Nero H264 AVC vs VP6.2.6

[SpaceV]

Hi guys,

compiled Shrek2 with Nero Standard-AVC 2pass @ 950 (all default settings) and VP6.2.6 with 1 pass best Quality 3/30 quantisizer @950
and hands down, VP6 has a much better quality, I can barely see room for improvement compared to DVD, no visible scene with pixels or strong detail loss. With Nero I saw many of those in the first few minutes of the movie. All original size 16:9 etc...

I set Nero to fit 80min CD but it only used 560MB, my VP6 is 650MB.

Both encodes took about 5 hrs on a AMD64 @ 2.45GHZ

Is ON2 ist still ahead of the game? I think if not VP7 will be.
What do u guys think?

Oh, and on the Nero encode the sound is not in sync, on VP6 it is. Both were encoded from the same source.

[bond]

Originally posted by SpaceV
compiled Shrek2 with Nero Standard-AVC 2pass @ 950 (all default settings) and VP6.2.6 with 1 pass best Quality 3/30 quantisizer @950
and hands down, VP6 has a much better qualitytell me: how do you want to proove this statement?
i want to tell you that claims alone are not enough on doom9 ;)

I set Nero to fit 80min CD but it only used 560MB, my VP6 is 650MB.so both actually dont use 950kbps as you stated at the beginning?

Both encodes took about 5 hrs on a AMD64 @ 2.45GHZso neros avc codec is twice as fast as vp6? (2pass vs. 1pass)

I can barely see room for improvement compared to DVD, no visible scene with pixels or strong detail loss. yeah, i even heard that vp3 shows no difference to the dvd source at 500kbps, come on... ;)

[SpaceV]

Both were set @ 950, before I started encoding.

The file size is the result. Seems like they are not both @950
so its not a real fair comparison I guess.

On Neros 1st pass I got about 35-40fps and on the 2nd 10-11fps.
VP6 I got about 7-9fps. So yes, Nero is faster.

Nero is probably as good as VP6 when they really use the same bandwidth but I dont think the quality of H264 is revolutionary compared to existing codecs. Maybe I encode with Nero again and use higher settings. I also need to find out why my audio is not in sync....

I think my video plays at 30fps... need more time.

The first post were just first impressions....
And I am also a little biased bc I am for the "little" guy.
That's why I have an AMD and like VP6 :)

[bond]

Originally posted by SpaceV
The file size is the result. Seems like they are not both @950
so its not a real fair comparison I guess.at least its not perfect, yes

On Neros 1st pass I got about 35-40fps and on the 2nd 10-11fps.
VP6 I got about 7-9fps. So yes, Nero is faster.interesting, are you sure you used neros avc codec, and not the normal asp one?

Maybe I encode with Nero again and use higher settings.i dunno the default settings, but make sure that you enable cabac and b-frames

I also need to find out why my audio is not in sync....strange indeed

[Taurus]

Originally posted by SpaceV
On Neros 1st pass I got about 35-40fps and on the 2nd 10-11fps.
VP6 I got about 7-9fps. So yes, Nero is faster.

I don't know what kind of super machine you are using :D ,
but as bond mentioned, are you sure, you used new nero avc codec?
Same settings used on my machine resulted 3-4fps on 2nd pass :devil: :D

Try again!

Cheers

Taurus

[Sharktooth]

He got an overclocked athlon 64.
That speeds are quite believable...

[chilledoutuk]

i dont see the point in on2's codecs to me its just another proprietry codec which nothing else supports. Its not even open source.

[stephanV]

Originally posted by chilledoutuk
i dont see the point in on2's codecs to me its just another proprietry codec which nothing else supports. Its not even open source.

and of course Nero is opensource! oh wait...

video quality evaluation has nothing to do with seeing the point of anything. one looks better, or they look about the same. be glad there is competition.

[iwod]

Well at least he should have conducted a fairer test. I mean there is like 90Mb Different!!!!!!!!!!

i get the point you are trying to say that H.264 is not revoluationary...... :D

[stephanV]

Originally posted by iwod
Well at least he should have conducted a fairer test. I mean there is like 90Mb Different!!!!!!!!!!
True, but someone's personal opinion about a company will contribute little to a fairer test. I think On2Tech deserves a little more than to be trolled away, as it *has* provided the basis of an open-source and totally free codec: theora. Personnaly, i do not understand why people always blindly follow the demi-gods called "open standard" and "open source". Of course having said that, it does not mean I think VP6.2 is better than Nero's H.264... or the other way around.


i get the point you are trying to say that H.264 is not revoluationary...... :D
If your comment is aimed at me, then you are wrong. I have made no comments about the quality of H.264 in general or Nero in particular in this thread.

[Sharktooth]

Well i like more Vp6.2 coz there is a VCM and for the ability to control sharpness from inside the codec. So even the high bitrate encodes can look as good as the original (grain and noise included).
The only remark i have for Ateme AVC is the inability to reproduce results similar to a 1:1 copy at high bitrates like other codecs do (xvid, vp6.2.x and others).
I prefer Vp6.2 also coz of the "lightweight" decoder and i can play backups on my xbox with xbmc...

