Commit Graph

35 Commits

Author SHA1 Message Date
bnnm
4e45cbd571 Add FFmpeg encoder delay logs 2018-10-07 21:16:57 +02:00
bnnm
65aca53cb5 Fix some XWB + FFmpeg subsongs 2018-10-07 03:26:05 +02:00
bnnm
9f6c1c4da2 Fix rare FFmpeg codecs decoding more samples after data EOF, cleanup 2018-10-06 22:39:34 +02:00
bnnm
804fceb5e7 Remove custom FFmpeg, as custom IO streamfile can be used instead 2018-09-10 02:19:34 +02:00
bnnm
74536e0933 Swap custom FFmpeg OPUS for custom streamfile IO 2018-09-10 01:23:34 +02:00
bnnm
9e54efe6d8 Remove FFMPEG_EA_XMA as it's now done with custom IO + layers 2018-08-25 13:22:57 +02:00
bnnm
8d7fb21c8f Check max filesize to avoid buggy reads
It was affecting some truncated Switch Opus that were reporting too many
samples
2018-08-15 22:47:22 +02:00
bnnm
9333a22c16 Remove some minor deprecated warnings
Tested with all major codecs we use for FFmpeg, all decode the same
2018-08-15 00:18:05 +02:00
bnnm
7a08ea263f Add FFmpeg fake header/header_size check to simplify some calls 2018-03-31 13:07:26 +02:00
bnnm
3ad6261208 Add check when calling free/seek/reset codec
It's possible to set a "coding" that triggers calls to the above while
codec_data is still null, when using special layouts like AAX
2018-03-10 16:59:00 +01:00
bnnm
42f44925f4 Fix possible segfault with some FFmpeg versions 2018-02-17 23:21:41 +01:00
bnnm
448d52fa4e Remove BGW ATRAC3 mode in FFmpeg in favor of custom streamfiles 2018-01-27 13:01:30 +01:00
bnnm
d09dc9739b Add encrypted ATRAC3 BGW [Final Fantasy XI PC] 2017-10-27 18:50:06 +02:00
bnnm
abaefd17cd Rename wwise_opus to switch_opus as it's apparently so 2017-09-30 00:16:40 +02:00
bnnm
6da7a86cc4 Add FFmpeg custom IO callbacks for data transmogrifiers
Those are custom read/seeks that can transform data on the fly, for
games that use variations of regular FFmpeg codecs. When FFmpeg tries to
read N bytes, the code may read less/more but still fill the buffer with
N transformed bytes. Must handle "real" (data) and "virtual"
(transformed) offset/sizes.

Works with fake headers too, but "virtual" offset/size now include the
fake header, as I found hard to keep in mind when the header_size was
being added/substracted (could be clearer).
2017-09-24 18:40:59 +02:00
bnnm
0103f449d0 Pass stream_index to FFmpeg manually for flexibility 2017-08-19 10:28:28 +02:00
bnnm
5cddab629d Add FFmpeg audio stream selection 2017-08-18 18:56:43 +02:00
bnnm
4263533ba9 Move FFmpeg code from ffmpeg.c to ffmpeg_decoder.c for consistency 2017-05-19 17:21:22 +02:00
bnnm
04f2cb0344 Moved FFmpeg RIFF utils to their own file
I'm going to add more later so it was getting kind of unwieldy
2017-02-25 13:54:05 +01:00
bnnm
248ec021d8 Add FFmpeg fake riff helpers: XMA1, XWMA, XMA2 tweaks 2017-02-25 12:25:10 +01:00
bnnm
9317ce9cc2 Add ffmpeg_set_skip_samples to manually fix encoder delay in some cases 2017-02-12 13:17:01 +01:00
bnnm
376b4acbb2 Minor func rename (ffmpeg_make_riff_xma2_from_fmt) 2017-01-14 22:25:15 +01:00
bnnm
f5750ba2cc Moved make_riff_x to ffmpeg_decoder.c (only actually used by FFmpeg)
Also added make_riff_xma2_from_fmt_be
2017-01-13 23:47:52 +01:00
bnnm
7e520162a3 FFmpeg: remove ACCURATE_LOOPING flag now it's battle-tested
It's fast enough and the older looping code would need to be redone
anyway
2017-01-08 02:54:18 +01:00
bnnm
cdfd47eb7b FFmpeg: expose block/frame size, rename totalFrames for clarity 2016-12-10 15:53:29 +01:00
bnnm
c5c97c4027 Fixed discard in edge cases 2016-12-03 18:17:37 +01:00
bnnm
0faa3286aa Don't manually check framesRead and rely on FFmpeg's EOFs
FFmpeg's duration isn't always reliable (ie. bad headers) and the
decoder detects EOFs already, extra decoding attempts should be ignored.
This way vgmstream can use other values without modifying
ffmpeg_codec_data's state.
2016-12-03 11:42:38 +01:00
bnnm
609bfb5d61 Minor fixes and cleanup 2016-12-01 23:49:00 +01:00
bnnm
f5da8281e2 Move FFmpeg reset to ffmpeg_decoder (cleanup) 2016-12-01 19:58:51 +01:00
bnnm
2754f05b11 Move FFmpeg looping to ffmpeg_decoder for further changes (cleanup) 2016-11-30 23:48:49 +01:00
bnnm
b9db20283d FFmpeg: keep block size constant (for edge loops), don't write on EOF 2016-11-29 20:22:20 +01:00
bnnm
cda61c3eeb simplify slightly discarding 2016-11-20 18:55:00 +01:00
bnnm
729658a19c fix discard samples (looping works now) + document 2016-11-20 15:39:21 +01:00
Christopher Snowhill
42c00f304b Add rollback to FFmpeg loop seeking, since FFmpeg doesn't seem to institute any rollback of its own. 2016-08-14 15:18:34 -07:00
Chris Moeller
cbe9bc3d6b Added FFmpeg support, and fixed a long standing issue with broken loops locking up the decoder. 2016-07-16 23:02:27 -07:00