I just think that open source codecs like LAME, Xvid, x264, Vorbis are highest quality codecs and wisely popular. They are better than their commercial competitors.
Many people help to improve and find bugs of open source codecs.
Why there is no open source high quality AAC encoder? What happens with FAAC?

Useless poll. Anyway as with all open source projects, they need people with the necessary knowledge and time to help improve them. If there's no (or not many) people to maintain the project then development stalls.

* There is no need for AAC format at all. Plenty of open source codecs for other formats.

In my opinion, Nero does an excellent job of updating their AAC encoder as they often come to these boards. They also enlist the help of HA members to test things out. It may not be open source but it is free, updated often, and they are open to input. Apple is a bit more shut out with their AAC encoder but you will find an Apple AAC developer on here every once in a while (or on the Apple boards).

All-in-all, I see no need for an open source AAC encoder as FAAC had a hard enough time as it is and Nero is doing a good job with their current development processes.

Personally, I see no need for a replacement to the mp3 format just yet and there are far too many formats out there. From personal testing, none of them have sound quality that is drastically above that of Lame 3.98b8 at the 128-192kbps bitrate range. That is just me though and I plan on switching to AAC once Nero's encoder comes out only if it is able to provide quality above what Lame 3.98b8 can offer at a certain setting.

Agree.

Ogg Vorbis aoTuV and several others are out there right now. Open sourced and with development teams all it needs is greater support.

Anything open source is welcomed.

However, as someone already stated, why? Most people find Lame MP3 to be transparent enough. Too many codecs, to many confusion. As long as MP3 is still universally supported, and the fact most people find Lame -V 5 to be transparent, it will be hard for anything to be the replacement of MP3. Even Ogg Vorbis is not gaining as much support as I wish it would.

yes, as true as most people find AAC to be transparent already at 96 kbit/s

Yes, Nero and Apple AAC has a good quality. But is there any commercial mp3 encoder better then LAME? The same way open source AAC encoder might have very high quality.


Nobody touched this item.

AAC Licensing Fees
(please refrain from discussing the death of musepack)

Open source community need not to support such format, given the fact that both transparent and "cellphone quality" bitrates are covered by other codecs.

I'm a bit confused about the negative responses here.

Nobody's suggesting a new format, and AAC is already about the second most widely supported format. Surely a FOSS implementation would be helpful for people doing FOSS audio work?

OTOH, would a truly FOSS implementation be possible? Aren't there patent problems?

Some have argued that FhG is better at some bitrates (and settings) than Lame with certain samples. So yeah, there are conditions where a commercial mp3 encoder is better than Lame. For the most part no, that is not the case. I think that FAAC had its chance but the ball was just dropped with it.



No, I am just stating that we first need to focus on what is out there before introducing yet another AAC encoder. We have the CoreAudio/Apple AAC encoder, Nero AAC, FAAC, Helix AAC, and a couple of others yet none of them are able to vastly improve beyond what Lame mp3 does in the 128kbps-192kbps bitrate range. So I think we should focus on improving what is already out there. Then, once AAC becomes more mature, introduce an open source AAC encoder with the potential to pass commercially developed encoders in quality. I think Nero is doing a pretty good job of enlisting the help of everyone here on HA though. Nero might be commercially developed but the developers remain frequent visitors here so it is almost as if they are open source.

I would expect so, for which I am sorely tempted to apologize.

You might as well complain about no open source WMA codec...

How can you compare? WMA is proprietary, AAC is standardized meaning the spec can be got (bought?). Also, don't confuse the term codec - WMA has an OSS decoder, but no an OSS encoder.

The only reason there is no OSS AAC encoder is because no-one has wanted to improve FAAC or start a new project. Most of the FAAC devs went to Nero, which resulted in the Nero codec people use today.

There is still a possibility that libavcodec would finally have an AAC encoder with a psymodel (GSOC)

Gabriel, it's kind of you to help Kostya with the aac codec. Hope that codec achieving good fruit eventually.

Um, there is a LOT of IP in any AAC encoder, tied to the bitstream, compression model, stereo coding model, stereo pair method, etc. A LOT.

This makes "open source" an "open source" of patent violations, I'd think. You can't make a bitstream without tripping over patents.

But other standards like MP3, MPEG 4 ASP and H.264 are also patented. And there is x264 (open source H.264 encoder) wich is the best encoder in quality terms at this moment.

That hasn't stopped LAME. A company that's willing to license the patent can use the source without issues (without the confines of the LGPL).

AAC and WMA are both proprietary, which is why you have to pay fees to use either. The spec is available for AAC though, which is nice.



There is an open source WMA encoder, though I doubt the quality is very good.



So basically, just like the situation with WMA . . .

Anyway I'm not sure what good an open source AAC encoder would really be. Nero and Apple are both pretty good, they're paying the license fees, and they're actually interested in improving support for the format. I don't think the open source community has much interest at all in AAC at the moment, so wishing for an open source encoder doesn't make much sense.

I wouldn't be so confident about your claims as Mainconcept is also doing a pretty good job. It is not free, though.

x264 has had a lot of recent psy enhancements which give it much better quality than Mainconcept or Ateme. If its SoC projects get a move on it should be even better.

AAC is not "proprietary". It is an ISO standard. Please keep the record straight.


An ISO standard can (and many do) require licensing patents to practice. This is the case for MPEG-AAC (any variety). Encoders are particularly tough in that there are patents that read on how to build the bitstream, etc.

You can evade some of them, but only by deliberately making the encoded quality worse.

Isn't FAAC open source?
FAAC can be high quality (not the best but high nowadays)

I can be made to pay to use it, thus it is literally proprietary. If there was no ownership, then I couldn't be made to pay.



I never said otherwise.

