cosmetics: Fix two common typos: wont --> will not, lets --> let us.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14372 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

cosmetics: typo/grammar fixes

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14371 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

cosmetics: Fix indentation after last commit.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14370 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Fix segfault with non existing input files.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14369 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Cosmetics: grammar nits

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14368 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

misc spelling/grammar fixes

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14367 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Remove AltiVec vector declaration compiler compatibility macros.
The original problem was that FSF and Apple gcc used a different syntax
for vector declarations, i.e. {} vs. (). Nowadays Apple gcc versions support
the standard {} syntax and versions that support {} are available on all
relevant Mac OS X versions. Thus the greater compatibility is no longer
worth cluttering the code with macros.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14366 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Use tkhd matrix for proper display in mov.
Patch by John Schmiederer jschmiederer on2 com
Original thread: [PATCH] Use tkhd matrix for proper display in mov
Date: 05/23/2008 09:31 PM

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14365 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Rename update() function to backward_filter() and add a doxy comment

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14364 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

More tables renaming/commenting

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14363 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Follow FFmpeg convention of returning negative values on error in eval_lpc_coeffs()

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14362 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Rename pred() to eval_lpc_coeffs() and add a doxy comment

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14361 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Give tables a more descriptive name

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14360 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Add comments to do_hybrid_window()

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14359 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Give parameters of do_hybrid_window() more meaningful names

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14358 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Rename co() function to the much more descriptive name of
do_hybrid_window().

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14357 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Simplify co(): write constant in a more readable way

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14356 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Simplify co(), use memcpy/memmove and colmult() when useful.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14355 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Reset prev_frame_num for idr frames as required by the spec.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14354 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Better explanation for buf_size==0 code.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14353 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

SMPTE 421M Annex L (aka .rcv) muxer

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14352 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

100l cross_idr simplification had a typo

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14351 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Remove truncated support, use a AVParser if you used that, its the same.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14350 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Simplify cross_idr calculation.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14349 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Reduce the loglevel on a log message in the Nellymoser decoder. Related to Roundup 447.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14348 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Remove redundant keyframe checks.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14347 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Reindent after last commit

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14346 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Do not rescale AV_NOPTS_VALUE in output_example.c

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14345 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

In-loop deblocking filter for VC-1 decoder

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14344 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

My commit at r14340 was not the right solution. For a monocromatic
frame, there will be only one centroid and it will be impossible to
find three distinct ones. It is better to just avoid shifting if
there are not three different centroids.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14343 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Do not merge declaration and initialization in elbg.c to
workaround an ICC bug.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14342 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

When picking a "high utility centroid" do not pick one
that has no corresponding points. Not only it is the
worst possible pick, but also the code was written
without this case in mind.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14341 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

The codebook generator algorithm involves picking three
different codebook centroids ("high utility", "low
utility" and "closest to the low utility one"). This
change avoid the corner case of choosing two times the
same centroid.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14340 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Support gaps in the frame num.
Fixes at least:
MR3_TANDBERG_B.264
MR4_TANDBERG_C.264
MR5_TANDBERG_C.264

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14339 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Reset frame_num for MMCO 5.
Fixes at least
MR2_TANDBERG_E.264

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14338 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

3rd try to get the loop filter ref/mv check working correctly.
Fixes at least:
src19td.IBP.264
CVWP3_TOSHIBA_E.264
cvmp_mot_picaff0_full_B.26l
CVMP_MOT_FRM_L31_B.26l
cvmp_mot_frm0_full_B.26l
CVMP_MOT_FLD_L30_B.26l
cvmp_mot_fld0_full_B.26l

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14337 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

qscale has a range of 0..51 we thus do not need a 256 entry table and neither need
to and it with 0xFF.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14336 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Setting chroma_qp to 0 for PCM blocks is really a interresting minor
optimization, more interresting would it have been had the author
thought about what value chroma_qp would have for the following MB.
Or failing that, had actually tested the code.
So this reverts this non functional optimization, and makes the code
work.
Fixes at least CAPM3_Sony_D.jsv

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14335 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Deobfusate chroma_qp_table[1] init code.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14334 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Enhance qblur and qcomp documentation.
Patch by tripp eliared yahoo com

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14333 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Revert r13542: Spatial direct mode probably also works with PAFF.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14332 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Loop filtering with multiple slice types fix.
Fixes at least:
CABAST3_Sony_E.jsv
CABASTBR3_Sony_A.jsv
CABASTBR3_Sony_B.jsv

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14331 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Make configure --help document the --enable-debug option.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14330 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Add complete listing of bitstream filters and 2 examples to the documentation.
patch by tripp, eliared yahoo com

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14329 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Correct b_quant offset comment in avcodec.h, it is the description for
b_quant factor. patch by tripp, eliared yahoo com

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14328 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Move --disable-debug to a better place within configure help output.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14327 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Use slice_type_nos instead of slice_type where it makes no difference.
This way we use slice_type_nos almost everywhere which means 1 variable less
for gcc to put in a register.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14326 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

SI/SP slice support. (not bitexact)

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14325 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

remove superflous ()

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14324 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Simplify slice type checks.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14323 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

