New Ken Kutaragi interview (1) @ PC Watch

scooby_dooby said:
Found it! My mistake, it was a h.264 comparison, not vc1.

The original master = 4.03 :)lol:)
24mbps MPEG2 = 3.59
20mbps H.264 = 3.9
16mbps H.264 = 4.0
12mbps = 3.71
8mbps = 3.65
http://ftp3.itu.ch/av-arch/jvt-site/2004_07_Redmond/JVT-L033.doc

This does point to 16-20mbps being the limit as far as perception goes, 16 actually scored a little bit higher than 20.
That's the one. It doesn't cover higher bitrates though as I thought, so maybe I was getting confused with another paper? Anyhow, when the difference between original source and compressed at 20 Mbps is so small, it begs the question why waste space on improvements that can't be detected. Though there is a caveat - we don't know the type of media being shown. Dark footage exhibits far worse artefacts than typical scenes, for example, and if the test cases didn't cover such situations, the areas where most improvements would be made weren't being explored.
 
scooby_dooby said:
It seems to be just an example of stupid company politics coming in the way of common sense. It's like Sony has invested so much time and money into MPEG2 propaganda that it behooves them to use MPEG2, even when it's clearly an ass-backwards solution to the problem, and it clearly is completely unsuitable for 25gb discs.

I like the extra space BR provides, but if they are going to completely waste it by using 10 year old, outdated codecs, what's the point?

Yes. I believe VC-1 and AVC are better than MPEG 2.

However if at higher bitrate, they perform similar to MPEG 2 and there are adequate space for it, then I can imagine business folks pushing for the later. They need to earn enough to continue to invest more in this market.

Also I think Sony's intention is not to force everyone to use MPEG 2. They just want to make sure they can still get revenue stream for it in the next gen platform. Naturally, the final choice lies with the studio and consumers. Regarding the claims that MPEG 2 can outperform AVC and VC-1, the proof will be in the pudding. I have no preference either way except for the final quality.
 
patsu said:
Yes. I believe VC-1 and AVC are better than MPEG 2.

However if at higher bitrate, they perform similar to MPEG 2 and there are adequate space for it, then I can imagine business folks pushing for the later. They need to earn enough to continue to invest more in this market.

Also I think Sony's intention is not to force everyone to use MPEG 2. They just want to make sure they can still get revenue stream for it in the next gen platform. Naturally, the final choice lies with the studio and consumers. Regarding the claims that MPEG 2 can outperform AVC and VC-1, the proof will be in the pudding. I have no preference either way except for the final quality.

That makes sense to me (it's all about the $$), but it's the initial 25gb releases using mpg2 that really have me scratching my head, clearly it's not adequate right now and shouldn't be in use.

They've created a funny situation where people buying BR movies are not getting real HD movies. Are they going to re-release the same movie's in the future? What abot the early adopters? Do they re-buy the same BR movie to get the quality they should have recieved initially? Talk about confusing customers.
 
scooby_dooby said:
That makes sense to me (it's all about the $$), but it's the initial 25gb releases using mpg2 that really have me scratching my head, clearly it's not adequate right now and shouldn't be in use.

They've created a funny situation where people buying BR movies are not getting real HD movies. Are they going to re-release the same movie's in the future? What abot the early adopters? Do they re-buy the same BR movie to get the quality they should have recieved initially? Talk about confusing customers.

Scooby the problem is h.264 is not ready yet. But it will in the very very very short future.
 
mckmas8808 said:
Scooby the problem is h.264 is not ready yet. But it will in the very very very short future.

Not ready yet!?!? That study I linked is from 2004!

Maybe they got caught with their pants down, ass-u-ming that DL-BR disc would be ready by launch, and now they have to scramble...still they should've have a much better 'Plan B' then this nonsense.
 
scooby_dooby said:
Not ready yet!?!? That study I linked is from 2004!

Maybe they got caught with their pants down, ass-u-ming that DL-BR disc would be ready by launch, and now they have to scramble...still they should've have a much better 'Plan B' then this nonsense.

It is a standard produced by a committee with representation from a broad cross section of the movie industry, that's why it takes a long time to agree on everything. The benefit is that it is vendor neutral which will mean it will be adopted across the board when it does come out, and it won't chop and change in order to force an upgrade or to give one vendor an advantage over the others like Microsoft's Office or Windows Media player or Realplayer or Apple Quicktime formats. That has to be a big bonus.
 
patsu said:
Yes. I believe VC-1 and AVC are better than MPEG 2.

However if at higher bitrate, they perform similar to MPEG 2 and there are adequate space for it, then I can imagine business folks pushing for the later. They need to earn enough to continue to invest more in this market.

