The --paused option will start the player in paused state. That means it
will start out with a still image of the first frame.
This can be useful in combination with --ss to inspect a certain frame.
Caveat: this plays a small bit of audio at the start, which might be
perceived as an annoying artifact. This is because this is implemented
by frame stepping after initialization in order to decode and display
the first video frame.
The vd_ffmpeg decode() function returned without doing anything if the
input packet had size 0. This meant that flushing buffered frames at
EOF did not work. Remove this test. Have the core code skip such
packets coming from the file being played instead (Libav treats
0-sized packets as flush signals anyway, so better assume such packets
do not represent real frames with any codec).
Libav has changed back to not modifying avctx->has_b_frames based on
the extra buffering caused by thread use. Add back the code to do the
adjustment on the player side once again.
The timing mode using the buffering info is no longer the default, so
in most cases having this right or not won't matter for playback.
Remove the private bswap and intreadwrite.h implementations and use
libavutil headers instead.
Originally these headers weren't publicly installed by libavutil at
all. That already changed in 2010, but the pure C bswap version in
installed headers was very inefficient. That was recently (2011-12)
improved and now using the public bswap version probably shouldn't
cause noticeable performance problems, at least if using a new enough
compiler.
Change demux_lavf to use CodecID -> RIFF tag mappings that are now
available through the public Libav API. Previously it used a copy in
ffmpeg_files/taglists.c. That can now be deleted.
Change various code to use the latest Libav API. The libavcodec
error_recognition setting has been removed and replaced with different
semantics. I removed the "--lavdopts=er=<value>" option accordingly,
as I don't think it's widely enough used to be worth attempting to
emulate the old option semantics using the new API. A new option with
the new semantics can be added later if needed.
Libav dropped APIs that were necessary with all Libav versions
until quite recently (like setting avctx->age), and it would thus not
be possible to keep compatibility with previous Libav versions without
adding workarounds. The new APIs also had some bugs/limitations in the
recent Libav release 0.8, and it would not work fully (at least some
avcodec options would not be set correctly). Because of those issues,
this commit makes no attempt to maintain compatibility with anything
but the latest Libav git head. Hopefully the required fixes and
improvements will be included in a following Libav point release.
Using type = -1 for terminating the command argument list was an
unnecessary complication. Change it to 0 to make use of default
initialization. Remove the explicit termination from the command
definitions. Now the argument list is automatically terminated by C
default initialization rules.
Also remove other redundant {0} initializers.
When an argument's default value is actually initialized to something other
than 0, make the field being initialized explicit (e.g. {1} becomes
{.i = 1}).
Try to make use of whitespace more consistent.
Setting the WM_NAME/WM_ICON_NAME window properties didn't always work:
apparently there are some characters that can't be represented in the X
STRING or COMPOUND_TEXT encodings, such as U+2013 EN DASH. The function
Xutf8TextListToTextProperty partially converts the string, and returns
a value different from 'Success'. This means vo_x11_set_property_string
didn't set these window properties.
On most modern window managers, this is not a problem, since these use
the _NET_WM_NAME/_NET_ICON_NAME and the UTF8_STRING encoding. Some older
WMs like IceWM don't read these, and the window title remains blank.
It's not clear what exactly we should do in this situation, but fix it
by setting set the WM_NAME/WM_ICON_NAME properties as UTF8_TEXT. This
violates the ICCCM, but at least IceWM seems to handle this well.
See also:
http://lists.freedesktop.org/archives/xorg/2004-September/003391.htmlhttp://lists.freedesktop.org/archives/xorg/2004-September/003395.html
Also fix a minor memory leak when conversion to COMPOUND_TEXT fails.
The default compression setting is 7, which is hopefully a good balance
between speed of compression, and resulting file sizes. The maximum png
compression will be very slow even on fast computers. On the other hand,
the lowest compression setting produces files of several MB size with
normal video resolutions, which should be avoided as well.
The screenshot image file type can now be selected with the
--screenshot-filetype option. The --screenshot-jpeg-quality option
controls the compression setting of the written JPEG image file.
Now the option --term-osd=force will cause mplayer to display all OSD
messages on the terminal, even if there is video.
Possible values for --term-osd:
- auto: use video OSD, or of there's no video, the terminal (default)
- off: always use video for OSD
- force: always use terminal for OSD
-term-osd and --term-osd are equivalent to --term-osd=force. This
changes the meaning of the option, since -term-osd used to enable the
OSD default behavior, i.e. --term-osd=auto.