When one takes the literal meaning of proprietary, yes, AAC is a proprietary format. However, one normally regards a proprietary format as being tied to one company and one standard such as ATRAC3, WMA, RealAudio, etc. There are multiple AAC encoders on the market which normally means that the format isn't proprietary. It is true that a party privately owns certain copyrights on the AAC format thus making companies pay to make their own AAC encoder. With that regard though, one can look at mp3 as being proprietary as FhG owns many copyrights on that format. Hence, I wouldn't regard AAC as being proprietary anymore so than mp3 or OGG.

AT&T, Fraunhofer, Dolby and Sony all claim rights to PATENTS on parts of AAC.


Perhaps Lucent, unsure on that.

Not copyrights.

Literal in this case appears to be a synonym for "correct".



Perhaps you do, but this isn't correct, since assets can be jointly owned.



Given that there are multiple WMA encoders made by different groups, it would seem your definition contradicts your assertion that WMA is proprietary. Unless you're willing to argue that WMA isn't proprietary, perhaps you should reconsider your definition.



MP3 is proprietary regardless of how one looks at it.



Well sure you can argue anything, but doing so is often quite stupid.

In context, there is absolutely no difference between the "proprietariness" of AAC and MP3. Both require patent licenses to make both encoders and decoders legally.

Both are ISO Standards, and as such, the bitstream format and decoder are not in the slightest proprietary in any usual sense of the word, meaning that the design is privately held and secret. There is not a bit secret about the bitstream format or decoder. Nothing. Zero.

This is a simple, testable, vefifiable fact.

So what's your objection?

The point is that there are 2 different ways to look at this:

*the industry way:
The industry ecosystem consider every mpeg standard (as backed up by ISO and ITU) as open standards because they are internationally recognized as such by national standard bodies, and are not tied to a specific single company.

*the Free Software (with capital letters) way:
They consider anything covered by patents without any explicit royalty exemption as being proprietary.

Both camps will probably not agree anytime soon, but I don't think intentionally and explicitly ignoring the POV of the other camp will lead to anything else than a flame war.

Btw, this is related to patents, which cover technology usage license, not copyrights, which cover technology implementation.

I do think there is a need for a free aac encoder, and definitely a market too. I already need to encode lots of movie content to mp4 container, and having to rely on commercial encoders makes things a bit harder, as mp4 allows only aac audio. Nero and Quicktime are both annoying to get to work on linux and other platforms.

I don't use it for music though, and I don't see much demand for it there. But AVC&AAC in MP4 is rapidly becoming the standard to encode video in, as support for it is already quite widespread (ipods, psp, ps3, xbox360) so an open source encoder for the audio would definitely help.

Try handbrake. It's free and works really well. You may get it to work in linux to, nut I wouldn't know.

Personally I don't think a open source AAC encoder is needed as Nero's encoder is free, in development, easy to get, high quality, and command line based making it work with a large range of rippers and encoding software. Sure it maybe commercial but will a true open source competitor change anything to what we already have with Nero's encoder.

Also seeing that x264 is an open source Mpeg4 codec, what audio encoder is used with that?

Wrong, MP4 allows: MP3, MP2 and a few others. http://en.wikipedia.org/wiki/MP4#Data_streams

Nero give you a linux static binary, how is that annoying?

Ah yes of course, I was only thinking of multichannel audio. MP3 works fine with stereo, true.


Oh, I wasn't even aware of that. Damn I'm blind.

Because it's only an executable and there is no library, thus forcing you to parse the output before multiplexing it?
Because it's x86 only and thus can't be used on PPC or ARMs?
Because its license is obviously only granted to end users?

The Nero command line encoder is perfect if you want to rip some CD and convert the tracks to AAC, but if you consider other use cases, it's not a proper solution anymore.

If Winamp is using Lame as its mp3 encoder, it's not likely because of its sound quality vs FhG encoders, but rather because it saves them 2.5 bucks per encoder. Why couldn't this apply to AAC?

Anyone can buy the standard and learn how parse a bitstream and make a decoder. In ITU an encoder is also standardized in most cases. (but not always) These are non-proprietary standards, I don't have to go to a particular company to get a secret sauce that is the decoder. The information is out there in public.

This is a misuse of the word "proprietary". I've had this argument with both Richard and John before, and calling an open standard "proprietary" is simply attempting to play both with the langauge and people's heads.

As much as "information wants to be free", when it's all free, nobody will bother to make any good information. We're already seeing some effects of that right here in our own little field, aren't we, now?

Now, all of the ISO MPEG standards are non-proprietary. The owners of the IP in the standards have agreed to RAND (Reasonable and Non-Discriminatory) licence the IP.

The encoders, to some extent, are a different story, because they are not part of the standard (beyond the "reference model" for which we can agree is not very interesting), and there may be additional IP in the reference model that is not RAND.

In that sense, a particular ENCODER for AAC can be both proprietary and not free to use. So there is potential for a proprietary issue in AAC encoders. There can be a "secret sauce". You might have to get it from a particular vendor.


Copyright covers the code. We're all talking about patents here, indeed.

There is some activity http://codecs.multimedia.cx/

i often use nero , but you can easily use FAAC if you really want to...

This feature list says nothing about the quality. Video is no different than audio: lossy compression must be evaluated in blind comparison.

J7N: I replied to an off-topic post that went to the recycle bin.

I am very grateful that the Nero team has made their encoder available for use for free. The quality is very good.

But over the years, it seems every open source project rapidly develops into the best available. Lame, Xvid, Open Office and H 264 are all examples.

There is something about the collaborative nature of open source that makes it attractive to me as well. When a group of individuals devote their time and skills for the development of a tool to be used by all, without seeking financial benefit, this makes me want to use the results and help tune them if I can.

$.02