[chilledoutuk]

Did i say nerodigital was open source oh wait...

No I was simply stating that I dont see the point in developing completly proprietry codecs as there very unlikely to be used over ones based on widely adopted mpeg or other open stndards.

oh and about theora yes thats all great and all but the codecs at a disadvantage already being a on2 hand me down (vp3 I believe) which obviosly they didnt want in the first place.

quite frankly i dont think you should judge the performance of h.264 soly on this recode release.

If you had experiance with the early versions of recode you would have noticed the massive improvent in the asp codec over the last year.

I expect they will also improve there h.264 codec in the same way.

[stephanV]

Originally posted by chilledoutuk
No I was simply stating that I dont see the point in developing completly proprietry codecs as there very unlikely to be used over ones based on widely adopted mpeg or other open stndards.
Used by whom? EA games for example? Quite unlikely that VP6 will be used by them indeed.

And what do you mean "widely adopted MPEG standards"? The only widely adopted and MPEG standard is MPEG2 in the form of DVDs. But the DVD standard is a severely limited form of MPEG2. And where are those MPEG4 ASP players?

oh and about theora yes thats all great and all but the codecs at a disadvantage already being a on2 hand me down (vp3 I believe) which obviosly they didnt want in the first place.
Yes Theora is based on VP3, but i dont see why that makes it bad or a disadvantage. They could have chosen not to release the source. Why do people always have to seek bad things behind a good deed?


quite frankly i dont think you should judge the performance of h.264 soly on this recode release.
Quite frankly i think this thread is called "Nero H264 AVC vs VP6.2.6", so Nero's AVC is exactly what should be judged. No one here was talking about judging standards in general. So please make comments about that, and don't give your weird opinions if there are points or not. You could question the point about this whole forum as there are more important things going on in the world. But if you think this thread is pointless, then waste your time on something else please.

[Manao]

Just a few comments :

StephanV : how did you manage not to reach the wanted size ? since the latest beta, nobody ever succeeded in making the rate control of the codec fail.

Also, when chilledoutuk speak of uses of the open standard codecs, he is obviously speaking of set up box & HD DVD.

Sharktooth : The only remark i have for Ateme AVC is the inability to reproduce results similar to a 1:1 copy at high bitrates like other codecs do (xvid, vp6.2.x and others).I disagree. AVC / H264 is quite capable of making a 1:1 copy of a source. Of course, it has troubles to reproduce mpeg2 artifacts ( ringing, quantification noise ), since its DCT differs totally from mpeg's one.

[stephanV]

Originally posted by Manao
Just a few comments :

StephanV : how did you manage not to reach the wanted size ? since the latest beta, nobody ever succeeded in making the rate control of the codec fail.

I did not, that would be SpaceV :)
Of course it also might have been VP6.2 that turned out to be too large, I wouldnt know.


Also, when chilledoutuk speak of uses of the open standard codecs, he is obviously speaking of set up box & HD DVD.
I was too:
And where are those MPEG4 ASP players?
The only 3 usuable standards we have *now* are forms of crippled MPEG1 and crippled MPEG2 (VCD, SVCD and DVD). The 4th and 5th ones would be DivX and (right about now) Nero Certified standards which are both crippled (again...) MPEG4 ASP. If there is one thing that is shown here, it is that MPEG does not really define the standards. Companies, however, do.

But that was never my point. Just don't like it when someone drops in a thread saying a codec is bad because it's not an open standard or because it is not *even* open source. Since when are open standards or open source an absolute guarantee for quality? By the same reasoning Philips and Sony should never have developed the compact disc because they are incompatible with my old grammophone. Not that i want to compare VP6.2 with a CD and Nero with an old LP... not at all.

[Manao]

My bad for the confusion Space / Stephan :rolleyes:

For the filesize, I checked the duration of the movie ( 92 minutes long ), and with the bitrate he gave, it's recode which is undersizing.

For h264 support, it will be greatly different from the mpeg asp mess, since h264 will be used in HD DVD, and for HDTV broadcast. Of course, it doesn't say anything about the quality of the codec, only of its usage.

[Sergei_Esenin]

Since when are open standards or open source an absolute guarantee for quality?

Not to get too off topic, but when comparing codecs I think it's fair to take into account their status as open source or open standards--not because of quality, but because of longevity. Open or common standards are no guarantee of quality--there's a lot of great proprietary software and formats out there. But open standards and sources are a guarantee of longevity and long-term usability. A decade from now I know all my rare multimedia, transferred from laserdisc, betamax, film, VHS, and unlikely ever to be released on DVD or its successors, will still be as accessible as ever without the added quality loss of an additional transcoding. That's because when I captured and cleaned them I saved them to high-bitrate, low-quant, standards-compliant MPEG-4 ASP, a format which has so many great open-source decoders across so many codebases and OSes that even if PC and OS architectures change radically in the future I'll still be able to easily use my files. Thus is the advantage of open standards and open-source code.

