Commits · 3eb57eeb5163b1b0c6b419b47b64dfa9a6a753e4 · videolan / ffmpeg-mt

21 Sep, 2008 3 commits

Make MBAFF temporal direct mode closer to the spec. · 3eb57eeb

michael authored Sep 21, 2008

Fixes at least:
CAMANL2_TOSHIBA_B
CVMANL2_TOSHIBA_B
camp_mot_mbaff0_full


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15375 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

3eb57eeb

Move direct_dist_scale_factor() and direct_ref_list_init() after fill_mbaff_ref_list() · 80d431f4

michael authored Sep 21, 2008

as future changes will need the mbaff lists inited.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15374 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

80d431f4

Set poc for field references in MBAFF. · 8f4be5e6

michael authored Sep 21, 2008


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15373 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

8f4be5e6

20 Sep, 2008 2 commits

Rename check_func2 to check_func_headers. · 39d1b623

diego authored Sep 20, 2008


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15372 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

39d1b623

Log an error message in case of invalid number of channels. · 8fab8ac3
stefano authored Sep 20, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15371 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
8fab8ac3

19 Sep, 2008 4 commits

Remove fixme that is based on the assumption of !direct_8x8_inference && MBAFF. · d5e29cac
michael authored Sep 19, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15370 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
d5e29cac

Remove check and error message for "MBAFF + !direct_8x8_inference is not implemented" · 3c2ff1d2

michael authored Sep 19, 2008

because this mode does not exist, H.264-2007 says "When frame_mbs_only_flag is
equal to 0, direct_8x8_inference_flag shall be equal to 1."

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15369 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

3c2ff1d2

Avoid undefined behavior for removing elements that where not in the tree. · 9b09b199
michael authored Sep 19, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15368 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
9b09b199

Ensure that muxer and encoder aspect ratios match. So no · 64ff2a71

michael authored Sep 19, 2008

invalid files are generated.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15367 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

64ff2a71

18 Sep, 2008 14 commits

Support comments in preset files. · 9f72ac92

michael authored Sep 18, 2008


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15366 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

9f72ac92

Make pps/sps scaling_matrix initialization closer to the spec. · 72ba36eb

michael authored Sep 18, 2008

This does not affect any of the conformance streams.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15365 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

72ba36eb

Zero 'idx' for each iteration of the coupling gain loop and as it is only used · 9ee54f52

superdump authored Sep 18, 2008

within this loop, we can move its declaration there too.

Based on a patch by Alex Converse (alex converse gmail com)

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15364 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

9ee54f52

Cosmetics: missing brackets · f9ab3613

superdump authored Sep 18, 2008


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15363 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

f9ab3613

Cast the return value of get_bits() to signed integer such that the subtraction · 916fd9a4

superdump authored Sep 18, 2008

is conducted as a signed calculation.

Patch by Alex Converse (alex converse gmail com)

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15362 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

916fd9a4

Ensure that extradata is not parsed twice. · cc970d4c

michael authored Sep 18, 2008

Fixes FRExt1_Panasonic.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15361 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

cc970d4c

add newline at the end of message · 7b02154c

kostya authored Sep 18, 2008

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15360 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

7b02154c

ZMBV encoder forgot to fill coded frame pointer · b39ac007

kostya authored Sep 18, 2008

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15359 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

b39ac007

small wording fixes · 9d61cf28

diego authored Sep 18, 2008


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15358 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

9d61cf28

Add fourcc for MS RLE4 · 8faf6f96

kostya authored Sep 18, 2008

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15357 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

8faf6f96

Factorize out code used for MS RLE format decoding in different decoders. · 78c9edb8
kostya authored Sep 18, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15356 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
78c9edb8
Remove duplication of TRANSPOSE8 macro, also removing some redundant load/stores · 42d5981f
conrad authored Sep 18, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15355 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
42d5981f
Make a separate macro for writing an entire block from registers to memory · 7b0528ea
conrad authored Sep 18, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15354 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
7b0528ea

Factorize SSE2_(Row|Column)_IDCT into one macro · f2e13cff

conrad authored Sep 18, 2008

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15353 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

f2e13cff

17 Sep, 2008 12 commits

revert r14983, value is not sample aspect ratio, it is display aspect ratio · 5b360c79
bcoudurier authored Sep 17, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15352 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
5b360c79

Cosmetics: remove excessive blank lines · 72640b9c

conrad authored Sep 17, 2008

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15351 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

72640b9c

Ensure MMX/SSE2 VP3 IDCT selection isn't disabled when only Theora is enabled · 63188a48
conrad authored Sep 17, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15350 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
63188a48

Cosmetics: reindent · 07aa70de

conrad authored Sep 17, 2008

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15349 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

07aa70de

libx264 ffpresets: · 33ab8f2d

superdump authored Sep 17, 2008

 * default - x264 CLI default values
 * fastfirstpass - disables options unnecessary for first pass
 * normal - Dark Shikari’s Q55 preset
 * hq - Dark Shikari’s Q75 preset
 * max - all options maxed out (at time of writing)


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15348 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

33ab8f2d

Rewrite SSE2 VP3 IDCT in inline asm · a25bd4cc

conrad authored Sep 17, 2008

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15347 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

a25bd4cc

Rewrite MMX VP3 IDCT in inline asm · 216b2f4f

conrad authored Sep 17, 2008

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15346 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

216b2f4f

MMX/SSE2 VP3 IDCT are bitexact now that the dequantization matrices are permutated correctly · d9fda003
conrad authored Sep 17, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15345 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
d9fda003
Take idct permutation into account when initializing dequantization matrices · 5c1a00c2
conrad authored Sep 17, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15344 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
5c1a00c2

Use SAMPLE_FMT_FLT instead of doing the float->int convertion in the · 1bbaadf1

vitor authored Sep 17, 2008

decoder. This changes very little the output and any difference should 
be inaudible.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15343 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

1bbaadf1

Second-stage gain and Moving Average Predictor codebooks. · 1134e790
voroshil authored Sep 17, 2008
```
git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15342 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
1134e790

Another set of approved changes for G.729: includes, formats structure, · 0654d1c1

voroshil authored Sep 17, 2008

several parts of main decoding routine and initialization code.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15341 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

0654d1c1

16 Sep, 2008 5 commits

Validate pulse position and error out if an invalid position is encountered. · f7205b05

superdump authored Sep 16, 2008

Patch by Alex Converse (alex converse gmail com)


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15340 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

f7205b05

Only print "invalid new backstep" when its really invalid. · 57e09361

michael authored Sep 16, 2008

Fixes issue541.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15339 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

57e09361

Fix regression test failure caused by golomb limit not being considered · f0cad86a
michael authored Sep 16, 2008
```
in r15334.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15338 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
```
f0cad86a

sync to latest x264 git, field was renamed · 9e4d607a

bcoudurier authored Sep 16, 2008

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15337 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

9e4d607a

Make decoder more robust by default against broken encoders. · b3d77ab9

michael authored Sep 16, 2008

Fixes issue540.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@15336 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

b3d77ab9