[FFmpeg-devel] [PATCH] avcodec/dcaenc: Initial implementation of ADPCM encoding for DCA encoder

[Даниил Чередник]

Hi.

This patch introduces initial implementation of subband ADPCM encoding for
DCA codec.

Some results:

sample:

https://yadi.sk/d/B_3sVskM3HZiWK - original

https://yadi.sk/d/7CK47Nt63HZiWf - without adpcm

https://yadi.sk/d/25q1JDV93HZiWq - with adpcm

chirp tone:

https://yadi.sk/i/tZKHoJ1d3HZk4c

Right now this feature is disabled by default. But it is ready to try
using -dca_adpcm 1 option.

There are some issues, should be solved before enabling this feature by
default:

1. Speed up: I am trying to find best filter in each subband. But with real
signal, usually only few subbands has significant prediction gain. The idea
is try to analyze FFT spectrum (which is already calculated), to check is
particular subband looks like tonal or noise. If subband is noise like - do
not try to find best LPC predictor.

2. Modify psychoacoustic to use prediction gain for bit allocation. Right
now ADPCM encoded block can get some extra bits.

3. Tuning the prediction gain threshold.


Thank you.
-- 
Daniil Cherednik
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-avcodec-dcaenc-Initial-implementation-of-ADPCM-encod.patch
Type: application/octet-stream
Size: 34368 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20170503/0122e692/attachment.obj>