@@ -608,7 +608,8 @@ is typically encoded in the 4 bit `block size' field. But for odd-sized frames - which often occur at the end of the stream - that value is stored as an 8 or 16 bit integer following UTF-8 encoded frame number. -Less commonly, odd sample rate values are stored as 8 bit (in kHz), +.PP +In addition, odd sample rate values are stored as 8 bit (in kHz), 16 bit (in Hz) or 16 bit (in 10s of Hz) prior to the CRC-8, should a predefined value not be available. .PP @@ -662,5 +663,49 @@ This subframe simply stores all the samples as-is, with no compression whatsoever. It is a `fallback' encoding method for when no other subframe makes one's data any smaller. - - +.SUBSUBSECTION "the FIXED subframe" +.PP +This subframe consists of `predictor order' number of unencoded +warm-up samples followed by a residual. +Determing which predictor order to use on a given set of input samples +depends on their minimum delta sum. +This process is best explained by example: +.TS +tab(:); +| c | c | c | c | c | c | c | +| c | r | r | r | r | r | r |. +_ +index:sample:@{\[*D] sup 0}@:@{\[*D] sup 1}@:@{\[*D] sup 2}@:@{\[*D] sup 3}@:@{\[*D] sup 4}@ += +0:-40: +1:-41:\fI-41\fR +2:-40:\fI-40\fR:\fI-1\fR +3:-39:\fI-39\fR:\fI-1\fR:\fI0\fR +4:-38:-38:\fI-1\fR:\fI0\fR:\fI0\fR +5:-38:-38:0:-1:1:-1 +6:-35:-35:-3:3:-4:5 +7:-35:-35:0:-3:6:-10 +8:-39:-39:4:-4:1:5 +9:-40:-40:1:3:-7:8 +10:-40:-40:0:1:2:-9 +11:-39:-39:-1:1:0:2 +12:-38:-38:-1:0:1:-1 +13:-37:-37:-1:0:0:1 +14:-33:-33:-4:3:-3:3 +15:-36:-36:3:-7:10:-13 +16:-35:-35:-1:4:-11:21 +17:-31:-31:-4:3:1:-12 +18:-32:-32:1:-5:8:-7 +19:-33:-33:1:0:-5:13 +_ +@|~{roman "sum"}~|@::579:26:38:60:111 +_ +.TE +.LP +Note that the numbers in italics play a part in the delta +calculation to their right, but do \fBnot\fR figure into the +delta's absolute value sum, below. +.PP +In this example, @\[*D] sup 1@'s value of 26 is the smallest. +Therefore, when compressing this set of samples in a FIXED subframe, +it is best to use a predictor order of 1. docs/codec_flac.ms 4ae5feba76b065715bf17cb179186398c01a539d Brian Langenberger bjl@usa.net http://github.com/tuffy/python-audio-tools/commit/51efce5f561cf4cdf5fcff3367849d7a8c6fd50d 51efce5f561cf4cdf5fcff3367849d7a8c6fd50d 2008-05-09T18:54:09-07:00 2008-05-09T18:54:09-07:00 Added an explanation and example for choosing the best FIXED subframe predictor order as part of the FLAC encoder documentation. f824d11f230bf8143e2de2a31fbd3b8d09de0543 Brian Langenberger bjl@usa.net