Now any action that stops playback of a file (even playlist navigation)
will save the position. Normal EOF is of course excluded from this, as
well as commands that just reload the current file.
The option name is now slightly off, although you could argue what the
word "quit" means.
Fixes#1148 (or at least this is how I understood it).
Run opening the stream and opening the demuxer in a separate thread.
This should remove the last code paths in which the player can normally
get blocked on network.
When the stream is opened, the player will still react to input and so
on. Commands to abort opening can also be handled properly, instead of
using some of the old hacks in input.c. The only thing the user can
really do is aborting loading by navigating the playlist or quitting.
Whether playback abort works depends on the stream implementation; with
normal network, this will depend on what libavformat (via "interrupt"
callback) does.
Some pain is caused by DVD/BD/DVB. These want to reload the demuxer
sometimes. DVB wants it in order to discard old, inactive streams.
DVD/BD for the same reason, and also for reloading stream languages
and similar metadata. This means the stream and the demuxer have to
be loaded separately.
One minor detail is that we now need to copy all global options. This
wasn't really needed before, because the options were accessed on
opening only, but since opening is now on a separate thread, this
obviously becomes a necessity.
Also recreate ASS_Library on every file played. This means we can move
the code out of main.c as well.
Recreating the ASS_Library object has no disadvantages, because it
literally stores only the message callback, the (per-file) font
attachment as byte arrays, and the set of style overrides. Hopefully
this thing can be removed from the libass API entirely at some point.
The only reason why the player core creates the ASS_Renderer, instead
of the subtitle renderer, is because we want to cache the loaded fonts
across ordered chapter transitions, so this probably still has to stay
around for now.
When the VO was moved it its own thread, responsibility for redrawing
was given to the VO thread itself. So if there was a condition that
indicated that redrawing was required, like expose events or certain
VOCTRLs, the VO thread was redrawing itself.
This worked fine, but there are some corner cases where this works
rather badly. E.g. if I fullscreen the player and hit panscan controls
with mpv's default autorepeat rate, playback stops. This happens because
the VO redraws itself after every panscan change command. Running each
(repeated) command takes so long due to redrawing and (involuntary)
waiting on vsync, that it never leaves the input processing loop while
the key is held down. I suspect that in my case, redrawing in fullscreen
mode just gets slow enough that it takes 2 vsyncs instead of 1 on
average, and the processing time gets larger than the autorepeat delay.
Fix this by taking redraw control from the VO, and instead let the
playloop issue a "real" redraw command to the VO if needed. This
basically reverts redraw handling to what it was before moving the VO to
a thread.
CC: @mpv-player/stable
Each subsystem (or similar thing) had an INITIALIZED_ flag assigned. The
main use of this was that you could pass a bitmask of these flags to
uninit_player(). Except in some situations where you wanted to
uninitialize nearly everything, this wasn't really useful. Moreover, it
was quite annoying that subsystems had most of the code in a specific
file, but the uninit code in loadfile.c (because that's where
uninit_player() was implemented).
Simplify all this. Remove the flags; e.g. instead of testing for the
INITIALIZED_AO flag, test whether mpctx->ao is set. Move uninit code
to separate functions, e.g. uninit_audio_out().
For the sake of libmpv. Might make things much easier for the user,
especially on Windows. On the other hand, it's a bit sketchy that a
command exists that makes the player access arbitrary memory regions.
(But do note that input commands are not meant to be "secure" and never
were - for example, there's the "run" command, which obviously allows
running random shell commands.)
Somewhat more flexible: now there's a separate overlay struct, and you
don't need to coerce all state into struct sub_bitmap. Also, removing
the previous mapping (munmap call) is now all in one place, the
replace_overlay function.
Makes the next commit easier to implement.
The messages "Audio: no audio" and "Video: no video" could be printed
twice each if initializing them failed. Prevent his silliness.
CC: @mpv-player/stable
Apparently this is what users want. When playing with normal speed,
nothing is done. When playing slower than normal, resampling is used
instead, because scaletempo (which does the pitch correction) adds
too many artifacts.
There's no real reason why audio_init_filter() should exist. Just use
af_init or af_reinit directly. (We lose a useless message; the same
information is printed in a quite close place with more details.)
Requires less code, and the way the filter chain is marked as having
failed to initialize allows just switching off audio instead of
crashing if trying to insert a volume filter in mixer.c fails, and
recreating the old filter chain fails too.
This would play some silence in case video was slower than audio. If
framedropping is already enabled, there's no other way to keep A/V
sync, short of changing audio playback speed (which would give worse
results). The --audiodrop option inserted silence if there was more
than 500ms desync.
This worked somewhat, but I think it was a silly idea after all. Whether
the playback experience is really bad or slightly worse doesn't really
matter. There also was a subtle bug with PTS handling, that apparently
caused A/V desync anyway at ridiculous playback speeds.
Just remove this feature; nobody is going to use it anyway.