-noterm-osd has the same effect as --term-osd=off, and is kept for
compatibility.
Implementation note:
The location for the OSD text was shared between the two code paths (it
was in osd_state.osd_text). We can't rely on the fact that the video-OSD
update code normally isn't run when --term-osd is called. When e.g.
panscan is updated, the video OSD code will draw the OSD anyway. This
would sometimes show unwanted OSD text on the video.
Deal with this by putting the current terminal-OSD text in a different
place (in MPContext.terminal_osd_text) to deal with this.
If an m_option_type_choice option is declared with M_OPT_IMPLICIT_DEFAULT
in its flags, it doesn't require a parameter. For example, if --opt is
such an option, it can be invoked as "--opt=val", "-opt", or "--opt".
The last two will set the option to the first choice the option declares.
Note that "-opt val" (using the old option syntax) is not allowed in this
case, as it would be ambiguous.
Normal option parsing should be unaffected.
Playing a .cue file directly will now parse the .cue file, and load and
play the file(s) referenced in the cue. If multiple files are referenced,
a timeline including all files will be created to create the impression
of a single, flat audio file containing all the tracks.
For each track, a chapter is created. The chapter navigation commands can
be used to jump between tracks. The chapter titles will use the string
provided by the track's TITLE cue command. (The -identify command can be
used to print all chapters in a not so user friendly way.)
Other than the chapter names, there is no attempt at displaying or exposing
any other meta data contained in the cue files yet.
The handling (or lack of thereof) of gaps (track pregaps and postgaps) is
probably not correct yet. In general, mplayer's mapping of tracks to the
source audio files can be verified by examining the timeline, which will
be printed when passing the -v switch.
Note that this has nothing to do with the old cue:// support. The old code
isn't touched, and is still only able to play .cue/.bin pairs. Prefixing a
.cue file with cue:// will always invoke the old code, while playing a .cue
file directly (i.e. "mplayer file.cue") will always use the new code.
Playing audio images (.cue/.bin pairs of files) doesn't work yet.
bstr_strip_ext and bstr_get_ext were taken from find_subfiles.c.
bstr_cut is extended to work like bstr_splice: passing a negative
argument will start counting from the end of the string, e.g.
bstr_cut("abc", -2) == "bc"
If digital pass-through is used, this supported setting the volume (just
mute, actually), but not getting the volume. This will probably lead to a
stuck mute state in the mplayer frontend. Make the code respond to volume
queries even if digital pass-through is used.
Ideally, ao_coreaudio should implement full mute control, but I can't
even test on OSX.
The mixer frontend code can now make use of a proper system mixer mute
toggle, if the audio output driver supports it.
The consequence is that, if support is available, mplayer will no longer
temporarily set the system volume to 0 if mute is enabled.
Generally, the code now deals with the following combinations of available
AO features:
- software volume control forced by user (--softvol / soft_vol flag)
=> if enabled, never touch the "hardware" controls
- "hardware"/driver volume control available (whether
AOCONTROL_GET/SET_VOLUME works)
=> if not available, enable volume controls by enabling softvol
- "hardware"/driver mute control (AOCONTROL_GET/SET_MUTE)
=> if not available, emulate by setting volume to 0
- whether the volume+mute controls are kept or not when the AO is closed
(indicated by ao->no_persistent_volume)
=> if not persistent, restore the volume/mute next time the AO is opened
The mplayer frontend (specifically, mixer.c) needs to know this. If the
audio output doesn't remember the volume across reinitialization, the
frontend will restore the volume settings. There is also the assumption
that the volume setting isn't global in this case (i.e. changing it
won't change the volume of other applications or annoy the user
otherwise).
None of these changes have been tested. I'm guessing that ESD and NAS do
per-connection non-persistent volume settings.
Changing the volume when softvol is enabled or if the audio output driver
doesn't support volume controls causes insertion of the "volume" filter.
This fails with AC3. Since the filter wasn't removed after that, and the
filter chain was in a bogus state, random crashes occured past this
point.
Fix it by reinitializing the filter chain completely on failure. Volume
controls simply won't work. (This can't be fixed, because AC3 is a
compressed format, and would require additional decoding/encoding passes
in order to support arbitrary volume changes.)
This also affects balance controls.
I'm not sure what's the point of this feature. Aside from that, the EDL
code is relatively buggy anyway, and I see no reason why such an obscure
feature should be left in, if it possibly causes bugs.