More correct ref comparission in the loop filter.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14322 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Check the allocated pointer instead of the given pointer as intended
in vhook/ppm.c:Configure.
patch by Erik Hovland, erik hovland org

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14321 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Clarify -minrate/-maxrate description in the documentation.
patch by tripp, eliared yahoo com

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14320 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Simplify spatial direct ref selection with FFMIN3()

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14319 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

cosmetics: Make libavcodec/ppc/dsputil_altivec.c conform to style guidelines.
This includes indentation changes, comment reformatting, consistent brace
placement and some prettyprinting.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14318 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

cosmetics: Reindent two misplaced braces.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14317 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

cosmetics: Reformat PPC code in libavcodec according to style guidelines.
This includes indentation changes, comment reformatting, consistent brace
placement and some prettyprinting.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14316 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Fix used references indexes in the loop filter.
Disable filter_mb_fast() as it optimized the incorrect code.
Fixes at least:
BA3_SVA_C.264
CABA3_SVA_B.264
CABACI3_Sony_B.jsv
CAFI1_SVA_C.264
camp_mot_frm0_full.26l
CAWP5_TOSHIBA_E.264
CVFI2_SVA_C.264
CVSE3_Sony_H.jsv
CVWP2_TOSHIBA_E.264
CVWP5_TOSHIBA_E.264
SL1_SVA_B.264

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14315 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Add skip* AVOptions, which seem to have been missed.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14314 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Decrease severity of 'invalid new backstep' log message to warning
in mpeg audio decoder.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14313 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Supply context argument to av_log() in mpeg audio decoder.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14312 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

read 'coupling in use' parameter from bitstream into decoder context first,
then copy to the local variable.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14311 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Fix h264_loop_filter_strength_mmx2() so it works with PAFF.
fixed at least:
CVFI1_Sony_D.jsv
CVFI1_SVA_C.264
MR6_BT_B.h264

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14310 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Prevent the qdm2 code from overreading/overflowing. Fixes Coverity ID 112 run 2

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14309 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Take the brain amputated spec litteral if the user asks for it (-strict 1)
That is, add 16 frames delay, cache trashing and av desync.
fixes at least the following reference bitstreams:
CABA3_Sony_C.jsv
CACQP3_Sony_D.jsv
CAMANL1_TOSHIBA_B.264
CANL3_Sony_C.jsv
CVBS3_Sony_C.jsv
CVMANL1_TOSHIBA_B.264

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14308 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Remove explicit strict_std_compliance support as it seems to work fine over
AVOptions, actually it works even finer as it also gets passed to the
decoder.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14307 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Make strict_std_compliance avaialable to decoders.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14306 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Fix for possible null pointer dereferencing, closes Coverity report 68 run 2.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14305 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Cosmetics: alignement

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14304 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Simplify co(): do not abuse pointer aritmetics

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14303 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Make the automatic has_b_frames increasing code increase it up to the
same maximum that can be achived by specifying the value in the bitstream.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14302 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Simplify co(): remove variables that are only used once

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14301 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Try to fix my latest 100l fate SVQ3 breakage.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14300 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Remove apparently unneeded out!=NULL check.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14299 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Simplify picture output code.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14298 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Remove useless #if 0 code.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14297 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Dont forget marking disposed delayed pics as unused.
Fixes issue335

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14296 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

add ogv extension

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14295 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

cosmetics, remove space

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14294 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

update swf video frame number when muxing done, fix #439

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14293 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

indent

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14292 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

simplify

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14291 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

cosmetic: use hex value instead of decimal to show relation to 0x7FFFFF

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14290 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Remove delayed_output_pic, i do not understand what this variable was good for.
Its simpler and might be faster now.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14289 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Drop out of order pics even if bitstream_restriction_flag is set.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14288 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Set avctx->coded_frame in RoQ encoder. At some point in
the SVN history this became mandated. Fix issue 548.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14287 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Make the ugly hack which uses an unused entry in the internal buffer
array actually use a unused one, so it does work.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14286 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Make ffmpeg -t X with X<=0 transcode nothing.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14285 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

minor simplification of qdm2_decode_fft_packets()

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14284 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Make frames non reference by default, so they are disposed of automatically
in MPV_frame_start() if we break out due to an error at a random place.
Fixes issue334

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14283 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Correcting displayed muxing overhead and video stream size.
Partially fixes issue348.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14282 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Support reading large metadata.
fixes issue187

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14281 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Make bitstream_* fields unsigned.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14280 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Only realloc() bitstream buffer when the needed size increased,
this is needed to prevent loosing bitstream data with large metadata.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14279 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

If metadata has been parsed goto end instead of trying to
decorrelate and output somehing nonexistng.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14278 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Fix a mem leak in vc1_decode_frame().
Patch by Erik Hovland erik hovland org

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14277 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Fix a mem leak in av_find_stream_info().
Patch by Erik Hovland erik hovland org

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14276 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

intreadwrite: support DEC compiler __unaligned type qualifier

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14275 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

MPEGTS: Improve probe function

When a sync byte is found, check that transport_error_indicator is zero,
and adaptation_field_control is valid (non-zero).

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14274 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

cosmetics, remove space

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@14273 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b