Also I think Sony's intention is not to force everyone to use MPEG 2. They just want to make sure they can still get revenue stream for it in the next gen platform. Naturally, the final choice lies with the studio and consumers. Regarding the claims that MPEG 2 can outperform AVC and VC-1, the proof will be in the pudding. I have no preference either way except for the final quality.

Using a higher rate Mpeg2 to provide the same quality as a 2x + lower rate VC1 negates the space advantage completely (that's to say that we'll see BD DL media anytime soon). VC1 will only continue to mature and get better over times meaning that it can run at even lower rates and still be transparent.

As for AVC, the HD DVD demo disc in encoded in Mpeg4/AVC but the VC1 encoded movies look better. Also the VC1 encoded movies look better than their AVC counterparts released in Japan.
 
SPM said:
It is a standard produced by a committee with representation from a broad cross section of the movie industry, that's why it takes a long time to agree on everything. The benefit is that it is vendor neutral which will mean it will be adopted across the board when it does come out, and it won't chop and change in order to force an upgrade or to give one vendor an advantage over the others like Microsoft's Office or Windows Media player or Realplayer or Apple Quicktime formats. That has to be a big bonus.

VC1 is also being pushed onto Mpeg-la for licensing/royalties meaning that the fear of MS changing it around freely goes out the window.
 
SPM said:
It is a standard produced by a committee with representation from a broad cross section of the movie industry, that's why it takes a long time to agree on everything. The benefit is that it is vendor neutral which will mean it will be adopted across the board when it does come out, and it won't chop and change in order to force an upgrade or to give one vendor an advantage over the others like Microsoft's Office or Windows Media player or Realplayer or Apple Quicktime formats. That has to be a big bonus.
This makes no sense. Once the standards are set, as they are now, then any PC capable of playing Blu-ray or HD DVD will be able to play the encoded material, whether the encoder originated from Microsoft or not. Otherwise, the hardware based decoders would require firmware upgrades when new movies came out. In my opinion, given the strong push from Microsoft for VC-1, it is more likely to become the standard just because they are being so helpful.
 
RobertR1 said:
Using a higher rate Mpeg2 to provide the same quality as a 2x + lower rate VC1 negates the space advantage completely (that's to say that we'll see BD DL media anytime soon). VC1 will only continue to mature and get better over times meaning that it can run at even lower rates and still be transparent.

As for AVC, the HD DVD demo disc in encoded in Mpeg4/AVC but the VC1 encoded movies look better. Also the VC1 encoded movies look better than their AVC counterparts released in Japan.

Taking the LOTR as an example, at 15 mbps we use up all 30 Gb without extra material. With BD DL, *assuming* BDA continues to push it, we can have 16-20 mbps (based on scooby's numbers) and still have extra material. Although this is only 1 (extreme) example, it shows me that BD DL can also be useful for VC-1 and AVC. The pristine quality and the convenience to avoid disk switching will be important should online movie distribution becomes prevalent later on.

As for the argument of wasted space.. for most other movies, doing it with MPEG2 using up all 50Gb vs doing it with VC-1 using only 30Gb makes no difference to me.

The MPEG2 proponents highlighted that less "compression" treatment will produce results more faithful to the original content. I was told VC-1 and AVC can adhere to the very same principle by turning off the more advanced techniques. In this sense it is scalable. However if MPEG 2 is more mature/optimized at this moment *and* we have sufficient space to toy with, it may not be a bad idea to use MPEG 2 for now. In my view, what people seem to disagree on is *when* to use (or even cut over to) VC-1 and AVC. This is an individual studio, per-title decision.

OTOH the business folks want to make money along the way to justify for their continual existence and further investment. Naturally both Sony and MS are guilty (if I can use such a word in this context) of it even though they use different reasons to justify their actions.

Personally, the only thing for me to do is to see the movies for myself at the end of the year or early next year. That's why I'm not tracking "Blu-ray vs HD-DVD" actively. I'd leave the major decisions to the good people in the industry. Afterall, they will need to fight another war against low-quality, online movie distribution after this one.
 
Sis said:
This makes no sense. Once the standards are set, as they are now, then any PC capable of playing Blu-ray or HD DVD will be able to play the encoded material, whether the encoder originated from Microsoft or not. Otherwise, the hardware based decoders would require firmware upgrades when new movies came out. In my opinion, given the strong push from Microsoft for VC-1, it is more likely to become the standard just because they are being so helpful.

MS is indeed doing a wonderful job pushing VC-1. I seldom hear about QuickTime and H.264 in the next gen DVD race (perhaps Apple Computer is betting the entire farm on the online distribution space, as opposed to the hi-def race).

I have many questions on VC-1.

