On Sat, Jan 7, 2012 at 12:08 PM, Ronald S. Bultje <[email protected]> wrote: > > Hi, > > On Fri, Jan 6, 2012 at 3:10 PM, Alex Converse <[email protected]> wrote: > > On Thu, Jan 5, 2012 at 4:33 PM, Alex Converse <[email protected]> > > wrote: > >> This keeps internal state variable accurate for LFEs and any other > >> elements which lack SBR data payloads. > >> --- > >> libavcodec/aacsbr.c | 2 ++ > >> 1 files changed, 2 insertions(+), 0 deletions(-) > >> > >> diff --git a/libavcodec/aacsbr.c b/libavcodec/aacsbr.c > >> index 81b0b4c..d38debc 100644 > >> --- a/libavcodec/aacsbr.c > >> +++ b/libavcodec/aacsbr.c > >> @@ -129,6 +129,8 @@ av_cold void ff_aac_sbr_init(void) > >> av_cold void ff_aac_sbr_ctx_init(AACContext *ac, SpectralBandReplication > >> *sbr) > >> { > >> float mdct_scale; > >> + sbr->sample_rate = 2 * ac->m4ac.sample_rate; //TODO use the nominal > >> sample rate for arbitrary sample rate support > >> + ac->m4ac.ext_sample_rate = 2 * ac->m4ac.sample_rate; > > > > This breaks HE-AAC in ADTS {fate-aac-ct-sbr_i-ps_i.aac} because > > m4ac.sample_rate is set by compatibility for > > headerless-AAC-sans-extradata. > > You mean s/breaks/changes/, where changes really means "improves"? Or > did I misunderstand?
I mean makes worse not better. We support a large number of hackey out of spec modes that don't always play nice together. _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
