Skip to main content
Topic: Can OGG get better than AAC? (Read 5147 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Can OGG get better than AAC?

Well, I've been reading some threads in this wonderful site and I came to the conclusion that nowadays the state-of-the-art audio format is AAC, but OGG is getting better for each release of a RC. Does OGG framework allows it to get better than AAC? if yes, how far is it?

Thx:)

Can OGG get better than AAC?

Reply #1
Anything is possible, even getting better than MPC. (Although, of course, that's unlikely.)

I believe you should wait to see what happens. It CAN get better, but it's impossible to say that it WILL.

Besides, AAC is a standard, there are lots of AAC encoders implementations. (Although we only have access to 3). So, it can get better than Psytel, and I guess it's already better than FAAC, but who knows if it can get better than FhG?

Regards;

Roberto.

Can OGG get better than AAC?

Reply #2
What software contains the fhg aac codec and where can I get said software?
r3mix zealot.

Can OGG get better than AAC?

Reply #3
There are 2 FhG AAC codecs.

FhG AACdemo 3.0 is the consumer - fast - slightly low quality AAC codec. It's bundled with Liquifier. Unfotunately, you can't create ISO AAC streams from liquifier. The files it creates (LQT) are encrypted.

FhG AACdemo 2.2 is the professional - slow - high quality AAC codec. It isn't bundled with any software, it only exists in the form of a "secret" command line encoder. Noone I know has this encoder.

You gan get Liquifier 5.0b24 in some warez FTPs. If you don't find it around, PM me <cough>.

Regards;

Roberto.

Can OGG get better than AAC?

Reply #4
Quote
Does OGG framework allows it to get better than AAC?


Streaming Layer:

The Ogg format has a nice simplified and powerful structure if you ask me. If you are talking in terms of the streaming layer, meaning the packet structure and API programming interface with various codecs such as Vorbis for audio and Tarkin for video when the time comes, then im convinced the structure is well ahead of AAC in that department. Especially with the idea of the packet structure consuming the least amount of overhead and the idea of Ogg Logical/Physical Bitstream Sequential and Concurrent Multiplexing of streams, then it definitly get's a plus in that section. I am sure the programming libraries will become more rhobust in time as well.

Anaylsis Layer:

The physcoacoustic model has come a long way and is still always undergoing work. Overtime it will gradually become better, I wouldn't say that both of the Codecs are better than each other in one way or another. I will however say that both physcoacoustic models differ from one another and have positive and negative aspects worth paying attention to in a sense. AAC has some interesting things in there physcoacoustic model that are worth looking at and Vorbis has diffrent physcoacoutic principals as well. It's up for you to decide which one you like. In my personal I opinion though I like Vorbis, because it open-source, free and the physcoacoustic model can be modified and changed over the course of time to either improve upon it or strengthen it. AAC is a standard however and looks to make various aspects of the codec better. New physcoacoustic principals and ideas to use for future designs. It is flawed however, because it is patented and serves no good to us in that sense.  From an economic standpoint yes, definitly not a technical one though. We don't want a hostile takeover on the net over corporate interest. Internet audio was made to be free along with system of vast information, as they say on xiph.org "It keeps everyone honest".
budding I.T professional

Can OGG get better than AAC?

Reply #5
AAC as a standard does not have "psychoacoustic model" - each implementation has its own psychoacoustic algorithm. ISO describes some basic model, called "Model II" - but any advanced implementation (fhg, psytel, etc..) uses more powerful processing.

That also applies to Ogg - if someone goes to real-time implementation of Ogg on a cheap DSP - he would propably have to use simplified psychoacoustic processing, etc..

Regarding file format - AAC file format is as flexible as it gets, when it comes to >audio only< bitstream. That is the purpose of ADTS/ADIF AAC stream. For mixing audio + video + other data (tags, html, 3D scene description, XML) - MPEG-4 file format is used.

MPEG-4 Systems layer (file format, etc..) is probably the most flexible multimedia container used today. Check out the:

http://www.envivio.com/solutions/etv/sample.jsp

To check what MPEG-4 file format could contain.



Note - I'm not saying here that AAC or MPEG-4 is better solution than Ogg - but think again - check out who is behind MPEG-4, the number of research institutes and leading companies, I'm quite sure that they know their work.

Can OGG get better than AAC?