The same can't really be said for proprietary standards with closed sources. Will encodes made with proprietary encoders to closed standards still be usable a decade from now, except on antique hardware running antique software, or with an additional quality-losing transcode from the antique format to a new one? It doesn't matter for many common uses, but for true archival purposes it's perhaps the most important issue. It's the reason I stick to open standards when I digitize rare material--MPEG-4 ASP right now, and hopefully AVC when encoders are mature and capable of transparency at high bitrates. Just food for thought.

[chilledoutuk]

Originally posted by stephanV
Since when are open standards or open source an absolute guarantee for quality?

No but there is a guarantee for universal usability though.

If on2 decided to make vp6 an open standard after they have refined it then great but is this really likely to happen?

I am not saying you shouldn’t use vp6 but I consider it risky to use it as a sole backup method as it is a codec totally dependant on one company for future compatibility and support.

For example if like say DiVX network were to go bankrupt (something that could happen) then because there codec's encodes were mpeg-4 compliant you would still be able to use an array of different playback filters that are also based on the mpeg-4 standard to watch and decode your encodes without any issues.

You and I seem to be looking at this from different aspects you clearly like the codec’s quality which is completely understandable because well I do as well. But I am looking at the technology and company on2 from a commercial aspect.

You see on2 is competing in a market that is very competitive and will only get more so. for there codec to be successful in adoption by companies making hardware for playback in users living rooms they would need to almost need to give the codec and licensing rights and patents away just in order to compete with standards such as mpeg-4 asp and avc.

Also remember that the mpeg-4 standard isn’t just a video codec it specifies an standard for complete multimedia content creation and delivery for almost anything that could be needed by content providers.
This will be a very influential aspect of any av standard and to me one of the most important as it enables the content provider the luxury of knowing that almost anything they may want to achieve will have an outlined and accepted method for doing so.

I believe that on2 make some very efficient codec’s I just wish they tried to base there work upon an open standard as sometimes efficient cooperation between many companies is extremely difficult if standards aren’t outlined and agreed upon.

[Soulhunter]

Originally posted by Manao
Sharktooth : I disagree. AVC / H264 is quite capable of making a 1:1 copy of a source. Of course, it has troubles to reproduce mpeg2 artifacts ( ringing, quantification noise ), since its DCT differs totally from mpeg's one.

Veto, it has also problems to reproduce film grain and normal* noise !!!

* noise that was already in the raw source (film) before compressing it to MPEG2...


Bye

[LordRPI]

Originally posted by chilledoutuk

For example if like say DiVX network were to go bankrupt (something that could happen) then because there codec's encodes were mpeg-4 compliant you would still be able to use an array of different playback filters that are also based on the mpeg-4 standard to watch and decode your encodes without any issues.


O_o

OT: My university, RPI, has some research thing going on with On2. http://www.eng.rpi.edu:8080/soenews/update.do?artcenterkey=287

[stephanV]

Originally posted by chilledoutuk
No but there is a guarantee for universal usability though.
No, there is not. DivX players do not play all XviDs, they do not even play all DivX content. So how useable is that? Useable, but certainly not universal.

If on2 decided to make vp6 an open standard after they have refined it then great but is this really likely to happen?

Don't know, don't care.

For example if like say DiVX network were to go bankrupt (something that could happen) then because there codec's encodes were mpeg-4 compliant you would still be able to use an array of different playback filters that are also based on the mpeg-4 standard to watch and decode your encodes without any issues.
1. On a PC this doesnt make any difference. I will still always have the playback filters installed.
2. Even so, in the past there have been multiple examples of MPEG4 playback filters of companies that b0rk on eachother encodes.

MPEG4 ASP has not provided the compatibility that was hoped for. Don't know who's fault that is, but it has never really been there. I think we are still waiting for a fully compliant MPEG4 ASP hardware device.

Again, its not MPEG that decides what gets used. It's the companies that do. If companies say: "no QPEL", then there will be no QPEL.


You see on2 is competing in a market that is very competitive and will only get more so. for there codec to be successful in adoption by companies making hardware for playback in users living rooms they would need to almost need to give the codec and licensing rights and patents away just in order to compete with standards such as mpeg-4 asp and avc.
ASP and AVC are not *that* cheap. And with a TV out one could still get pretty far, certainly in the future. It would be a better solution anyways, as the only thing current hardware devices do is cripple you in your choices.

Also remember that the mpeg-4 standard isn’t just a video codec it specifies an standard for complete multimedia content creation and delivery for almost anything that could be needed by content providers.
I'm not a content provider, I don't care. And I'm certainly not a fan of all-in-one packages, which are again crippling my choice. Perhaps nice for a content provider, certainly not nice for me.

The difference between you and me is, that you are thinking as a company or industry. I'm thinking as an individual. :)

Not saying you are wrong in doing that, but you have by no means a valid arguementation that On2techs codecs are pointless.