Commit 64b7811c tried to do the "right thing" with respect to whether
keyboard input should be enabled or not. It turns out that X11 does
something stupid by design. All modern toolkits work around this native
X11 behavior, but embedding breaks these workarounds.
The only way to handle this correctly is the XEmbed protocol. It needs
to be supported by the toolkit, and probably also some mpv support. But
Qt has inconsistent support for it. In Qt 4, a X11 specific embedding
widget was needed. Qt 5.0 doesn't support it at all. Qt 5.1 apparently
supports it via QWindow, but if it really does, I couldn't get it to
work.
So add a hack instead. The new --input-x11-keyboard option controls
whether mpv should enable keyboard input on the X11 window or not. In
the command line player, it's enabled by default, but in libmpv it's
disabled.
This hack has the same problem as all previous embedding had: move the
mouse outside of the window, and you don't get keyboard input anymore.
Likewise, mpv will steal all keyboard input from the parent application
as long as the mouse is inside of the mpv window.
Also see issue #1090.
We inserted these filters with fixed parameters, which was ok. But this
also didn't change image parameters for the filters down the filter
chain and the VO. For example, if rotation by 90° was requested by the
file, we would insert a filter and rotate the video, but the VO would
still receive image parameters that direct rotation by 90°.
This wasn't a problem, but it could become one.
Fix this by letting the filters automatically pick up the image params.
The image params are reset on application. (We could probably also
always try to apply and reset image params in a filter, instead of
having special "auto" parameters. This would probably work, and video.c
would insert a "rotate=0" filter. But I'm afraid this would be confusing
and the current solution is cosmetically slightly nicer.)
Unfortunately, the vf_stereo3d.c change turned out a big mess, but once
the "internal" filter is fully replaced with libavfilter, most of this
can be radically simplified.
Until now, creating the input_ctx was delayed until the command line
and config files were parsed. Separate creation and loading so that
input_ctx is available from start.
This should make it possible to simplify some things. For example,
some complications with Cocoa were apparently only because input_ctx
was available only "later". (Although I'm not sure if this is still
relevant, or if the Cocoa code should even be organized this way.)
This warning makes absolutely no sense. Passing an empty string to
printf-like functions is perfectly fine. In the OSD case, it just sets
an empty message, practically clearing the OSD.
set_osd_bar_chapters() always cleared the OSD bar stops, even if the
current bar was not the seek bar. Obviously it should leave the state of
the bar alone in this case.
Also change the function control flow so that we can drop one
indentation level, and do the equivalent change for the other OSD bar
functions.
Eliminate the remains of the OSD message stack. Another simplification
comes from the fact that we do not need to care about time going
backwards (we always use a monotonic time source, and wrapping time
values are practically impossible). What this code was pretty trivial,
and by now unnecessarily roundabout.
Merge get_osd_msg() into update_osd_msg(), and add_osd_msg() into
set_osd_msg_va().
There's no need to update OSD messages and the terminal status if nobody
is going to see it. Since the player doesn't block on video display
anymore, this update happens to often and probably burns slightly more
CPU than necessary. (OSD redrawing is handled separately, so it's just
mostly useless text processing and such.)
Change it so that it's updated only on every video frame or all 50ms
(whatever comes first).
For VO OSD, we could in theory try to lock to the OSD redraw heuristic
or the display refresh rate, but that's more complicated and doesn't
work for the terminal status.
When using --force-window (and no video or cover art), this heuristic
prevents any redrawing during seeking. It should be applied only if
there is any form of video.
This makes subtitle display somewhat work if no video is displayed, but
a VO window exists (--force-window or cover art display).
The main problem with normal subtitle display is that it's locked to
video: it uses the video PTS as reference, and the subtitles advance
only if a new video frame is displayed. In audio-only mode on the other
hand, no video frame is ever displayed (or only 1 in the cover art
case). You would need a workaround to adjust the subtitle PTS, and you
would have to decide with what frequency to update the display. In
general, there is no "right" display FPS for subtitles. Some formats
(ASS) have animations parameterized by time, and any refresh rate could
be used.
Sidestep these problems by enabling the text OSD-based subtitle
mechanism. This is similar to --no-sub-ass, and updates and renders
subtitles with plain OSD. It has some caveats: no bitmap subs, somewhat
incorrect timing, no formatting. Timing in particular is a bit strange
and depends how often the audio output asks for new data, or other
events that happen to wakeup the playloop.
Before this commit, there was AF_FORMAT_AC3 (the original spdif format,
used for AC3 and DTS core), and AF_FORMAT_IEC61937 (used for AC3, DTS
and DTS-HD), which was handled as some sort of superset for
AF_FORMAT_AC3. There also was AF_FORMAT_MPEG2, which used
IEC61937-framing, but still was handled as something "separate".
Technically, all of them are pretty similar, but may use different
bitrates. Since digital passthrough pretends to be PCM (just with
special headers that wrap digital packets), this is easily detectable by
the higher samplerate or higher number of channels, so I don't know why
you'd need a separate "class" of sample formats (AF_FORMAT_AC3 vs.