Reply #6
Everyone is entitled to there own opinion, however I think people should give Vorbis a fair chance which many never do because, it is not considered a "standard". I am quite aware of the MPEG-4 AAC container format and the work Fraunhoffer IIS does with perceptual audio coding, etc. I saw the official design on there website, It still serves no good to use though it's patented right? they must have also mentioned something about watermarking in the container format as well. Who wants that?. It's just another way for corporate companies to put chains around your wrists. They are good from the perspective with the work they do perceptual audio coding and the physcoacoustics algorithms they work with and design, but what good is it if they run around patenting everything? They can't go back and modify it. In Vorbis we can do that. When the time comes I am sure Vorbis will be just as good , I hadn't been referring to the actual container format. I was speaking of the simplicit design within Vorbis streaming model. That's what I had been intending when I said Vorbis is well ahead of AAC in that department, after reading the Vorbis documentation it had spoke of the MPEG packet structure being complex and difficult to work with and how a the much needed simplier design in the Ogg Streaming Model would benefit as it does in my opinion. The MPEG packet structure can be complex to work with I have seen it before.
budding I.T professional

Can OGG get better than AAC?

Reply #7
Quote
Originally posted by HotshotGG
Everyone is entitled to there own opinion, however I think people should give Vorbis a fair chance which many never do because, it is not considered a "standard". I am quite aware of the 


Well not being a "standard" is a very big problem for telco markets and other applications in standard-based environment, like DAB or 3G Wireless.

However, I don't see any problem in PC-based applications, vorbis is as "standard" as many other formats (RealAudio, WMA) and it offers advantages like being open and free.

Quote
MPEG-4 AAC container format and the work Fraunhoffer IIS does with perceptual audio coding, etc. I saw the official design on there website, It still serves no good to use though it's patented right? they must have also mentioned something about 


Not particularily - you would be surprised how many technologies are actually patented. Big companies don't care much.

Patenting things is really becoming a hype - and now all companies patent their IP in order to get a chance for revenue at some point in the future. Fortunately for Ogg, there are many algorithms that were invented in 20-30 years ago and missed their chance to become patented

Quote
watermarking in the container format as well. Who wants that?. 


Industry - digital cinema, record labels - the people who actually made money out of the content. Certainly not the end-users

Quote
It's just another way for corporate companies to put chains around your wrists. They are good from the perspective with the work they do perceptual audio coding and the physcoacoustics algorithms they work with and design, but what good is it if they run around patenting everything? They can't go back and modify 


What's good? It is called $$$, and for some applications - sh*t load of $$$ - so it is very attractive to patent everything that your engineers invent. I know - that sucks from a OpenSource point of view, but that's how corporate industry work for decades - I am not sure if anybody could change that

Quote
it. In Vorbis we can do that. When the time comes I am sure Vorbis will be just as good , I hadn't been referring to the actual container format. I was speaking of the simplicit design within Vorbis streaming model. That's what I had been intending when I said Vorbis is well ahead of AAC in that department, after reading the Vorbis documentation it had spoke of the MPEG packet structure being complex and difficult to work with and how a the much needed simplier design in the Ogg Streaming Model would benefit as it does in my opinion. The MPEG packet structure can be complex to work with I have seen it before.


Unfortunately - I can't agree that MPEG systems structure is much more complex than Vorbis - it is not - and there are already many applications with MPEG-4 streaming over the packed-based and synchro. networks (ISDN, ATM) - Vorbis certainly has advantages in openess, good quality and being free, but it is not simplier than MPEG.

On the other hand - MPEG is very well documented, in a form of an official ISO standard, with reference software and many implementators - this is certainly the part where Vorbis could improve.

Can OGG get better than AAC?

Reply #8
Quote
there are many algorithms that were invented in 20-30 years ago and missed their chance to become patented.  


If you are referring to MPEG ISO Standards MPEG-1 Layer 3 is only 15 years old pre-liminary work started in 1987. MPEG-2 AAC was a proposed standard that started with pre-liminary work in 1995. MPEG-4 AAC, was talked about at the 2002 ceBit conference about two months ago. Vorbis is free-open source codec, that is release under the GNU and GNU Lesser Public Liscense, The goal of creating the codec was for educational as well non-patented purposes to stop any corporate hostile takeover. if somebody whoever is interested in the codec, hardware developer of some sort wishes to take the codec they can take the physcoacoustics algorithm and modify it and sell it for there own, in return xiph.org get's some of the money in return to keep the project going. Like you had mentioned above. I don't think Monty had wanted a patent on the codec anyhow, if he did I am sure he and xiph.org would already be trying to get one. That is the purpose of the lesser public liscense.