Before this commit, the mute state was only reset when either mute was
explicitly cleared, or the volume was changed via mplayer controls. If
the volume controls are connected to the system mixer, and the system
mixer volume is changed otherwise (e.g. with alsamixer), the mute
setting was inconsistent.
Avoid this by checking the volume. If the returned volume is not 0, the
mute flag is considered invalid. This relies on system mixers always
returning a volume of 0 when mplayer has set the volume 0.
Possible caveat: if the audio output's volume control don't return a
volume of exactly 0 after 0 was written, enabling mute basically won't
work. It will set the volume to silence, forget the previous volume, and
report that mute is disabled.
Muting audio is implemented by setting the volume controls to 0. This
has the annoying consequence that attempting to change the volume while
the audio is muted will reset the user's volume setting. E.g. increasing
the volume while muted will start from 0, instead from the volume that
was set before muting. Changing the volume while muted effectively resets
the volume to 0 (which is not very useful), with no possibility of
restoring the old voume.
This commit makes mplayer always report the volume that was set when mute
was enabled while mute is still active.
Caveat: this might be have confusing effects when the volume control is
directly connected with a system wide mixer setting. Now it's even less
obvious (and thus more confusing) that muting will set the mixer volume
to 0.
Also always clip input volumes, and remove some minor code duplication.
When you mute audio, mplayer is supposed to restore the volume controls
on exit. This affects when --softvol isn't used and the audio output
driver volume controls directly affect the system wide volume controls.
This wasn't done in some cases.
Some audio outputs don't provide access to a system-wide mixer control, and
do per-application audio mixing. Further, some of these forget the volume
as soon as the audio device is closed. This can be annoying, because
mplayer will "forget" the volume when playing a new file or when crossing
ordered chapter boundaries. Support restoring the volume on audio
reinitialization if an audio output driver knowingly behaves this way.
(This doesn't change that mplayer never writes any settings into the config
file, including volume settings.)
This commit doesn't yet change any actual output driver to use this code.
Hopefully make some logic in the volume restore code a bit more robust.
At least in the case when switching to no audio track and then switching
back, the volume settings were not restored with --softvol. Fix this by
moving the call restoring the settings to a better place.
There were some slight differences between what input.conf mapped, and
what was in input.c def_cmd_binds[]. Make them match.
Add some minor documentation improvements in input.cfg.
Also remove double comments ('##'), because they were confusing.
At least on some keyboards, the key between '0' and 'Enter' on the
key pad is mapped to KP_Separator. Since X11 VOs accept unicode
input, the mplayer keycode this key generates depended on the numlock
state, and with numlock enabled this mapped to an ASCII character.
This is probably not what the user wanted, since two physical keys
will always map to the same key code.
Map it to KP_DEC.
This assumes the terminal uses UTF-8. If invalid UTF-8 is encountered (for
example because the terminal uses a legacy encoding), the code falls back
to the old method and feeds each byte as key code to the input code.
In theory, UTF-8 input could randomly fail, because the code in getch2.c
doesn't try to fill the input buffer correctly with input sequences
longer than a byte. This is a problem with the design of the existing
code.
This change allows using non-ASCII keys with X11. These keys were ingored
before.
Technically, this creates an invisible, non-interactive input method
context. If creation fails, the code falls back to the old method, which
allows a subset of ASCII only.
This moves all key codes above the highest valid unicode code point (which
is 0x10FFFF). All key codes below MP_KEY_BASE now directly map to unicode.
Configuration files (input.conf) can contain unicode characters in UTF-8
to map non-ASCII characters/keys.
This shouldn't change anything user visible, except that "direct key codes"
(as used in input.conf) will change their meaning.
Adding these flags improves the quality of the YUV->RGB conversion when
screenshots are taken. It trades precision for performance.
This doesn't affect any other swscale uses, such as vf_scale or vo_x11.
Based on a patch by cantabile. Fixes#140.
When --shuffle was used, the pt_step -1 jumped to the next file, instead
of the previously played file. This was because the playtree entries were
never actually shuffled, but instead a random unplayed file was picked.
Fix this by actually shuffling the playtree in advance. I couldn't see any
clear location where exactly this should happen, so it's done when a
playtree iterator is created.
Not removing the old code, since new playtree entries could be added while
an iterator is active.
This simplifies the code. It also can make loading of mkv files faster, as
initializing two ASS_Render instances required scanning mkv font
attachments twice.