Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: What features would you like in a new lossless codec? (Read 9452 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

What features would you like in a new lossless codec?

What kind of features would you like to see in new lossless codec (or added to an existing codec)? Of course, apart from obvious desirable aspects like high compression ratio, short decoding time, cross-platform, etc.

I am looking for ideas like parallel processing support, whether it be CPU (ie. multi-core) or GPU (eg. CUDA), or on-the-go upsampling (PCM output at higher sampling rate and bit depth, minimizing cost compared to FIR solutions), etc. I'm looking for recent and novel features that would enhance the lossless experience.

I also consider different architecture schemes that would make files more practical. For example this could be along the lines of MP3HD, where residual data is stored in metadata. Pros: hardware already implemented, fast extraction of lossy files for portable use, seamless transition for average user. Cons: lower compression ratio, need for strict regulation, potentially easily hacked, etc.

So, what features would you like see in the future?

What features would you like in a new lossless codec?

Reply #1
on-the-go upsampling (PCM output at higher sampling rate and bit depth, minimizing cost compared to FIR solutions),


Is this actually possible?

What features would you like in a new lossless codec?

Reply #2
Cons: .... potentially easily hacked, etc.
Seems to imply a proprietary solution with a sprinkling of DRM.... Not for me, thanks.
lossyWAV -q X -a 4 -s h -A --feedback 2 --limit 15848 --scale 0.5 | FLAC -5 -e -p -b 512 -P=4096 -S- (having set foobar to output 24-bit PCM; scaling by 0.5 gives the ANS headroom to work)

What features would you like in a new lossless codec?

Reply #3
on-the-go upsampling (PCM output at higher sampling rate and bit depth, minimizing cost compared to FIR solutions),


Is this actually possible?

Maybe, maybe not. It depends on the internal decoding operations. You could insert filters into intermediate stages of decoding. Although with existing codecs this could be a real puzzle, with a new architecture designed from scratch, it may well be possible.

What features would you like in a new lossless codec?

Reply #4
Cons: .... potentially easily hacked, etc.
Seems to imply a proprietary solution with a sprinkling of DRM.... Not for me, thanks.

This is definitely not my intention. I was just brainstorming; any architecture idea is welcome. I'm mainly looking for features with the overall objective of practicality.

What features would you like in a new lossless codec?

Reply #5
on-the-go upsampling (PCM output at higher sampling rate and bit depth


In a proper, responsibility-driven design these features don't belong into a codec but can be implemented much more elegantly further down the pipeline. Else stuff like sample accurate searching calls for a somewhat messy implementation in contrast to a cleanly laid out pipeline.

Windows and Mac OS X can convert output samples to higher bit depth without codec assistance already, Linux probably, too.

I also don't see any benefit of doing resampling within a codec in contrast to further down the pipeline (player plugin or OS).

What features would you like in a new lossless codec?

Reply #6
on-the-go upsampling (PCM output at higher sampling rate and bit depth, minimizing cost compared to FIR solutions),


Is this actually possible?

Maybe, maybe not. It depends on the internal decoding operations. You could insert filters into intermediate stages of decoding.


Heh.  Well if we don't have to be constrained by actually being possible, can I ask for it to undo poor mastering ?

What features would you like in a new lossless codec?

Reply #7
on-the-go upsampling (PCM output at higher sampling rate and bit depth


In a proper, responsibility-driven design these features don't belong into a codec but can be implemented much more elegantly further down the pipeline.

Like i said, i'm just throwing ideas here. Nothing has been evaluated.

What features would you like in a new lossless codec?

Reply #8
I am looking for ideas like parallel processing support, whether it be CPU (ie. multi-core) or GPU (eg. CUDA), or on-the-go upsampling (PCM output at higher sampling rate and bit depth, minimizing cost compared to FIR solutions), etc. I'm looking for recent and novel features that would enhance the lossless experience.


Do you have the intention to create the ultimate "audiophile" lossless codec? I also don´t think up-sampling belongs to a lossless codec at all.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

What features would you like in a new lossless codec?

Reply #9
I'd like to see difference files against arbitrary AAC files, that when combined represent a lossless version of the original, where ('AAC size' + 'diff size') * 1.05 <= 'comparable FLAC size'. Compressing a regular difference with FLAC doesn't work, btw. You'd get 'AAC size' + a diff of roughly 0.9 * 'comparable FLAC size'. You have to tap in deeper for anything remotely useful.

What features would you like in a new lossless codec?

Reply #10
I am looking for ideas like parallel processing support, whether it be CPU (ie. multi-core) or GPU (eg. CUDA), or on-the-go upsampling (PCM output at higher sampling rate and bit depth, minimizing cost compared to FIR solutions), etc. I'm looking for recent and novel features that would enhance the lossless experience.


Do you have the intention to create the ultimate "audiophile" lossless codec? I also don´t think up-sampling belongs to a lossless codec at all.

Noted.

What features would you like in a new lossless codec?

