
6000k wasn't that much of a compression considering how compressible the source is.

The source with grain, should render metric tests quite useless at high compression.

Personally i'd like to see perceptual analysis by trained eyes.

ty

[Golgot13]

The source with grain, should render metric tests quite useless at high compression.

Agree


Personally i'd like to see perceptual analysis by trained eyes.


There is lot of developpment and research about it. It 's why there is on professional encoder HVS/ psychovisual option.
But the efficiant of H264 (codec or implementation) give a good advantage...

[MfA]

Personally i'd like to see perceptual analysis by trained eyes.
Expert viewers can know what codec they are looking at, you run a risk of measuring average bias rather than quality when the differences get small.

[Manao]

And even worse than that, "trained eyes" don't see the same artifacts "common people" see. Or rather, they see all the artifacts, but weight them differently. So one has to ask whether their point of view is more pertinent than the point of view of the "average user".

[benwaggoner]

I think it's eager anticipation for the grain source, making people a bit twitchy.
Yeah, fully understood. We'll be using the same clip internally, so I'm on pressure from all ends. But I've been given some other higher priority projects I've got to get done first (which is why I'm working on a gorgeous Saturday afternoon).

[45tripp]

Expert viewers can know what codec they are looking at, you run a risk of measuring average bias rather than quality when the differences get small.
Impressive.
Still, when is anything ever truly unbiased?

And even worse than that, "trained eyes" don't see the same artifacts "common people" see.
I'd find it interesting.


Or rather, they see all the artifacts, but weight them differently. So one has to ask whether their point of view is more pertinent than the point of view of the "average user".
That could be considered educational.
Helping the average user better spot and understand various artifacts.

And why not align your preferences and biases to those of an advanced user?
I don't see how it can be worse than, or even construed different to, aliging to the bias of a mathematical model.

ty

[Golgot13]

@Zambelli

- Nice MS update this link about superior quality of VC1 (compare at H264 and MPEG2):
http://www.microsoft.com/windows/windowsmedia/howto/articles/vc1techoverview.aspx#VC1ComparedtoOtherCodecs

But I'm not agree with this:
" On all three measures (resolution, noise, and overall impression), the quality of the Microsoft VC-1 encoder was judged closest to the original D5 master. By comparison, the H.264 encoder that was tested rated as comparable only to MPEG-2 on two of the three measures (resolution and overall impression), and was rated somewhat worse than VC-1 on noise. "

- MS like to say something and don't give any graphs or informations about this affirmation.
- About H264 and noise, I can say that with my all PSNR graph show that H264 is better
than VC1 for video movie with noise (tests release with PeP 1.06).
- The difference is more than 7% PeP 1.07 make (compare with PeP 1.06).



@Ben
- We wait the video to show if H264 or VC1 is better for video movie with noise.
- I don't see lot of post from you on this forum but you answer quickly on DVDlist
when some professional want to use H264 codec (than VC1)...
- To my mind, H264 is better for HDDVD use and VC1 on BD (VC1 can be compare at H264 on HDDVD,
more space so more bitrate to have same PSNR)....


@Sharktooth
You may be say true about MS "dodge" for video because we wait three month soon (for the reign of VC1)...

[Sharktooth]

yup... have you ever had any doubt about it?

[Inventive Software]

Bear in mind that they're (Microsoft employees, Mr Waggoner in particular) trying to feed themselves and their families, not just your insatiable encoding habits. ;) It'll come when it comes. I anticipate it as much as anybody else, but I don't bang on about it. ;)

[Sharktooth]

@Inventive Software: it's just a matter of 5 minutes...

[zambelli]

But I'm not agree with this:
" On all three measures (resolution, noise, and overall impression), the quality of the Microsoft VC-1 encoder was judged closest to the original D5 master. By comparison, the H.264 encoder that was tested rated as comparable only to MPEG-2 on two of the three measures (resolution and overall impression), and was rated somewhat worse than VC-1 on noise. "

- MS like to say something and don't give any graphs or informations about this affirmation.
- About H264 and noise, I can say that with my all PSNR graph show that H264 is better
You might've missed this part of the text a paragraph earlier:

"For example, the DVD Forum conducted tests in the winter of 2002 to select codecs for the next-generation optical disc standard. Viewers from Hollywood film studios and major consumer electronics companies rated video clips on a scale of 1 to 5 for resolution, noise, and overall impression. Multiple codec implementations were tested, including MPEG-2, VC-1 (provided by Microsoft), H.264, and MPEG-4 Advanced Simple Profile. The baselines against which the codecs were compared were D5 masters and D-VHS (24 Mbps). During the tests, viewers were not told which codec was used to encode each of the clips."

The statements about VC-1 being better than H.264 specifically refer to the comparison in question: 2002's DVD Forum comparison. The article isn't making any generalizations, it's simply stating that Microsoft's VC-1 encoder was judged to look better than some H.264 encoder in a 2002 DVD Forum comparison. It's a simple statement of something that actually did occur.

