An earlier revision of the prediction patch used the calc_is_encoding_err()
to check whether the predictor state is valid for both channels.
It _might_ be reused in the future as it is a generally useful function
able to tell the compatibility of channels in a pair.
I would hold off on applying this for now, though I'll keep it in mind.
On 22 August 2015 at 02:41, Ganesh Ajjanagadde <gajja...@mit.edu> wrote:
> On Fri, Aug 21, 2015 at 9:21 PM, Timothy Gu <timothyg...@gmail.com> wrote:
> > It's not used anywhere else.
> > ---
> > libavcodec/aacenc_is.c | 10 +++++++++-
> > libavcodec/aacenc_is.h | 11 -----------
> > 2 files changed, 9 insertions(+), 12 deletions(-)
> >
> > diff --git a/libavcodec/aacenc_is.c b/libavcodec/aacenc_is.c
> > index 4ae1054..2a0ed86 100644
> > --- a/libavcodec/aacenc_is.c
> > +++ b/libavcodec/aacenc_is.c
> > @@ -30,7 +30,15 @@
> > #include "aacenc_is.h"
> > #include "aacenc_quantization.h"
> >
> > -struct is_error calc_is_encoding_err(AACEncContext *s, ChannelElement
> *cpe,
> > +struct is_error {
> > + int pass; /* 1 if dist2 <= dist1 */
> > + int phase; /* -1 or +1 */
> > + float error; /* fabs(dist1 - dist2) */
> > + float dist1; /* From original coeffs */
> > + float dist2; /* From IS'd coeffs */
> > +};
> > +
> > +static struct is_error calc_is_encoding_err(AACEncContext *s,
> ChannelElement *cpe,
> > int start, int w, int g, float
> ener0,
> > float ener1, float ener01, int
> phase)
> > {
> > diff --git a/libavcodec/aacenc_is.h b/libavcodec/aacenc_is.h
> > index abb9b32..e213ef0 100644
> > --- a/libavcodec/aacenc_is.h
> > +++ b/libavcodec/aacenc_is.h
> > @@ -31,17 +31,6 @@
> > /** Frequency in Hz for lower limit of intensity stereo **/
> > #define INT_STEREO_LOW_LIMIT 6100
> >
> > -struct is_error {
> > - int pass; /* 1 if dist2 <= dist1 */
> > - int phase; /* -1 or +1 */
> > - float error; /* fabs(dist1 - dist2) */
> > - float dist1; /* From original coeffs */
> > - float dist2; /* From IS'd coeffs */
> > -};
> > -
> > -struct is_error calc_is_encoding_err(AACEncContext *s, ChannelElement
> *cpe,
> > - int start, int g, int w, float
> ener0,
> > - float ener1, float ener01, int
> phase);
> > void search_for_is(AACEncContext *s, AVCodecContext *avctx,
> ChannelElement *cpe);
> >
> > #endif /* AVCODEC_AACENC_INTENSITY_STEREO_H */
> > --
> > 1.9.1
>
> Patch itself LGTM.
> However, this was added just today, so these may be exposed very soon.
> Thus I would hold off for some time until the aac stuff is ironed out.
> If no replies come within next few days, go ahead with this.
>
> >
> > _______________________________________________
> > ffmpeg-devel mailing list
> > ffmpeg-devel@ffmpeg.org
> > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
