Hey guys!

I'm a CSU student doing a research paper on "Codecs for Streaming HD Audio" for my Audio Recording class. The research has turned out to be very tricky. I think I've found all of the lossless audio codecs that support streaming and high fidelity (see below and correct me if I'm wrong), but I have had a difficult time finding accurate, objective sources for this topic. I just fired off an email to Apple's media help department regarding ALAC. Do any of you know of any books or reliable websites that might have this sort of information? My professor said not to rely on Wikipedia too much. Thank you all tremendously!

Casey


For the line-up I have:
ALAC, FLAC, MPEG4 ALS, OptimFrog, RealAudio, WavPack, and WMA.

What is your definition of "High Definition" audio? Anything that is lossless and can be streamed? I'm not sure that is the common definition.

When I say high definition, I am talking about fidelity of the source audio. For this purpose, high definition = better-than-CD quality, ideally 24-bit/96KHz/multichannel.

I guess I should add TAK to the bunch. I see it was added to the comparison table.

client side at least:
a. you can't do that with flash
b. apple docs are waste of time

i would try to research
a. the wma part of the story
b. java players might just support that kind of thing?

(I'm thinking about web embedded methods here only)

p.s.
with some standalone clients it should be easy via http, for example, try

(which is 24bit stereo flac, i assume it should work with 5.1 streams as well)

Well that depends on what the purpose of the project is and it's applications? Do you want to know how they work from an engineering perspective or is this strictly multimeda driven? There are plenty of engineering papers out there you can skim through to get a conceptual understanding of the topic. You might want to start with Information Theory and work your way around. In terms of multmedia I don't understand what you are looking for? "High Definition Audio" is a rather widespread term that can mean anything. The first thing that comes to mind is radio not lossless data compression. You might want to just say "lossless audio".

Information Theory:

http://mit.edu/6.933/www/Fall2001/Shannon2.pdf
http://www.inference.phy.cam.ac.uk/mackay/...ory/course.html

Linear Prediction:

http://caltechbook.library.caltech.edu/227/

Golumb-Rice Coding:

http://urchin.earth.li/~twic/Golombs_Original_Paper/
http://flac.sourceforge.net/documentation_...t_overview.html

Obviously this is not an exhaustive list, but I am sure you can find more information out there for your paper.

Thank you for those links. Whoa!

The audience for this paper will be my instructor and my classmates. So it should be informative, but probably not overly coded, although I may have to resort to discussing the programming of compression and decompression methods if I cannot find enough information. I was hoping to find more general information about the codecs, such as features/capabilities, advantages and disadvantages. I suppose I could try to interpret the programming into language that is easier to understand. But since some of these codecs are commercial, non-open source, unless someone has reverse engineered the process and has documentation, I won't know how they work, right?

Do you know if they all use the same methods to accomplish compression/decompression? Are Linear Prediction and Golumb-Rice standard techniques?

Thanks again for your input and say hello to Mass for me.

I think if you are looking for a general outline this might work for you:

http://wiki.hydrogenaudio.org/index.php?ti...less_comparison

The three areas I mentioned are common to all lossless compression techniques, although albeit somewhat more complicated you will need to consult the documentation for some. The first one is the theoretical foundation for which compression is done. The second two are subsets of Information Theory. FLAC for instance uses 0-32nd order LPC predictors that's where linear prediction comes in the mathmatical transform step (technically it's called adaptive FIR filtering). It then uses a technique called Golomb-Rice Coding which is a second stage coding technique known as "entropy coding" in order to save bits. This same technique is used by Shorten, FLAC, Apple Lossless, and MPEG-4 ALS. MPEG-4 ALS is very similiar to FLAC. The other codecs I am not as sure for instance I am not familiar with how Wavpack hybrid lossy mode works, but you are only focusing on lossless codecs so that's beyond the scope of this discussion. Boston gives you a warm welcome also!