There's a paragraph even earlier in the article that makes it very clear that every comparison of codecs is a comparison of codec implementations and is not useful for generalizing.

[Golgot13]

You're true about the state in 2002. I sure that VC1 (aka WMV9) was better than H264
(I was surprise by the quality of WMV on WMVHD title). But in 2002, H264 was in developpement
(not really finish) and today (since 2005) H264 is the best codec for new HD disc format.

I don't understand why you update your website to put old information ???
We are in 2007, and this codec challenge can show if this data information is available today.

[Inventive Software]

@Golgot: It's called "spin". ;)

[IgorC]

But in 2002, H264 was in developpement
(not really finish) and today (since 2005) H264 is the best codec for new HD disc format.
Since 2004.

I don't understand why you update your website to put old information ???

"We forget it for obvious reason". Marketing.

[Golgot13]

I forget

Bear in mind that they're (Microsoft employees, Mr Waggoner in particular) trying to feed themselves

I'm not agree, because if you read my old post you will know that MS use "forum"
to say at professional (on training) and magazine that his technology is better.
So I don't surprise to see some post of MS guys...


@IgorC
You don't test VC1 and H264 encoder in 2004 (?).
I think you talk about FRExt, but it was available at the end of 2004...

[IgorC]

You don't test VC1 and H264 encoder in 2004 (?).
I think you talk about FRExt, but it was available at the end of 2004...
I think I talk about Ateme H.264
There wasn't need to test VC1 in 2004 simply because it wasn't released.

[zambelli]

I think I talk about Ateme H.264
There wasn't need to test VC1 in 2004 simply because it wasn't released.
VC-1 didn't really exist as a finished standard until 2006. Most of the codec comparisons between "VC-1" and other codecs between 2002 and 2005 were done on WMV9, essentially. The DVD Forum comparison, for example, actually compared WMV9 Main Profile against H.264 and MPEG-2. All the WMV-HD movies were Main Profile encoded too.

"We forget it for obvious reason". Marketing.
I think you guys are nitpicking now. There's nothing in that article that actually says "VC-1 is better than H.264". There are examples given of comparisons in which VC-1 won or did well, but that's all they are - examples. They're not there to demonstrate general superiority, but to simply show VC-1 doesn't suck. :)

[Sharktooth]

those are examples of VERY OUTDATED comparisons expecially choosen to make ppl think VC-1 is good and maybe better than h.264...
the reality is a bit different though...

[Golgot13]

those are examples of VERY OUTDATED comparisons expecially choosen to make ppl think VC-1 is good and maybe better than h.264...
the reality is a bit different though...


Hi Sharktooth,

no way with MS, H264 is the best codec MS know it so we will not see any video to make a comparison
and all informations on MS website which talk about H264 are from H264 in 2002-2004.
Because today VC1 can not be compare with H264 (but with MPEG2 it can).

I'm very disappointed by MS (you know it Zambelli) about the true of VC1
and politic with some authoring studio (in europe, I'm not alone...).

Lot of studio use and will use H264 on new HD optical disc format.


I can show the quality result with last developpment of VC1 but it is impossible
the company, when I work, signed a NDA (and Ben don't say me if I can)....

[Golgot13]

Hi all,

we will never see something from Ben...
Amirm say that VC1 is the best only... :(

@Sharktooth
You said true, MS dodge the challenge because they can make
better than H264 (we waited one month to have a VC1 file from MS
and saw than 1.07 couldn't do better than H264).

I hope to see more HDDVD in H264 like Transformer (which used Toshiba encoder).

[zambelli]

What exactly are you waiting for, Golgot13? An HD source? Just any HD source? Does it need to be shot on film, or will HD cam do? How long does it need to be?

I'm not sure why this notion that encoder testing can't continue until a video source is provided. It's not like anybody needs Microsoft's permission to continue testing encoder quality.

[Golgot13]

What exactly are you waiting for, Golgot13?

The truth. You can say simply that H264 codec produce better quality than VC1 codec.
Ben promised video source with "grain" to proove than VC1 have best result than H264,
but I understand that we will never see it...
"Sharktooth" said true "MS dodge", because you can not produce a better quality than H264 encoder.

I tested the version 1.07 of your encoder and you can not compare it today with H264 encoder.
MS said at everybody that there are more HD title because VC1 is best codec but it's because PeP was/is free...

MS can not produce everytime the "must":
- HDi is good better than BDJ (because problem of implementation between of different BD hardware player)
- VC1 codec is not better than H264 (but on PeP there is some good tool like Xscaler).


All industry use H264: Divx buy Mainconcept (for H264); There are HD camcorder on H264;
some APN can record in H264; H264 is support by 3G mobile phone for videoconference;
HD channel broadcast in H264,....
For VC1, except HD disc (which support both H264 codec too) I saw only Tatari and Inlet's Fathom card
(Inlet card can encode in H264 too...).


I 'm really disappointed by MS behavior.

[honai]

I 'm really disappointed by MS behavior.

