Hi. I am glad to inform everyone here that support for decoding using libcelt was just pushed to ffmpeg. It raised a few API questions or issues: * At least while it's experimental, having a constant in the library for the supported bitstream version would probably be useful. Something like that comes to mind: extern celt_int32 celt_bitstream_version; or, even better: int celt_bitstream_version_supported(celt_int32 version); Since dynamic linking is supported, a build-time constant is not suitable. Currently, I resorted to have libcelt generate a dummy header to get the version_id field. * The header_size is set to 56, while the header is actually 60 bytes. Is it normal? * It seems that the celt_*_destroy all accept NULL without choking, just like libc's free. If this is on purpose, could it be stated in the documentation? Regards, -- Nicolas George -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 198 bytes Desc: Digital signature Url : http://lists.xiph.org/pipermail/opus/attachments/20110420/4265874d/attachment-0002.pgp
On Wed, Apr 20, 2011 at 8:10 AM, Nicolas George <nicolas.george at normalesup.org> wrote:> Hi. > > I am glad to inform everyone here that support for decoding using libcelt > was just pushed to ffmpeg. > > It raised a few API questions or issues: > > * At least while it's experimental, having a constant in the library for the > ?supported bitstream version would probably be useful. Something like that > ?comes to mind: > > ?extern celt_int32 celt_bitstream_version; > > ?or, even better: > > ?int celt_bitstream_version_supported(celt_int32 version);We removed that functionality in ff96b165feadf5c0bb5c75c3013d3447bd57298a on march 21st as we no longer intend to change the CELT bitstream itself (the Ogg encapsulation is another matter).
Apparently Analagous Threads
- [PATCH] Fix CNG when effEBands is less than nbEBands.
- [PATCH] Eliminate the ec_int32 and ec_uint32 typedefs.
- [PATCH] Pitch now quantised at the band level, got rid of all the VQ code.
- [PATCH] Moved the application of the pitch gain to (un)quant_bands(). This doesn't change anything to the codec, but will make the next changes easier.
- V0.8.0 Problems