this will find the coefficients which minimize the sum of the squared errors,
levinson-durbin recursion OTOH is only strictly correct if the autocorrelation matrix is a
toeplitz matrix which it is only if the blocksize is infinite, this is also why applying
a window (like the welch winodw we currently use) improves the lpc coefficients generated
by levinson-durbin recursion ...
optionally (use_lpc>2) support iterative linear least abs() solver using cholesky
factorization with adjusted weights in each iteration
compression gain for both is small, and multiple passes are of course dead slow
Originally committed as revision 5747 to svn://svn.ffmpeg.org/ffmpeg/trunk
1) search for optimal rice parameters and partition order. i also
modified the stereo method estimation to use this to calculate estimated
bit count instead of using just the pure sums.
2) search for the best fixed prediction order
3) constant subframe mode (good for encoding silence)
Note that the regression test for the decoded wav file also changed.
This is due to FFmpeg's FLAC decoder truncating the file, which it did
before anyway...just at a different cutoff point. The generated FLAC
files are still 100% lossless.
With this update, FFmpeg's FLAC encoder has speed and compression
somewhere between "flac -1" and "flac -2". On my machine, it's about
15% faster than "flac -2", and about 10% slower than "flac -1". The
encoding parameters are identical to "flac -2" (fixed predictors, 1152
blocksize, partition order 0 to 3).
Originally committed as revision 5536 to svn://svn.ffmpeg.org/ffmpeg/trunk