Why exactly do you think Microsoft would care what you and a few other posters at doom9.org think about VC-1? The point is that you didn't even pay anything for this, while Microsoft is a business. So why should they feel obliged to prove anything to you as a non-paying non-customer? I don't get it, to be honest.

[Razorholt]

The point is that you didn't even pay anything for this.

Maybe because they forgot to put a price on it? Bill is loosing his head, I'm telling you.

The Win OS is included within the computer's price, did you know that? And M$ codecs as well as their other freebies are just marketing tools for you to choose their platform - Nothing wrong with that tho.

I'm also not impressed in VC-1, sorry. Can I say that? Please?

[45tripp]

good edit!

a nice new accessible source will be nice for everyone to play around with,
regardless of codec 'battles'

[honai]

@Razorholt

Of course you are entitled to your own opinion. But you are not entitled to Microsoft actually caring about your opinion.

That is the difference.

[Inventive Software]

Can I just say that I think VC-1 handles interlaced encoding better than MPEG-4 ASP and AVC? I'm attempting to reduce bitrates and make the encodes appear nicer, after recording from my DVD recorder. It records interlaced, and so far, encoding it to VC-1 beats the s**t out of H.264 and MPEG-4 ASP. Except for the time it takes. Tis why I'll leave it going through the night. ;)

[Shinigami-Sama]

Dear lord
I just read this whole thread(2 hours!)
and the last hour of its was zamballie defending himself again
granted he walks into it most of the time

but can we get back on topic?
I was rather enjoying this thread

I will ask around my graphics artist buddies for any good hi-def sauces though

[Golgot13]

Why exactly do you think Microsoft would care what you and a few other posters at doom9.org think about VC-1? The point is that you didn't even pay anything for this, while Microsoft is a business. So why should they feel obliged to prove anything to you as a non-paying non-customer? I don't get it, to be honest.


It was not true, MS gave the encoder at few authoring studio companies last year
(thanks for no competition, only big studio receive it no standard/small companies...).
We receive it many months after because there is not enough a real market for HDDVD...
About money, when VC1 is used on HD disc for each disc there is some royalties
which is received by owners of the format (like MPEG2, H264, AC3,...).

[Beastie Boy]

And your point is what exactly?

You seem to have hijacked this thread just to complain about MS because you feel left out by them.

[Golgot13]

And your point is what exactly?

You seem to have hijacked this thread just to complain about MS because you feel left out by them.


No, I want that thread finish and show what codec is the best.
I made many test on movie source (with and without grain) but I can not give file
(H264 and VC1 stream are signed). The company, where I work, signed a NDA with MS for VC1 encoder.
I can only say that H264 win with my all source (I have sometime some problem
with HDDVD player with H264 stream). But I prefer that it proove on this thread
(I waited the source promised by Ben)

[honai]

I realize that you feel strongly about this, and that it's kind of a personal crusade for you to prove it. I have the utmost respect for you perseverance. And I'm sorry to say this, but even if you proved that, it would be absolutely meaningless, pretty much like a bag of rice toppling down somewhere in China.

Nobody cares.

[Golgot13]

I have the utmost respect for you perseverance.

To my mind, it's my stupid stubbornness

And I'm sorry to say this, but even if you proved that, it would be absolutely meaningless, pretty much like a bag of rice toppling down somewhere in China.


But after some people will know and use it (H264 codec :p)


Nobody cares.

No there is many professionals here (MS made some advertisement
for this forum and AVS Forum about feedback that HDDVD is good).

So at all authoring studio, HDDVD is really the best support (today), HDi is good
level on interactivity (good job of MS) and H264 can give the best HD video result.

[honai]

You assume that the industry is headed by CEOs and CTOs who think logical and reasonable. They don't. ;)

The point is that there are many factors in the decision making process which codec to chose for HD discs, and doom9.org is not one of them.

[Golgot13]

and doom9.org is not one of them.

Yes but compressionnist can open their eyes with this thread.
Here we are purist because many professional compresionnist who made/make
HDDVD and BD (more on BD disc ?) use the default preset for all movie,
same preset for all movie they only change the bitrate !!!
Except all movie encoded by MS team: really good quality I bought all HDDVD from them

[Amdh]

So, As Expected, x264 Beats All The Other Formats, That's Perfect .. I'm Using It Since 2 Years And It's Just Excellent, It's True That MPEG2 Is Still Faster And Receives Much More Support From Hardware Producers But x264 Is Still My Choice. Thanks For The Extreamly Intresting Results.

Hey By The Way, I've Not Even Tried VC-1 Coz I Know That The Quality It Gives Won't Compare With H264

[Shinigami-Sama]

You assume that the industry is headed by CEOs and CTOs who think logical and reasonable. They don't. ;)

The point is that there are many factors in the decision making process which codec to chose for HD discs, and doom9.org is not one of them.

then who ever is making the decisions needs to give their heads a shake. Theres so much data holed up here in doom9 it makes some universities cry
especially with the codec shoot outs