[Sharktooth]

On2tech codecs are used for streaming content and movies, video games (see EA's NFSU2 for example) and more...
On2 is "active" company and have their own and well supported video/audio solution (VP6 and AVC... well the audio codec name may be confusing some ppl but that's it...).
I can't say the same thing for "MPEG-x standard" codecs. They're badly supported and sometimes badly screwed (depends on the developers), not following standards at all or best of all there are cases when they're "constrained" in poor developed GUIs (look at Recode 2.2... the AVC codec is good but the GUI is horrible, limited, full of problems, doesnt exploit the full potential of the ateme codec and is pervaded by bugs with serial number checking and in the installing phase...).
The only exceptions in this MPEG world are the opensource developers that try to follow the standards as much as they can. Results are codecs like Xvid (the best mpeg4 ASP codec)... but, for a company view, relaying on an open source project is like shooting in your feet with a 12 gauge shotgun. But sometimes they do, and without the author's permission (look at Kiss DivX players)...

[chilledoutuk]

people who think there are problems with compliancy within the mpeg-4 standard should note that the only reason for this is because some companies think the mpeg-4 standard is a pick and mix choosing what features they want without regard as to whether there decoder will be mpeg-4 compliant or not ( ahem DiVX network).

one thing i will say about ateme and ahead Nero Digital is that it as a product abides to the mpeg-4 standards better than any other solution around at the moment.

A testament to this is that a mpeg-4 encode from recode will play flawlessly in any mpeg-4 compliant media player such as VLC, mplayer etc without any decoding issues what so ever.

Now considering that the decoding filters these players used were designed in coordination with the specified mpeg-4 standard and nothing else it clearly shows why standards are needed.

It also shows recodes ability to make mpeg-4 standard compliant videos and the ffmpeg group’s ability to decode them without any issues.

I hope you’re not questioning the importance of standards to consumers. as people have trouble as it is keeping up with technology developments if standards were not created we would have lots of different incompatible technologies and would lead to consumer confusion. Ultimately causing consumer anxiety and hesitation meaning one thing a badly performing tech industry.

@stephanV I find your constant dissection of my post immature please find a more interesting way of conveying your point.

on the issue with DiVX players I don’t consider these players MPEG-4 ASP players as they cont comply with the MPEG-4 ASP standard fully if you take a media player that does comply with MPEG-4 ASP like XBMC (basically mplayer which uses ffmeeg) then you will see that the issue isn’t that the standard is a turd but rather that the hardware manufactures haven’t caught up yet and gained full compliancy but they will eventually. Then they will be able to play any mpeg-4 ASP compliant video stream.

Now everyone knows that XViD is an open source codec and if a company was to rip there code then that would be bang out of order. But I don’t think there would be an issue if a content provider decided to use XViD to create mpeg-4 compliant streams as long as they open about using an open source project codec for there encoding. (I don’t think the devs would mind as long as they were recognised for there work).

Because of this thread i recently done some encode with vp6 and I have to say that XVID is clearly a better quality codec to my eyes. Perhaps I just like the look of mpeg4.

[Sergei_Esenin]

DivX players do not play all XviDs, they do not even play all DivX content. So how useable is that? Useable, but certainly not universal.

Then use a less limited, more standards-compliant decoder like XviD or ffdshow. DivX only implements a smaller subset of ASP features.

MPEG4 ASP has not provided the compatibility that was hoped for. Don't know who's fault that is, but it has never really been there.

There are complete standards-compliant implementations for encoding and/or decoding like XviD and ffdshow, and have been for some time. The common newbie-friendly DivX encoder produces standards-compliant output (using fewer options), and has for a long time. There's even a tool now for converting old non-standards-compliant DivX 3.11/Microsoft MPEG-4 files into standards-compliant files without recompression

I think we are still waiting for a fully compliant MPEG4 ASP hardware device.

I don't know or care about hardware devices. They're limited and I see no point in a tech-savvy media lover using one when they're easily bettered by a Media PC or Home Theater PC. If I need interoperability for sharing content with people who use only limited set-top box hardware devices, burning a DVD downconversion of my MPEG-4 masters is a one-step process using tools like NeroVision or similar packages.

But it boils down to this: ASP never gained universal hardware support only because there was no mass-market consumer multimedia format associated with it, the way MPEG-2 had DVD and MPEG-4 AVC will have HD-DVD and Blu-ray. Even so, as an open standard for which so much open code has already been written there will always be compliant decoders for current platforms, and from a long-term usability standpoint that's what's paramount. Yes, for the short-term other factors may intrude, but for carrying your files forward indefinitely open standards are key.

[stephanV]

Originally posted by chilledoutuk
@stephanV I find your constant dissection of my post immature please find a more interesting way of conveying your point.
It is not a dissection of your post. It's just to make clear to which point i am responding. I can't see anything immature about that, lots of people respond to posts that way. You only have to look at this thread to see that. But ok, suit yourself. You have brought nothing new to the discussion other than that you like MPEG4 more than VP6.2 (big surprise! but very unfounded though...) and made some weird statements that everyone is free to use XviDs code as long as they leave a "thank you"-note. I think MPEGLA would have something to say about that; licences need to be paid for. "Open standard" does not always equal "free".

