Fraunhofer has begun marketing HD-AAC, ISO MPEG-4 AAC-LC & SLS Options

[rpp3po]

Fraunhofer IIS just presented HD-AAC at the Consumer Electronics Show in Las Vegas. It's a hybrid format which can be played back as conventional AAC on legacy devices. HD-AAC capable devices also read the SLS layer and get a lossless reconstruction of the original signal.

More details here.

MPEG-4 SLS is not really news, but Fraunhofer finally seems to be pushing this out to a broader public now. Decoders are supposed to be ready within 12 weeks for literally all current (mobile and stationary) hardware platforms.

Seems to be a great product, I hope Apple picks this up!

It's probably not necessarily better than other hybrid solutions. But it would be nice if it got enough momentum to be implemented into the top line of mobile players and their ecosystems. For example, if you would be able buy HD-AAC in the iTMS and then let it copy just the AAC-LC part to your iPod.

This post has been edited by rpp3po: Jan 10 2010, 22:35

[Brent]

You know, if only for the ridiculous ease of use when managing my iPod. Hope Apple will use this!

[mixminus1]

For example, if you would be able buy HD-AAC in the iTMS and then let it copy just the AAC-LC part to your iPod.

...or just buy ALAC files and then have iTunes encode to AAC on-the-fly when you sync your iPod (like it already can do with the shuffle).

This would seem to be the simpler solution as it utilizes an existing format, allows the end user to choose the AAC bitrate, and should also be more efficient from a downloading and storage standpoint (assuming the HD-AAC files would be larger than ALAC due to their hybrid nature).

[/mnt]

Hopefully this might work better then the HD Mp3 format.

For example, if you would be able buy HD-AAC in the iTMS and then let it copy just the AAC-LC part to your iPod.

...or just buy ALAC files and then have iTunes encode to AAC on-the-fly when you sync your iPod (like it already can do with the shuffle).

Sadly it only transcodes with the shuffle, the feature would be very usefull for low storage iPods such as the current 8gb iPod Touch (a shameless rip-off from Apple) and iPhones; which would make it easy for Apple to sell their music in ALAC through iTMS. iTunes transcodes MP4 videos into a compatible format on the fly though.

[C.R.Helmrich]

This would seem to be the simpler solution as it utilizes an existing format, allows the end user to choose the AAC bitrate, and should also be more efficient from a downloading and storage standpoint (assuming the HD-AAC files would be larger than ALAC due to their hybrid nature).

I heard/read somewhere that the compression performance of HD-AAC is supposed to be on par with FLAC and similar dedicated lossless coders.

Chris

[Garf]

I heard/read somewhere that the compression performance of HD-AAC is supposed to be on par with FLAC and similar dedicated lossless coders.

Chris

If we don't include speed into the comparison. But yes, compression ratio is OK. One can wonder what the use of MPEG4 ALS is once you have SLS.

MPEG4 SLS (HD-AAC) will win if you need to transcode, because it only needs minimal processing to "slice" the stream to the bitrate you want, and the AAC or AAC+SLS layers are backwards compatible to LC-AAC.

[C.R.Helmrich]

One can wonder what the use of MPEG4 ALS is once you have SLS.

Good question

The more I think of it: with HD-AAC gaining in popularity, I think LossyWAV will have a very powerful contender.

Chris

[odyssey]

Garf: Will Nero release an encoder capable of utilizing this?

[andy o]

Will be interesting to see how this affects the audiophile landscape if it becomes popular. The DTS-HD Master Audio format (bluray) works very similarly, and the inability to reverse engineer and decode the MA extension properly on the PC has had many people paranoid that their sound is not "as good as it could be" because they're probably hearing the DTS "core". People started doubting that they were hearing decoded full DTS-HD MA even though the programs (commercial bluray players like TotalMedia Theatre and PowerDVD) were saying that they were decoding it.

[Sebastian Mares]

Garf: Will Nero release an encoder capable of utilizing this?

I doubt Garf can answer this question not being a Nero employee.

[odyssey]

Garf: Will Nero release an encoder capable of utilizing this?

I doubt Garf can answer this question not being a Nero employee.

Wasn't he previously? I don't remember too well Who is? (you maybe?)

[odyssey]

People started doubting that they were hearing decoded full DTS-HD MA even though the programs (commercial bluray players like TotalMedia Theatre and PowerDVD) were saying that they were decoding it.

So why would they need the HD, if they can't hear the difference anyway?

[KFal]

You know, if only for the ridiculous ease of use when managing my iPod. Hope Apple will use this!

MPEG4 SLS (HD-AAC) will win if you need to transcode, because it only needs minimal processing to "slice" the stream to the bitrate you want, and the AAC or AAC+SLS layers are backwards compatible to LC-AAC.

Yes! I would love such a solution where one only has to manage one file for both archiving and mobile listening purposes.

[Garf]

Wasn't he previously? I don't remember too well Who is? (you maybe?)

I quit Nero about 3 years ago. I think you want to ask muaddib here, or maybe he'll comment in this thread himself.

[SebastianG]

The more I think of it: with HD-AAC gaining in popularity, I think LossyWAV will have a very powerful contender.

I dunno... It has pros and cons. LossyWAV-like formats are fairly easy to decode. On the other hand LossyWAV's "base layer" is currently not nearly as efficient as AAC (and it never will be). But I think it could perform much better with a simple, good and conservative psychoacoustic model + dynamic noise shaper (I can even offer code for designing and using frequency-warped all-pole lattice filters as noise shaper).

Cheers,
SG

[odyssey]

I quit Nero about 3 years ago.

LOL, I wasn't wrong, just a bit outdated I'll try him, thanks

[mixminus1]

MPEG4 SLS (HD-AAC) will win if you need to transcode, because it only needs minimal processing to "slice" the stream to the bitrate you want, and the AAC or AAC+SLS layers are backwards compatible to LC-AAC.