Reply #11
I'd difference files against arbitrary AAC files, that when combined represent a lossless version of the original, where ('AAC size' + 'diff size') * 1.05 <= 'comparable FLAC size'. Compressing a regular difference with FLAC doesn't work, btw. You'd get 'AAC size' + a diff of roughly 0.9 * 'comparable FLAC size'. You have to tap in deeper for anything remotely useful.


That would be MPEG SLS (aka AAC-HD).  Unfortunately, basing on AAC + difference signal means that lossless compression is not particularly efficient, so the format hasn't caught all that much traction. 

What features would you like in a new lossless codec?

Reply #12
AFAIK, MPEG SLS means singular files that take up a lot of space on your flash based device. Personally I never saw a use for that.

What features would you like in a new lossless codec?

Reply #13
AFAIK, MPEG SLS means singular files that take up a lot of space on your flash based device. Personally I never saw a use for that.


Yeah, thats sort of the problem with the whole idea.  You get to choose between efficient lossy compression or efficient lossless compression in a hybrid scheme.  It doesn't seem possible to have good efficiency in both which makes the whole thing pretty much useless.

What features would you like in a new lossless codec?

Reply #14
Well, depends on what you call efficient. There's a scientific article on HD-AAC. In pure lossless mode, i.e. no lossy core, it's at least as efficient as e.g. FLAC at its highest setting (measured that myself). With a 128-kb AAC core, I think it's in the range of FLAC -2 or -3. Not perfect, but still much better than PCM Wave or zip or Shorten. And I don't think you can get better than that given that you need some redundant side information in order to be able to play the lossy AAC core separately. So to me, it's quite efficient.

Chris
If I don't reply to your reply, it means I agree with you.

What features would you like in a new lossless codec?

Reply #15
What about tweaking?


What features would you like in a new lossless codec?

Reply #17
Yeah, thats sort of the problem with the whole idea.  You get to choose between efficient lossy compression or efficient lossless compression in a hybrid scheme.  It doesn't seem possible to have good efficiency in both which makes the whole thing pretty much useless.


I meant having two files, diff and aac, is preferable over one large file (MPEG SLS).

What features would you like in a new lossless codec?

Reply #18
Id like to have a mastered lossy "core", and a lossless diff against pre-mastering (or at least "audiophile" mastering) :-)

-k

What features would you like in a new lossless codec?

Reply #19
1. Fast error detection (that is: checksumming compressed stream).
2. Error correction (it's on ofr TODO. Too bad ofr seems dead...)
3. More flexible container. For example I could see use for a video in metadata. Or multiple audio streams. Or a program. They all can be found on CDs and DVD-As....Actually I thought about creating a codec that uses a FreeArc or similar container. Think of opening the file with a regular archiver and decompressing a wav file directly from it. Or adding some metadata.
There might be troubles with streaming support though...

What features would you like in a new lossless codec?

Reply #20
Id like to have a mastered lossy "core", and a lossless diff against pre-mastering (or at least "audiophile" mastering) :-)

-k

You mean something like this in one file?

Chris
If I don't reply to your reply, it means I agree with you.

What features would you like in a new lossless codec?

Reply #21
1. Fast error detection (that is: checksumming compressed stream).
2. Error correction (it's on ofr TODO. Too bad ofr seems dead...)
3. More flexible container. For example I could see use for a video in metadata. Or multiple audio streams. Or a program. They all can be found on CDs and DVD-As....Actually I thought about creating a codec that uses a FreeArc or similar container. Think of opening the file with a regular archiver and decompressing a wav file directly from it. Or adding some metadata.
There might be troubles with streaming support though...


Re: 3 surely you just want a container then, codecs shouldn't be containers. Simple metadata is fine, something like video is not.

What features would you like in a new lossless codec?

Reply #22
Id like to have a mastered lossy "core", and a lossless diff against pre-mastering (or at least "audiophile" mastering) :-)

-k

You mean something like this in one file?

Chris

Yeah. Does not really matter that they are in the same file, it was just an opportune point to repeat my wish =)

But perhaps some kind of meta-data-approach would make it practically easier to satisfy both music-lovers and radio stations.


Reminds me of groceries where I live. Buying a piece of meat might cost some sum. Buying what appears to be the same kind of meat only with added spices costs less. Perplexing that adding more effort lowers the price...

-k

What features would you like in a new lossless codec?

Reply #23
1. Fast error detection (that is: checksumming compressed stream).
2. Error correction (it's on ofr TODO. Too bad ofr seems dead...)
3. More flexible container. For example I could see use for a video in metadata. Or multiple audio streams. Or a program. They all can be found on CDs and DVD-As....Actually I thought about creating a codec that uses a FreeArc or similar container. Think of opening the file with a regular archiver and decompressing a wav file directly from it. Or adding some metadata.
There might be troubles with streaming support though...


Re: 3 surely you just want a container then, codecs shouldn't be containers. Simple metadata is fine, something like video is not.


Oh..yes, you're right, I mixed codec with a file format. All my comments apply to container actually...