Originally posted by Sergei_Esenin
Then use a less limited, more standards-compliant decoder like XviD or ffdshow. DivX only implements a smaller subset of ASP features.

I was talking about DivX hardware players. :)

[Paazabel]

Originally posted by Sergei_Esenin
I don't know or care about hardware devices. They're limited and I see no point in a tech-savvy media lover using one when they're easily bettered by a Media PC or Home Theater PC.
There are many problems, here ...

First, I have yet to find a PC-based 30i video output that works well in a home theater system. I have made nice PCs with nice videocards, but when I compare that to my KISS player, the KISS wins hands down on how it looks on the TV.

Second, although PCs are getting cheaper all the time, this is still a more expensive road than buying a consumer player with embedded support. KISS is a fairly expensive option, but I've recently seen DivX players in the $80 range (compared to $299 I paid for the DP-500).

Third, the tech savvy do not rule the marketplace. Therefore, we do not always (unfortunately) drive the standard. The problem DivX has is that it has not accomplished, as you point out, the "triple play" -- a player, a camcorder, and Hollywood content all using their format. If On2 is able to accomplish that with VP6, then it will become a "standard" no matter how much better XviD really is (see VHS vs. Beta).

Bottom line, the "best" technology doesn't always "win." More often, it is most widely adopted technology, not the best technical or the most open standard which ends up in everyone's living rooms. To this end, this thread is an interesting comparison not because one is open and one is not, but just to see how they stack up one to the other. If VP7 comes out and is the end-all-be-all of video encoding, but comsumer devices and camcorders and content providers have all embraced MPEG-4 ... then it will go the way of the Betamax; a better technology that just never caught on.

A final thought on various MPEG-4 codecs. They look more similar now than different. I can see differences, but not huge, glaring, black-tile-crawling-across-the-screen differences like there used to be. This tells me that MPEG-4 is reaching a maturity level that may soon gather it some support from content providers or brand manufacturers like Pioneer and Sony. In the face of this, I'm not sure that VP6 would remain interesting, but that is for On2 to decide.

[Andrey]

>>1. On a PC this doesnt make any difference. I will still always
>>have the playback filters installed.
Hmm ? No offence,but how long are you in industry ?
I have had a tons of filters that did not work on win2k and will not work on any next M$ system... (example - many bik playing filters)
This could be a real problem, for sure.
Sometimes it forces me to use win95/win98 emulators to get those files played (which is obviously illegal, at least M$ is sure about it)

[stephanV]

tons? which ones? i've never used bik sorry.

[Paazabel]

Sorry, but I've been monkeying with this suff since Intel's Indeo was going to take the world by storm. I never have to leave WinXP ...

[ObiKenobi]

Originally posted by chilledoutuk
Now everyone knows that XViD is an open source codec and if a company was to rip there code then that would be bang out of order. But I don’t think there would be an issue if a content provider decided to use XViD to create mpeg-4 compliant streams as long as they open about using an open source project codec for there encoding. (I don’t think the devs would mind as long as they were recognised for there work).

Hate to be immature by dissecting your post, but this is complete bull. To be able to use XviD, or any other mpeg-4 codec for that matter, a content provider has to purchase a license to use that specific codec, they aren't just able to use it cause they give recognition to the developers. If one doesn't pay these licensing fees and are found out one can get into quite a bit of trouble with MPEGLA.

[Andrey]

>>To be able to use XviD for your content you have to pay mpeg-4
>>licensing fees, you aren't just able to use it cause you give
>>recognition to the developers
Things were not that simple, at least year ago.
You were able to use any mpeg4 encoder for personal use, on other hand, it was illegal to distribute one w/o paying the fees. That's why xvid was distributed as source code...
There were defined the next fees: for streaming, for encoder/decoder binary, for encoding a content.
But encoding a content fee begins on a huge hours number, so one person can encode his material w/o paying any fees (if person have appropiate encoder).
Another solution is to download or buy Divx and use Xvid instead. This is the question of court and lawers of course, but from my perspective I do not see any problems with it - encoder/decoder fee is payed.
Distributing xvid binary - is what is obviously illegal in our case.
P.S. I did not read latest MPLAA releases - may be fee politics have been changed since...

[stephanV]

Originally posted by Andrey

Things were not that simple, at least year ago.
You were able to use any mpeg4 encoder for personal use, on other hand, it was illegal to distribute one w/o paying the fees. That's why xvid was distributed as source code...
There were defined the next fees: for streaming, for encoder/decoder binary, for encoding a content.
But encoding a content fee begins on a huge hours number, so one person can encode his material w/o paying any fees (if person have appropiate encoder).
One person ok... he was talking about a content provider... not one person.

Another solution is to download or buy Divx and use Xvid instead. This is the question of court and lawers of course, but from my perspective I do not see any problems with it - encoder/decoder fee is payed.
Your license is not a general one, it is given for a specific encoder. You cannot let DivX pay for you using XviD. You can of course, but only if it explicitly says so in your license agreement with DivX. And then there are still are some ethical issues you have to deal with.