What exactly did MS do to qualify VC-1 as a standard ? What standards body owns the VC-1 specification today ? Who owns the VC-1 brand now ? These are business assets that can be used against MS's competitors later (e.g., obsoleting VC-1, preventing contribution to VC-1 by other vendors, deploying and marketing a "VC-2" ahead of others, etc.).

The MPEG-LA body is only an administrative, negotiation and collection entity right ? Does it hold any power w.r.t. the evolution of VC-1 ?
 
thenefariousone said:
I think the homebrew community is a very vocal but very small group. Sony has no real interest in this group. They are only doing a few things to give the illusion of being on the side of the mod-scene.

The mod-scene may be a small group. But given the right climate the mod-scene can do a lot FOR you, for FREE. (You just need to look at valve or bethesda for confirmation of that). If a game dev truly could release a mod SDK for their game on PS3 that sufficiently insulates the user from the intricacies of programming for cell that could be a huge win.
 
soylent said:
The mod-scene may be a small group. But given the right climate the mod-scene can do a lot FOR you, for FREE.
Indeed. In the early days it was the free demo scene that pushed the Amiga's profile. The enthusiam you get from hardcore homebrew, an exotic architecture to attract them, and a programmer base potentially in the thousands all experimenting on techniques and trying to outdo each other, is going to produce some stunning results.
 
patsu said:
What exactly did MS do to qualify VC-1 as a standard ? What standards body owns the VC-1 specification today ? Who owns the VC-1 brand now ? These are business assets that can be used against MS's competitors later (e.g., obsoleting VC-1, preventing contribution to VC-1 by other vendors, deploying and marketing a "VC-2" ahead of others, etc.).

The MPEG-LA body is only an administrative, negotiation and collection entity right ? Does it hold any power w.r.t. the evolution of VC-1 ?
It looks like SMPTE owns it, according to wikipedia:

VC-1 is the informal name of the SMPTE standard 421M describing a video codec based on Windows Media Video version 9. On April 3, 2006, SMPTE announced the formal release of the VC-1 standard as SMPTE 421M.
 
Silly me. I should have checked wikipedia or google first. The press release link on the wikipedia site (http://www.smpte.org/news/press_releases/003_06.cfm) also answered some of my concerns regarding reference source availability. Although MS is the only company with a complete VC-1 implementation, the Blu-ray momentum would bring about other independent implementations.

Thanks for the help again. The pressure is on Sony to release a dual layer Blu-ray media asap.
 
Sis said:
This makes no sense. Once the standards are set, as they are now, then any PC capable of playing Blu-ray or HD DVD will be able to play the encoded material, whether the encoder originated from Microsoft or not. Otherwise, the hardware based decoders would require firmware upgrades when new movies came out. In my opinion, given the strong push from Microsoft for VC-1, it is more likely to become the standard just because they are being so helpful.

The problem comes when standards can be changed by whoever controls it to benefit one particular vendor as happens constantly in PC media players. For example Microsoft could improve VC-1 in future (say VC-2 standard) and refuse to license it to anyone else so that you could only use it on Windows Media Center or Windows Vista. Alternatively, they could charge a very high price for it's use so that other player manufacturers would be at a disadvantage. Another thing they could do is to incorporate seemingly innocuous licensing terms that prevent competing vendors from using it. An example of this is the open source programs (eg. embedded Linux) which the majority of embedded device manufacturers use in set top boxes and media decoders which require the license to be sublicensed on to the end user. By putting in a requirement for a written and signed license to be taken out directly with Microsoft for every sublicense, Microsoft can make it impractical to use VC-2 with embedded any media player other than Windows CE. This very tactic has been used only a couple of months ago by Microsoft to prevent it's supposedly open standard for MS Office formats OpenXML and it's Windows networking protocol to prevent it's only real competitors OpenOffice and Samba from using them.

No one other than Microsoft can improve VC-1 because they own all the patents to it and control the standards "commitee" that is supposed to manage it.

The answer to your post is no: a "standard" controlled and owned by a single company is not a standard, but a monopoly lock-in. Standards should be controlled by a neutral standards body which no single vendor can control. Ideally, to be a truely open standard, it should also be royalty free (because high royalties can be used to disadvantage competing vendors) and require each contributor to grant a royalty free patent grant for use in connection with that standard in perpetuity for any patents it holds in relation to that standard (because whatever the standards body says, the owner of the patent can still sue for patent infringement and obtain an injunction to prevent use of the patent by whoever they choose, and these patents don't need to be revealed).

As to whether Microsoft is being helpful, if Microsoft really wants to be helpful, it should allow a royalty free patent grant of any patents it holds in connection with the standard, and with future improvements to the standard whether made by Microsoft or others.

