| Branch: | Revision:

ffmpeg / tests / ref / acodec @ 7e0a284b

# Date Author Comment
7e0a284b 03/08/2011 06:44 PM Justin Ruggles

ac3enc: shift coefficients to 24-bit following MDCT rather than using an exponent offset.

This makes channel coupling more accurate, increasing quality for stereo
content. It also simplifies exponent extraction and mantissa quantization
by no longer needing to apply an offset to the exponents....

6159bd0d 02/18/2011 06:52 PM Justin Ruggles

ac3enc: fix bug in stereo rematrixing decision.

The rematrixing strategy reuse flags are not reset between frames, so they
need to be initialized for all blocks, not just block 0.

Signed-off-by: Mans Rullgard <>
(cherry picked from commit 5b54d4b376fbd2c291b8a6dba7d425e28f654bcf)

660ab0e6 02/16/2011 10:00 PM Justin Ruggles

ac3enc: change default floor code to 7.

This is to match the value in every (E-)AC-3 file from commercial sources.
It has a negligible effect on audio quality.

Signed-off-by: Mans Rullgard <>
(cherry picked from commit 50d7140441b61809ad4ba4ea7b4ad8fe77bc6888)

10030226 02/04/2011 02:08 AM Justin Ruggles

ac3enc: Change EXP_DIFF_THRESHOLD to 500.

This patch changes the exponent difference threshold in the exponent
strategy decision function of the AC-3 encoder. I tested lowering in
increments of 100. From 1000 down to 500 generally increased in quality
with each step, but 400 was generally much worse....

dc7e07ac 01/08/2011 11:21 PM Justin Ruggles

Add stereo rematrixing support to the AC-3 encoders.
This improves the audio quality significantly for stereo source with both the
fixed-point and floating-point AC-3 encoders.
Update acodec-ac3_fixed and seek-ac3_rm test references.

Originally committed as revision 26271 to svn://

6fd96d1a 01/04/2011 11:53 AM Justin Ruggles

Change the AC-3 encoder to use floating-point.
Fixed-point AC-3 encoder renamed to ac3_fixed.
Regression test acodec-ac3 renamed to acodec-ac3_fixed.
Regression test lavf-rm changed to use ac3_fixed encoder.

Originally committed as revision 26209 to svn://

ec44dd5f 12/29/2010 07:17 PM Justin Ruggles

Change the default dB-per-bit code from 2 to 3.
This gives slightly better quality in PEAQ tests.
Code 3 gives a dBpb value of 2816 = -132dB (128 psd units = -6dB), which
corresponds to 22 bits. Since the exponents have an offset applied, the
16-bit source looks like 24-bit source to the bit allocation routine....

295ab2af 12/21/2010 09:18 PM Justin Ruggles

Change FIX15 back to clipping to -32767..32767.
This avoids a 16-bit overflow in mdct512() due to a -32768 value in costab.
References updated for acodec-ac3, lavf-rm, and seek-ac3_rm tests.
Thanks to Måns Rullgård for finding the bug.

Originally committed as revision 26071 to svn://

918cd225 12/14/2010 02:51 PM Justin Ruggles

Simplify fix15().
Turn it into 2 macros, and use av_clip_int16() and lrintf().
This matches the int16 to float sample conversion in audioconvert.c.
The regression test output is different due to lrintf() rounding.

Originally committed as revision 25956 to svn://

c7d89948 09/11/2010 07:52 PM Justin Ruggles

Set a constant frame size for encoding G.726 audio.

Originally committed as revision 25107 to svn://

c43d77c1 07/09/2010 04:06 PM Måns Rullgård

tiny_psnr: skip wav headers on input files

The byte count printed excludes the header, and offsets are applied
after the the headers are skipped.

Reference files updated to reflect new output. Some stddev/psnr values
have changed slightly due to headers no longer being compared....

cb0067ec 07/09/2010 12:40 AM Vitor Sessak

tiny_psnr: print max absolute difference between files

Regression test reference updates are due to the extra output
from tiny_psnr.

Patch by Vitor Sessak

Originally committed as revision 24132 to svn://

e7ddafd5 06/04/2010 10:41 PM David Conrad

matroskaenc: Don't write a second seekhead for the clusters; mkvalidate agrees
with me that it's unnecessary.

Originally committed as revision 23478 to svn://

ac9baa71 06/04/2010 10:40 PM James Zern

matroskaenc: Mux clusters better

Start them on keyframes when reasonable, and delay writing audio packets
to help ensure that there's audio samples available for the first frame in

Patch by James Zern <jzern at google>

Originally committed as revision 23473 to svn://

85e86b68 05/22/2010 02:13 AM David Conrad

Update regression tests after removing track timecode scale from mkvenc

Originally committed as revision 23248 to svn://

66061a12 03/20/2010 01:36 PM James Darnley

Add VorbisComment writing to FLAC files.
Patch by James Darnley <james darnley at gmail>.

Originally committed as revision 22605 to svn://

30f06a58 03/04/2010 08:53 AM David Conrad

Simplify starting and ending clusters

Originally committed as revision 22199 to svn://

cc3e2472 03/02/2010 09:41 PM Måns Rullgård

Place regression test output files in subdirs per family

Originally committed as revision 22155 to svn://

c676895f 01/30/2010 09:47 PM Måns Rullgård

Separate audio-only tests so they are only run once

Originally committed as revision 21556 to svn://