Im still waiting for all those DS-filters that don't work on your system anymore BTW :)

[chilledoutuk]

Originally posted by ObiKenobi
Hate to be immature by dissecting your post, but this is complete bull. To be able to use XviD for your content you have to pay mpeg-4 licensing fees, you aren't just able to use it cause you give recognition to the developers. If you don't pay your licensing fees and our found out you will get into quite a bit of trouble with MPEGLA.

hate to be imaature and Disect YOUR post but i said

decided to use XViD to create mpeg-4 compliant streams

however you decided to distort what I said. I was simply pointing out that if people were officially to create mpeg-4 asp streams for comercial use that they could use xvid.

I did not however mention anything about negleting to pay licensing fees for using the technoloy at hand and to assume that is what I meant is well just basically stupid and narrow minded.

You see if you had read the other posts you would understand the context within which my statment was made.
Because something was mentioned earlier about a company know as kiss ripping of an open source project mplayer i think and selling its product based around that open source technology for a profit.

Therefore when i made the statement.

Now everyone knows that XViD is an open source codec and if a company was to rip there code then that would be bang out of order. But I don’t think there would be an issue if a content provider decided to use XViD to create mpeg-4 compliant streams as long as they open about using an open source project codec for there encoding. (I don’t think the devs would mind as long as they were recognised for there work).

I was refering to people using opensource projects (ie XViD) instead of comercial products (ie DiVX) to enable them to create the products and services that consumers need.

please dont make assumptions about the meaning of any thing that is not clear in any of my statments simply ask and i will attempt to clarify.

[stephanV]

Dissecting time again :)

This is what chilledoutuk said:
But I don’t think there would be an issue if a content provider decided to use XViD to create mpeg-4 compliant streams as long as they open about using an open source project codec for there encoding. (I don’t think the devs would mind as long as they were recognised for there work).

The biggest as long as (or at least as important)is if they pay the licensing fees to MPEGLA, not if they say thank you to XviD. Not mentioning this is quite strange (and therefor you got that response).

It's your job to be clear, not ours to guess what you actually mean.

[Nic]

Ok, lets keep this thread civil, and try and keep words like "stupid", "bull", etc down to a minimum. Otherwise it will get closed.

The discussion is interesting though...as long as tempers don't flare :)

-Nic

[ObiKenobi]

Originally posted by chilledoutuk
I did not however mention anything about negleting to pay licensing fees for using the technoloy at hand

But all you said was that there would be no issues as long as the company gave a mention to the developers, and that would be wrong. They would still have to pay mpeg-4 licensing fees regardless of the fact that the codec was an open-source project or not.

Originally posted by chilledoutuk
I was refering to people using opensource projects (ie XViD) instead of comercial products (ie DiVX) to enable them to create the products and services that consumers need.

Yeah I understand what you were talking about, I was just clarifying that it doesn't really matter if you are using an open-source or closed-source codec you still have to pay licensing fees if the codec is being used for a commercial purpose. Just so you know as well, a project being open-source doesn't necessarily mean it isn't a commerical product, those aren't one in the same. And in reference to your quote again:

Now everyone knows that XViD is an open source codec and if a company was to rip there code then that would be bang out of order. But I don’t think there would be an issue if a content provider decided to use XViD to create mpeg-4 compliant streams as long as they open about using an open source project codec for there encoding. (I don’t think the devs would mind as long as they were recognised for there work).

I'm sorry but there would be issues if a company was using the XviD codec for providing content if they didn't have the license to do so. I can't speak for the developers but they may or may not care (but that's not really the issue), but the MPEGLA certainly would which was my point. Assuming they did have the license then yes there would be no problems in using it except if they failed to include the original source upon distribution of the codec, but that's a completely different issue dealing with the GPL.

[chilledoutuk]

I am getting a little tired of being persecuted because i failed to mention something that I thought was common knowledge.

The licensing fees have always existed for commercial applications of MPEG-4 that I have known for some time and didn’t think that I needed to mention it the he context of my statement.

My statement was really in response to something someone said further back about using opensource projects to create commercial products. What they were referring to was someone copying open source code from mplayer and using it in DiVX hardware player I was simply saying that the majority of devs of opensource projects don’t mind people using there code as long as they don’t try to pass it off as there own and give them recognition for there hard work.

I would greatly appreciate people not reading between the lines when I make posts as what they have assumed was a deliberate omission was simply an oversight.

If any thing I post on these forums does not make sense then please do ask questions in order to attempt to clarify our understanding of each others points of view.

And remember only take heed from what people say not what they don’t say.

