Re: [FFmpeg-devel] [PATCH V3 1/2] avcodec/vorbisenc: Add pre-echo detection

2017-07-28 Thread James Almer
On 7/28/2017 11:59 AM, Tyler Jones wrote: >>> --- a/libavcodec/vorbisenc.c >>> +++ b/libavcodec/vorbisenc.c >>> @@ -33,6 +33,7 @@ >>> #include "mathops.h" >>> #include "vorbis.h" >>> #include "vorbis_enc_data.h" >>> +#include "vorbispsy.h" >>> >>> #include "audio_frame_queue.h" >>> #include

Re: [FFmpeg-devel] [PATCH V3 1/2] avcodec/vorbisenc: Add pre-echo detection

2017-07-28 Thread Tyler Jones
> > --- a/libavcodec/vorbisenc.c > > +++ b/libavcodec/vorbisenc.c > > @@ -33,6 +33,7 @@ > > #include "mathops.h" > > #include "vorbis.h" > > #include "vorbis_enc_data.h" > > +#include "vorbispsy.h" > > > > #include "audio_frame_queue.h" > > #include "libavfilter/bufferqueue.h" > > @@ -136,6

Re: [FFmpeg-devel] [PATCH V3 1/2] avcodec/vorbisenc: Add pre-echo detection

2017-07-27 Thread James Almer
On 7/27/2017 6:22 PM, Tyler Jones wrote: > The encoder will attempt to determine the existence of transient > signals by applying a 4th order highpass filter to remove dominant > low frequency waveforms. Frames are then split up into blocks > where the variance is calculated and compared with block

[FFmpeg-devel] [PATCH V3 1/2] avcodec/vorbisenc: Add pre-echo detection

2017-07-27 Thread Tyler Jones
The encoder will attempt to determine the existence of transient signals by applying a 4th order highpass filter to remove dominant low frequency waveforms. Frames are then split up into blocks where the variance is calculated and compared with blocks from the previous frame. A preecho is only like