Ah, didn't get that part from the block diagram in the Fraunhofer PDF.

So you can take an HD-AAC file, choose to "slice" it to 256 kb/sec or 128 kb/sec, and either of those resulting files would be playable by any device capable of LC-AAC playback?

[rpp3po]

Both the HD-AAC and the sliced LC-AAC file can be played back on a legacy device. The LC-AAC portion of the HD-AAC file is only encoded once, though. So you can only slice a new file out at the bitrate that was originally encoded, e. g. 256 kbit/s. For other bitrates you need to re-encode the lossless data.

This post has been edited by rpp3po: Jan 11 2010, 17:22

[Garf]

Both the HD-AAC and the sliced LC-AAC file can be played back on a legacy device. The LC-AAC portion of the HD-AAC file is only encoded once, though. So you can only slice a new file out at the bitrate that was originally encoded, e. g. 256 kbit/s. For other bitrates you need to re-encode the lossless data.

To clarify, you can only slice out a new file at a higher bitrate than what the LC-AAC part was originally encoded with. So applications of HD-AAC would generally encode the LC-AAC at something lower like 96 or 128kbps, rather than a higher bitrate.

If the LC-AAC part of the file was encoded at 128kbps, you can still slice out a 256kbps HD-AAC file, which would be 128kbps LC-AAC plus 128kbps correction layer. The correction layer is very efficient and the combination could reasonably be compared to a 256kbps LC-AAC file.

The resulting file will play back as a 256kbps file in a HD-AAC player and like 128kbps in a LC-AAC player.

I hope this made it clearer instead of the other way around

[mixminus1]

Yep, got it, thanks!

[rpp3po]

While everything you write is correct, I think we have used different meanings of "slicing out". There are three types of files:

1. HD-AAC with complete SLS content: lossless
2. HD-AAC with partial SLS content: high quality lossy.
3. Plain LC-AAC "sliced" or stripped from a HE-AAC file.

All three can be played back on legacy AAC players without modification. The latter will only access the embedded LC-AAC part, though. While being compatible that means wasting a lot of space, which one could use otherwise on those devices. So you can strip of all HD-content from an HD-AAC file and just upload the plain LC-AAC part to those devices.

Players with real HD-AAC support can play category 1 & 2 files with all features enabled. Category 2 files are a kind of a cross-breed. They are lossy but one can allocate an arbitrary amount of SLS data for higher quality. Up to 100% when they become category 1 files. So the user can choose exactly how much space he is willing to spend.

Personally, if I would be able to chose options myself, I would not go for a <= 128kbit/s LC-AAC base stream, but for about ~192kbit/s, That's my intended rate for mobile use and I don't expect older iPods ever to get updated for HD-AAC. Partial SLS content would be inaccessible for those. I don't expect category 2 to find broad acceptance anyway.

Request this feature here (apple.com) if you want HD-AAC support in iTunes! Apple really pays much attention to customer feedback.

This post has been edited by rpp3po: Jan 11 2010, 19:43

[rpp3po]

I think, if Fraunhofer wants this to be a success, they must have an unlicensed encoder and decoder out in the wild soon. I don't think that they can otherwise build enough momentum while the technology is still hot. The market is pretty saturated with formats. FLAC has already convinced some bigger labels. And the little extra-scalability-feature won't turn the tide of the world just because they are Fraunhofer and had success once.

Microsoft and Adobe have build monopolies on business facts like that. But I don't have much hope for Fraunhofer. Those institutions usually don't have the organizational structure, where a decision like that can really be made. Scientists are used to having to cover their asses consistently for every step into the unknown. And who would voluntarily take responsibility for a step like that. By contrast for real businesses like MS and Adobe taking risks is routine business.

This post has been edited by rpp3po: Jan 11 2010, 20:37

[Garf]

I think, if Fraunhofer wants this to be a success, they must have an unlicensed encoder and decoder out in the wild soon.

I'm not sure what you mean by unlicensed, but MPEG 4 SLS is an ISO standard and can be implemented by anyone. No need for Fraunhofer for that. They're just one firm that implemented it, and try to sell their codec which is named HD-AAC. Just like the HE-AAC standard was sold by Coding Technologies as aacPlus.

This standard is probably covered by a boatload of patents, and will for sure be covered by those for LC-AAC. Fraunhofer doesn't have any particular decision power about how they will be licensed.

[benski]

I think, if Fraunhofer wants this to be a success, they must have an unlicensed encoder and decoder out in the wild soon.

I'm not sure what you mean by unlicensed, but MPEG 4 SLS is an ISO standard and can be implemented by anyone. No need for Fraunhofer for that. They're just one firm that implemented it, and try to sell their codec which is named HD-AAC. Just like the HE-AAC standard was sold by Coding Technologies as aacPlus.

This standard is probably covered by a boatload of patents, and will for sure be covered by those for LC-AAC. Fraunhofer doesn't have any particular decision power about how they will be licensed.

http://www.vialicensing.com/licensing/SLS_fees.cfm

[saratoga]

Microsoft and Adobe have build monopolies on business facts like that. But I don't have much hope for Fraunhofer. Those institutions usually don't have the organizational structure, where a decision like that can really be made. Scientists are used to having to cover their asses consistently for every step into the unknown. And who would voluntarily take responsibility for a step like that. By contrast for real businesses like MS and Adobe taking risks is routine business.

I tend to agree. MS and Apple have very little reason to support something like this, so its hard to see it catching on.

Actually no one in the business of selling hardware and charging more for larger storage has much reason to hope scalable encoding succeeds. Much better to try and market people pure lossless formats and up-sell them on larger capacity, higher margin hardware.