Commits
| Author | Commit | Message | Date | Builds |
|---|---|---|---|---|
|
Added tag release-1_1_7 for changeset 70badbfe3674
|
|
||
|
|
Added tag transcode-1_1_7 for changeset 3b7e39310c40
|
|
||
|
|
[build][version] transcode 1.1.7
with vid.stab version 0.80. Another speed boost.
|
|
||
|
|
[filter][vid.stab] minor cosmetics.
Bump to 0.80.
|
|
||
|
|
[filter][vid.stab] bump vid.stab to 0.79
Apply missing changes from patchset 0.77.
|
|
||
|
|
Added tag release-1_1_6 for changeset a22b117aced2
|
|
||
|
|
[filter][vid.stab][style] minor nitpicks.
|
|
||
|
|
[filter][vid.stab] correct the crop behaviour.
Patch courtesy of Georg Martius <georg.martius a web d de>
|
|
||
|
|
[doc][changelog] ChangeLog bump.
|
|
||
|
|
[filter][vid.stab] revert uncareful filter_stabilize change.
|
|
||
|
|
lter][vid.stab] in crop mode, use border from previous stabilized frame.
The pixel border not suitable for stabilization are not equal to the preceding frame
as should be and is claimed in the description of the "crop" flag.
The output should be initialized using the *previous* output frame instead of the
input frame, which is not stabilized.
Pointed out by Guido Torelli.
|
|
||
|
|
[filter][vid.stab] in crop mode, use border from previous stabilized frame.
The pixel border not suitable for stabilization are not equal to the preceding frame
as should be and is claimed in the description of the "crop" flag.
The output should be initialized using the *previous* output frame instead of the
input frame, which is not stabilized.
Pointed out by Guido Torelli.
|
|
||
|
|
greatly simplify the ffmpeg libraries detection.
Removed some old cruft. No functional regressions spotted.
Also, do not hardcode the ffmpeg libraries names and leverage
some pkg-config flexiblity (and we already depend on it).
Patch courtesy of Nicolas Chauvet <kwizart a gmail d com>
|
|
||
|
|
bump bundled header to linux 2.6.38.6
|
|
||
|
|
split configure checks for video4linux and video4linux2
the v4l checks are interwined by accident. Moreover, v4l v1 code
is just a legacy, and provided for retrocompatibility.
Now the v4l and v4l2 checks are performed indipendently
if --enable-v4l is issued and the import modules are build according
to the interface avalability.
|
|
||
|
|
ffmpeg 0.8 compatibility.
patch courtesy of Nicolas Chauvet <kwizart a gmail d com>
|
|
||
|
|
[avi][portability] off_t is not guaranteed to be 64 bit wide.
loff_t should, so use it instead.
|
|
||
|
|
[misc][patch][debian][ffmmpeg] ffmpeg 0.8 compatibility: partly apply patchset from debian package.
applied:
02_toolame.diff
06-ffmpeg-0.8.diff
|
|
||
|
|
[ffmpeg][encoder] initial preset support.
Using transcode with ffmpeg+h264 causes some trouble with libx264 erroring out
with "broken ffmpeg default settings detected". libx264 further suggests: "use
an encoding preset (vpre)".
The attached patch adds two config parameters to ffmpeg.cfg, "vpre" and
"ffmpeg_datadir".
The former sets the video presets and accepts a comma separated list, e.g.
"fast,baseline", default being "medium".
The lat…
|
|
||
|
|
[avi][tools] better maximum size detection for avisplit
When using avisplit with the "maximum size" option on videos with large
variable gaps between keyframes many files were over the limit, some by
5Mb.
I found that the maximum size was tested against the current filesize +
1MB. A frame size was taken and stored in the variable "bytes", but
that was just the single frame so it was very small, I imagine that plus
filesize never gave correct results …
|
|
||
|
|
[filter][vid.stab] bump vid.stab to 0.79
|
|
||
|
|
[filter][vid.stab] bump vid.stab to 0.76
|
|
||
|
|
[filter][vid.stab] bump vid.stab to 0.77.
Patch courtesy of Georg Martius <georg dot martius at web dot de>
|
|
||
|
|
[cosmetic] print separate counters for Audio and Video buffers.
This affects only the verbose, user-friendly output.
(opposed as raw mode, machine-friendly).
This change, apart being nicer (!) slightly helps in detecting
deadlock problems.
|
|
||
|
|
[fix][threading][deadlock] fix potential deadlock when skipping frame range is outside the source limits.
Example: 100 frames source, skip range like 1-101.
There is something more evil lurking out here.
This basic case goes like there:
1. Video Import detects EOS and ends cleanly
2. Audio Import exits with unknown code (?! still investigating)
3. The encoder loop has 10 Video frames waiting, 1 Audio frames waiting.
4. After the last Audio frame consumed, the encoder tries to do another
loop, not knowing the import thread have finished.
5. It get s…
|
|
||
|
|
Removed tag release-1_1_6
|
|
||
|
|
[version][tag][release] HG tag for transcode 1.1.6
|
|
||
|
|
Removed tag release-1_1_6
|
|
||
|
|
[tag][version][build] transcode 1.1.6.
|
|
||
|
|
transcode 1.1.6
|
|