[Sharktooth]

This discussion is going a way off from the original post...
However MPlayer guys are not happy at all of the code stealing (http://www.mplayerhq.hu/homepage/design7/news.html#kiss05) by KISS Technology.
Now, returning back to the AVC vs VP6.2.x discussion, i would also clarify there are other "minor" standards based on proprietary codecs.
For example the EVD is based on on2's VP6 video compression technology. I said "minor" 'cause the DVD Video is the most recognized and diffused standard for distribuiting video and data, but i would like to remember you all there are plenty of DVD formats and not all of them are standardized yet (DVD+R/RW formats for exmple) and however there are companies that sells BADLY mastered DVDs (that means the content is not as "standard" as it should be).
Open standard are theoretically good if companies strictly follows them, but when things tends to get more complicated (for example: too many profiles, to much expensive hardware is needed etc...) companies prefers the much less complicated "substandards" like the DivX profiles.
So all the fuss is about implementing something that is cheap but still considered a "standard" (or being part of a "bigger" standard).
Now what i'm asking myself is what the heck were MPEG consortium guys thinking when they developed the MPEG-4 advanced simple profile...

[Tommy Carrot]

Originally posted by Sharktooth
Now what i'm asking myself is what the heck were MPEG consortium guys thinking when they developed the MPEG-4 advanced simple profile...

Afaik mpeg4 was originally developed for mobile phones and other handheld devices, their intention was not to make a codec for high quality contents, so it's not a big surprise that the mpeg consortium didn't care about the hardware support for a long time. But at the end mpeg4 took a completely different route from what they originally intended. I guess they learned about the mistakes, and it won't happen again with AVC.

[Sharktooth]

It is already happening. Nero digital profiles...

[Andrey]

>>They would still have to pay mpeg-4 licensing fees regardless of
>>the fact that the codec was an open-source project or not.
Of course.
The fees are defined not by the project nature, but by MPEGLA consorcium :)
>>It is already happening. Nero digital profiles...
Hmmm...
I think, the problem here lies in the really huge range of devices that are interested in the mpeg4 implementation.
If you try to play 512x384 file on a mobile phone, you will obviously fail because of hardware limitations (of computing power). That's the main reason why profiles are needed...
But a lot of the mess is created by hardware manufacturers, that produce strange chips that are able to work with a strange feature combinations, like only 1bframe or no GMC.
Profiles are not evil, if used correctly...
Just my 2 cents :)

[Sharktooth]

Yes, there are already existing profiles for mpeg4.
Companies like DivX and Ahead are redefining those profiles (read my "rant" in the previuous page) with their own for convenience.
If companies continue in making their own profiles and claiming their codecs as MPEG-4 compliant then there will no longer be a standard anymore...
At this point MPEG consortium should revise (to accomodate hardware producers) or impose the current MPEG4 profiles.

[babayaga]

Originally posted by Sharktooth
Yes, there are already existing profiles for mpeg4.
Companies like DivX and Ahead are redefining those profiles (read my "rant" in the previuous page) with their own for convenience.
If companies continue in making their own profiles and claiming their codecs as MPEG-4 compliant then there will no longer be a standard anymore...
At this point MPEG consortium should revise (to accomodate hardware producers) or impose the current MPEG4 profiles.
When encoder makers try to find suitable decoders, they face a problem : none are really compliant to the claimed profile/level.

This was already the case for MPEG-2 (in interlaced support), of course MPEG-4 (I'm not sure one is fully compliant with SP and ASP, not even speaking of levels) and there is very little chance that this will be solved in AVC/H.264.

That's why there is a very strong need of sub-profiles/levels per application : the problem is not the same for movie backup, broadband, broadcast, video-surveillance etc.

[SladeCraven]

Totally true. The HW (for example Envivio products) and SW (divx, xvid, FFMPEG, 3ivx, ...) codecs available are NOT at all compliant with the level of tools required by the ASP level. Do you know RVLC and data partitioning in these products/tools ??? No. But RVLC and DP are part of the ASP profile. An i don't speak about the interlaced support....
So for example in H264, people will claim to be MP@L3 or MP@L4 compliant, but without any MBAFF support (for example...).
For example, what is the status of the ateme codec included in Nero regarding the MBAFF ??

[bond]

Originally posted by stephanV
If there is one thing that is shown here, it is that MPEG does not really define the standards. Companies, however, do.i wouldnt say that. mpeg indeed defines clear "substandards" to make their beast more usable: the profiles and levels
still the problem is that the companies dont follow these profiles/levels, but develop their own. of course this hurts interoperability (which is, and is often forgotten, one of the main goals of the mpeg standards)

imho there are the following reasons for private profiles/levels:
1) exclusion of competitors:
- dxn wants the players to be sold as "divx players", nero wants them to be sold as "nerodigital players".
- they have no interest in having their competitors supported or even (indirectly) mentioned, which would be the case if the players would be promoted as "mpeg-4 players"
2) limitations in the used technologies:
- the chips simply dont support the mpeg-4 asp profile as mpeg defined it (mpeg's fault?)
- note that the only mpeg-4 asp feature not supported on the 2nd generation chips is 3warppoint gmc, the rest is fully supported
- companies want their stuff to be compatible with as most players as possible (thats why they use as benchmark the worst chip existing: dont even allow qpel aso in their normal usage profiles) -> hurts quality
- dxn released a vfw based codec, which means their tools have to suffer the limitations caused by this (hard to implement b-frames, subtitles, chapters, more advanced audio formats, menus aso), nero has the advantage to have entered the market later (-> doesnt have to suffer these limitations by using newer technology)
3) making sure that things really work for even the "dumbest" user
- still this would also be possible when following mpeg-4 profiles (on the encoder and (hardware) decoder side)