I am not trying to put down the MPEG standards, becuause they are official standards and by all means where else would you get any research documention? I just like the idea of a non-patented open-source free codec, you can learn something from it and not have to worry about Fraunhoffer sending you a a nice big letter telling you to stop your project, because you are infringing on there patents in some way.

budding I.T professional

Can OGG get better than AAC?

Reply #9
Quote
Originally posted by HotshotGG
If you are referring to MPEG ISO Standards MPEG-1 Layer 3 is only 15 years old pre-liminary work started in 1987. MPEG-2 AAC was a proposed standard that started with pre-liminary work in 1995. MPEG-4 AAC, was talked about at the 2002 ceBit 


It is not the 'MPEG' standards that are patented, but some of the unavoidable underlying algorithms

MPEG-2 AAC, MPEG-4 AAC and MPEG-1/2 Layer III basically share the most of the joint patents. These patents were filed mostly in 1987-88 (FhG), 1992 (by AT&T/Dolby) and some new ones (like new tonality measure) in 1996.

So - nobody actually patented a codec - but unfortunately many of the technologies are based on patented processes.

MPEG itself is not responsible and does not perform any licensing - it just requires that patent rights are granted on a fair and reasonable basis to third partes. If you think twice - this is as good as it could be, because if MPEG required royalty-free models, nobody from the industry would be interested to participate in MPEG.

Quote
conference about two months ago. Vorbis is free-open source codec, that is release under the GNU and GNU Lesser Public Liscense, The goal of creating the codec was for educational as well non-patented purposes to stop any corporate hostile takeover. if somebody whoever is interested in the codec, hardware developer of some sort wishes to take the codec they can take the physcoacoustics algorithm and modify it and sell it for there own, in return xiph.org get's some of the money in return to keep the project going. Like you had mentioned above. 


That is very good, and I (just like the most of the community) am very glad that Monty managed to build a codec without infringing a single patent - that is certainly an excellent achievment.

But, while the Vorbis was done by one man, MPEG is not working that way - and thus it must require commercial developers (and patents, too)

Quote
I don't think Monty had wanted a patent on the codec anyhow, if he did I am sure he and xiph.org would already be trying to get one. That is the purpose of the lesser public liscense. 


The problem in MPEG is that not one company is "controlling" the codec development - you never know if some company would just pop up with the so-called "submarine" patent - this is the case with MPEG-x Video codecs where many companies claim patent rights on some parts of the algorithm.

Quote
I am not trying to put down the MPEG standards, becuause they are official standards and by all means where else would you get any research documention? I just like the idea of a non-patented open-source free codec, you can learn something from it and not have to worry about Fraunhoffer sending you a a nice big letter telling you to stop your project, because you are infringing on there patents in some way. 


I agree - patent-free codec is certainly a very nice thing - but limiting to something that is not patented is also limiting your research abilities - for example, LAME team could implement anything they want, because they don't care about patents

For this moment - patented codecs are reality, because the most intense development is done in research institutes, and they are connected with MPEG - I also regret that most of the MPEG members like patent-based revenue generation, but there are some exceptions - like for MPEG-4 SA (Structured Audio) where entire algorithm and even implementation was released in public-domain (that is - completely free, you can do whatever you like with it)!

Can OGG get better than AAC?

Reply #10
Quote
Originally posted by Ivan Dimkovic
MPEG itself is not responsible and does not perform any licensing - it just requires that patent rights are granted on a fair and reasonable basis to third partes.

I suppose the 'fair and reasonable' went away in MPEG4

Can OGG get better than AAC?

Reply #11
Hehe.
budding I.T professional

Can OGG get better than AAC?

Reply #12
MPEG-4 AAC license (Audio) is pretty 'fair and reasonable' for commercial business.

MPEG-4 Video license wasn't so good in the announcement, but MPEG-LA announced that it will review the terms and add new business models in order to accelerate the adoption.

Also, "fair and reasonable" is a very very undefined term - fair and reasonable for whom? Big businesses, small businesses or the opensource community?

By the way - it is the patent holders that are causing problems, not the MPEG itself.

 
SimplePortal 1.0.0 RC1 © 2008-2019