I've read on the FLAC webpage the description of the format. While very interesting, it doesn't really go into the mathematics behind it.

At a given point, A. J. Robinson's work on Shorten is acknowledged and it is stated that his paper is a good starting point to understand the basic workings of FLAC. There is a link to what would seem to be his paper on the subject, but the link is outdated/broken.

So... I want to understand the mathematical details of the algorithm used by FLAC, but I haven't found any good online links and/or explanations.

If anybody can explain or point to an explanation (preferably online) I would be grateful. I'm interested in a technical discussion so long as it is complete and thorough -- i.e. something that is not extremely technical at first but eventually get's down and "dirty" with the details later would be my preference.

Since I have a Ph.D. in physics, I'm used to mathematics. I have some understanding on the subject of DSP (I did a master's thesis involving the use of autocorrelation to find weak periodicities in very noisy signals, for example, though this is not my only source of DSP knowledge), but I really am not current or an expert on the subject. I do catch on rather fast, though.

And quite frankly, I would actually prefer if a technical discussion would follow this post, as a good set of explanations is always better than the best written article/book.

Well, I thank anybody in advance who can help me out.

yeah, that paper moves around a lot; I managed to dig it out of the wayback machine:
http://web.archive.org/web/20040531020607/...nson_tr156.ps.Z

Josh

Thanks!

I'll give it a good read.

I am planning to impliment FLAC on FPGA platform, can some provide me with technical details of immplimentation, link mentioned above I was not able to open...

Thank you

Would that be for encoding or decoding or both?

The paper can be found in any of the shorten source distributions, and is available in both PostScript and LaTeX formats. You can grab them here.