from my point of view i think that nero pushes the support of mpeg-4 very close to how it was meant to be by mpeg (by supporting .mp4), enabling interoperability when the same coding settings are used (in no matter what encoder), still they also went the same way as dxn with defining own profiles, to ensure that there is a big "nerodigital" logo on every player supporting them and of course to simply circumvent problems with existing technology (basically now only the limitations on the hardware side)

still it shouldnt be forgotten that the only feature missing for full mpeg-4 asp support on the current hardware players is 3wp gmc
now with .mp4 and aac being supported we are only one small step away from having reached real interoperability as meant by mpeg :D ;)

[SladeCraven]

) exclusion of competitors: - dxn wants the players to be sold as "divx players", nero wants them to be sold as "nerodigital players". - they have no interest in having their competitors supported or even (indirectly) mentioned, which would be the case if the players would be promoted as "mpeg-4 players"
Exactly ! I would add that these companies could not afford advanced R&D in the video domain, so they can't gain royalties from the patents owned by Toshiba, Sony, ST, Thomson, Broadcom, Microsoft, ..., ...

still it shouldnt be forgotten that the only feature missing for full mpeg-4 asp support on the current hardware players is 3wp gmc
In "pure" MPEG terms, ASP includes data partitioning and Reverse VLC. I don't think that this consumer ASIC support these tools. An the interlaced mode does not work with the current ESS and Sigma Designs SoC.
So, according to my knowledge, there is NO full ASP compliance on the consumer market.
But i think that the MPEG group badly defines the ASP profile, by merging tools for low bitrate broadcasters and for high quality video. This is why the ASP profile is quite "difficult" to respect for consumers applications.

[babayaga]

Originally posted by SladeCraven
Do you know RVLC and data partitioning in these products/tools ???
The only SP and ASP fully compliant SW decoder (with a wide diffusion) I'm aware of, is of course NeroDigital's MPEG-4 decoder :D


Originally posted by SladeCraven
For example, what is the status of the ateme codec included in Nero regarding the MBAFF ??
In the current version, the NeroDigital AVC decoder is fairly well compliant with Baseline and Main Profile, this includes MBAFF and PAFF.

Levels is another issue for a SW decoder since it depends on the CPU power. Full compliancy to level 3.0 is probably beyond reach for current PCs.

[bond]

Originally posted by SladeCraven
I would add that these companies could not afford advanced R&D in the video domain, so they can't gain royalties from the patents owned by Toshiba, Sony, ST, Thomson, Broadcom, Microsoft, ..., ...lol, indeed ;)
still as a consumer i dont care about this. i want to know what the player is able to play when i buy it and saying that "it plays divx5 and nerodigital profile xxx" doesnt tell me the whole story

In "pure" MPEG terms, ASP includes data partitioning and Reverse VLC.not necessarily, to my knowledge error recovery was placed in own extensions to the normal profiles (not sure tough)

I don't think that this consumer ASIC support these tools.the error resilience tools resync markers and data partioning are supported on the available mediatek mpeg-4 decoder chip, widely used in hardware players

So, according to my knowledge, there is NO full ASP compliance on the consumer market.indeed

[temporance]

Originally posted by SladeCraven
Exactly ! I would add that these companies could not afford advanced R&D in the video domain, so they can't gain royalties from the patents owned by Toshiba, Sony, ST, Thomson, Broadcom, Microsoft, ..., ...These companies did not even exist when most of the research was done and the patents were filed! If their existence is commercially viable and they are paying their dues to the above mentioned dinosaurs then everyone's a winner.

But i think that the MPEG group badly defines the ASP profile, by merging tools for low bitrate broadcasters and for high quality video. This is why the ASP profile is quite "difficult" to respect for consumers applications. Agreed: and with this in mind, who can blame the likes of DXN for bastardizing the levels/profiles. Anyone can define their own levels / profiles - and it's not something that can be protected by patents. The real test of the usefulness of levels/profiles is whether they become widely adopted. It doesn't look like ASP is going to pass this test.

Btw, we have all this to come with AVC. Let's hope the AVC profiles are better mapped to hardware realities. There, nearly back on topic ;)

[stephanV]

Originally posted by bond
i wouldnt say that. mpeg indeed defines clear "substandards" to make their beast more usable: the profiles and levels
still the problem is that the companies dont follow these profiles/levels, but develop their own. of course this hurts interoperability (which is, and is often forgotten, one of the main goals of the mpeg standards)

So again I say:

If there is one thing that is shown here, it is that MPEG does not really define the standards. Companies, however, do. :)

If MPEG defines substandards/profiles and companies make their own afterwards anyway, then what do things like ASP really mean to me?