On 05.09.2006, at 22:51, Josh Coalson wrote:> --- taktik <taktik@renoise.com> wrote: >> What exactly is the reason why libFlac and/or the FLAC command-line >> en/decoder currently doesn't support 32bit integer samples? >> >> The comment in "FLAC/format.h" -> of >> FLAC__REFERENCE_CODEC_MAX_BITS_PER_SAMPLE notes that >> the FLAC format supports 32bit, but the reference en/decoders >> currently do not. Does >> this mean that libFlac or just the commandline tool is limited to 24 >> bit? >> >> We would love to use FLAC as a total replacement for WAV files in our >> >> product (in exported song >> and instrument content), but really need to have 32bit support, as >> our internal sample format >> is 32 bit. >> I already hackily tried to change >> FLAC__REFERENCE_CODEC_MAX_BITS_PER_SAMPLE to 32u and ran >> some tests using libFlac. All seemed to work well, so this limitation >> is maybe (no longer?) >> needed? > > if this works for you, then you are lucky! the missing parts > are 32-bit handling in flac itself (parsing PCM samples), and > tuning for 32-bit in libFLAC. otherwise the current libFLAC > should handle 32-bit but I don't have any good material to > tune against.Great. Assuming I would send you a patch that integrates the 32bit pcm im/ export in the commandline tool, adjust libFlac to handle 32bit int samples, and add some test for all this: Could this then be integrated into the official release? Greets, eduard -------------- next part -------------- A non-text attachment was scrubbed... Name: PGP.sig Type: application/pgp-signature Size: 186 bytes Desc: This is a digitally signed message part Url : http://lists.xiph.org/pipermail/flac-dev/attachments/20060906/12da1fca/PGP.pgp
--- taktik <taktik@renoise.com> wrote:> > On 05.09.2006, at 22:51, Josh Coalson wrote: > > > --- taktik <taktik@renoise.com> wrote: > >> What exactly is the reason why libFlac and/or the FLAC > command-line > >> en/decoder currently doesn't support 32bit integer samples? > >> > >> The comment in "FLAC/format.h" -> of > >> FLAC__REFERENCE_CODEC_MAX_BITS_PER_SAMPLE notes that > >> the FLAC format supports 32bit, but the reference en/decoders > >> currently do not. Does > >> this mean that libFlac or just the commandline tool is limited to > 24 > >> bit? > >> > >> We would love to use FLAC as a total replacement for WAV files in > our > >> > >> product (in exported song > >> and instrument content), but really need to have 32bit support, as > >> our internal sample format > >> is 32 bit. > >> I already hackily tried to change > >> FLAC__REFERENCE_CODEC_MAX_BITS_PER_SAMPLE to 32u and ran > >> some tests using libFlac. All seemed to work well, so this > limitation > >> is maybe (no longer?) > >> needed? > > > > if this works for you, then you are lucky! the missing parts > > are 32-bit handling in flac itself (parsing PCM samples), and > > tuning for 32-bit in libFLAC. otherwise the current libFLAC > > should handle 32-bit but I don't have any good material to > > tune against. > > > Great. > > Assuming I would send you a patch that integrates the 32bit pcm im/ > export in the > commandline tool, adjust libFlac to handle 32bit int samples, and add > > some test > for all this: Could this then be integrated into the official > release?yes, but I would say hold off until the next release because the number of changes in src/flac/encode.c since 1.1.2 will make it hard to patch. next release should be sometime this month. Josh __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com