It is foolish to trust any company not to abuse control over "standards" they control. Why shouldn't they? After all it is so easy to do and they are in business to make money after all. Hell, if I controlled a standard, I would do it. However in the case of Microsoft, you are dealing with a felon convicted in three continents, and an unrepentant serial offender who is still unabashedly at it - it is unwise to the extreme. Just take a look at what they have been doing just a few months ago in relation to PC media players and licensing their "standards" to competitors.

http://news.zdnet.co.uk/software/windows/0,39020396,39233873,00.htm
http://news.com.com/Critics+Microsoft+server+license+snubs+open+source/2100-7344_3-5555078.html
 
SPM said:
if Microsoft really wants to be helpful, it should allow a royalty free patent grant of any patents it holds in connection with the standard, and with future improvements to the standard whether made by Microsoft or others.

"Agreed"
-Memory stick, Bluray, Minidisc, UMD, BetaMax :)
 
Last edited by a moderator:
SPM said:
The problem comes when standards can be changed by whoever controls it to benefit one particular vendor as happens constantly in PC media players. For example Microsoft could improve VC-1 in future (say VC-2 standard) and refuse to license it to anyone else so that you could only use it on Windows Media Center or Windows Vista. Alternatively, they could charge a very high price for it's use so that other player manufacturers would be at a disadvantage. Another thing they could do is to incorporate seemingly innocuous licensing terms that prevent competing vendors from using it. An example of this is the open source programs (eg. embedded Linux) which the majority of embedded device manufacturers use in set top boxes and media decoders which require the license to be sublicensed on to the end user. By putting in a requirement for a written and signed license to be taken out directly with Microsoft for every sublicense, Microsoft can make it impractical to use VC-2 with embedded any media player other than Windows CE. This very tactic has been used only a couple of months ago by Microsoft to prevent it's supposedly open standard for MS Office formats OpenXML and it's Windows networking protocol to prevent it's only real competitors OpenOffice and Samba from using them.

No one other than Microsoft can improve VC-1 because they own all the patents to it and control the standards "commitee" that is supposed to manage it.

The answer to your post is no: a "standard" controlled and owned by a single company is not a standard, but a monopoly lock-in. Standards should be controlled by a neutral standards body which no single vendor can control. Ideally, to be a truely open standard, it should also be royalty free (because high royalties can be used to disadvantage competing vendors) and require each contributor to grant a royalty free patent grant for use in connection with that standard in perpetuity for any patents it holds in relation to that standard (because whatever the standards body says, the owner of the patent can still sue for patent infringement and obtain an injunction to prevent use of the patent by whoever they choose, and these patents don't need to be revealed).

As to whether Microsoft is being helpful, if Microsoft really wants to be helpful, it should allow a royalty free patent grant of any patents it holds in connection with the standard, and with future improvements to the standard whether made by Microsoft or others.

It is foolish to trust any company not to abuse control over "standards" they control. Why shouldn't they? After all it is so easy to do and they are in business to make money after all. Hell, if I controlled a standard, I would do it. However in the case of Microsoft, you are dealing with a felon convicted in three continents, and an unrepentant serial offender who is still unabashedly at it - it is unwise to the extreme. Just take a look at what they have been doing just a few months ago in relation to PC media players and licensing their "standards" to competitors.

http://news.zdnet.co.uk/software/windows/0,39020396,39233873,00.htm
http://news.com.com/Critics+Microsoft+server+license+snubs+open+source/2100-7344_3-5555078.html
First, as noted in a previous post, SMPTE is the standards body for VC-1. If Microsoft wants to create a VC-2, which they have every right to do, they can. But it won't be used by HD DVD or Blu-ray, so it has no bearing on anything in this discussion.

Second, asking one corporation to give up patent rights while everyone else in the patent pool doesn't is a tad hypocritical.

Third, breaking anti-trust laws is not actually illegal in the criminal sense, so Microsoft is not actually a felon.
 
patsu said:
Silly me. I should have checked wikipedia or google first. The press release link on the wikipedia site (http://www.smpte.org/news/press_releases/003_06.cfm) also answered some of my concerns regarding reference source availability. Although MS is the only company with a complete VC-1 implementation, the Blu-ray momentum would bring about other independent implementations.

Thanks for the help again. The pressure is on Sony to release a dual layer Blu-ray media asap.

If this is like all the other "standards" bodies that Microsoft works with, the other members are only able to make suggestions and implementations, and Microsoft controls the show. Implementations are irrelevant. Unless it is secret, anyone can create an implementation of any specification. What is important is who owns the patents that control the standard. I bet Microsoft and nobody else owns all the patents relating to VC1 - no cross licensing, no others contributing IP.
 
Back
Top