AF_FORMAT_IEC61937) to distinguish them. Actually, this whole thing is
just a mess.
Simplify this by handling all these formats the same way.
AF_FORMAT_IS_IEC61937() now returns 1 for all spdif formats (even MP3).
All AOs just accept all spdif formats now - whether that works or not is
not really clear (seems inconsistent due to earlier attempts to make
DTS-HD work). But on the other hand, enabling spdif requires manual user
interaction, so it doesn't matter much if initialization fails in
slightly less graceful ways if it can't work at all.
At a later point, we will support passthrough with ao_pulse. It seems
the PulseAudio API wants to know the codec type (or maybe not - feeding
it DTS while telling it it's AC3 works), add separate formats for each
codecs. While this reminds of the earlier chaos, it's stricter, and most
code just uses AF_FORMAT_IS_IEC61937().
Also, modify AF_FORMAT_TYPE_MASK (renamed from AF_FORMAT_POINT_MASK) to
include special formats, so that it always describes the fundamental
sample format type. This also ensures valid AF formats are never 0 (this
was probably broken in one of the earlier commits from today).
E.g. --loop-file=2 will play the file 3 times (one time normally, and 2
repeats).
Minor syntax issue: "--loop-file 5" won't work, you have to use
"--loop-file=5". This is because "--loop-file" still has to work for
compatibility, so the "old" syntax with a space between option name and
value can't work.
We generally want 2 things:
1. minimal wakeups for decoding each frame
2. minimal number of frames decoded on continuous seeking
Commit 35810cb8 changed this a bit, and fixed 1. But it broke 2., and
now it decodes 2 frames instead of 1 when you keep seeking (arrow key
held down or such). This made seeking appear slower.
Fix this by making the logic more explicit. In particular, call the
filters only if we actually try to get a new frame.
When playing with --no-audio and all other distractions disabled (like
OSC), it still wakes up 2 times per frame - but the second time is
merely because the VO didn't accept the new frame yet.
Be less annoying, print the actual OSD level instead of something
meaningless, but still clear the OSD if OSD level 0 (no OSD) is set.
Remove the special handling for terminal OSD, that was just dumb.
This means that if a property not listed in property_osd_display[] is
changed, it will be shown on the OSD as "name: ${name}".
Properties that are listed in property_osd_display[] and have osd_name
not set stay invisible by default. This is used for "pause" and
"fullscreen", which (like before this commit) are not shown by default,
because it would be annoying.
The defaults still can be changed with command prefixes (osd-msg,
no-osd, others).
Probably not many user-visible changes. One notable change is that the
terminal OSD code for OSD bar fallback handling is removed with no
replacement. Instead, terminal OSD gets the same text message as normal
OSD. For volume, this is ok, because the text message is reasonable.
Other properties will look worse, but could be adjusted, and there are
in fact no other such properties that would be useful in audio-only
mode.
The fallback message for seeking falls away as well, but that message
was useless anyway - the terminal status line provides all information
anyway.
I believe the show_property_osd() code is now much easier to follow.
If no VO was open, these options couldn't be changed or even queried.
Although these properties are nearly useless if no VO exists, there's
actually no good reason to forbid querying or setting them. Also, even
if the VO is created, it doesn't mean the VO window was created.
Why bother?
Also, since now some properties could be mapped to non-existing options,
but mp_property_generic_option() is used, deal with this case and return
a not-found error code.
If there's a command that uses the OSD by default, then always print the
associated message (or a fallback made of name + value), even if the
command has an associated OSD bar.
This means volume, gamma, panscan, etc. all show both a message and a
OSD bar.
Also, add a '%' to the volume message. The extra_msg thing is not needed
anymore.
See issue #1103.
It's just confusing; users are encouraged to edit input.conf instead
(changing the argument to the "add" command).
Update input.conf to keep the old behavior.
Follow up to previous commit.
This is probably confusing from a user point of view, since this field
shouldn't show up normally anymore. (Before this commit, it could show
up sporadically when a slow operation was performed during playback,
such as switching fullscreen.)
Normally, feeding a packet to the decoder should always return a frame
_if_ we received a frame before. So while we can't know exactly whether
a frame was dropped, at least the normal case is easily detectable.
This means we display something closer to the actual framedrop count,
instead of a bad guess.
This is the "old" framedropping mode (derived from MPlayer). At least in
the mplayer2/mpv source base, it stopped working properly years ago (or
maybe it never worked properly). For one, it depends on the video
framerate, which assume constant framerate. Another problem was that it
could lead to freezing video display: video could get so much behind
that it couldn't recover from framedrop.
Make some small changes to improve this.
Don't use the current audio position to check how much we are behind.
Instead, use the last known A/V difference. last_av_difference is
updated only when a video frame is scheduled for display. This means we
can keep stop dropping once we're done catching up, even if video is
technically still behind. What helps us here that this forces a video
frame to be displayed after a while. Likewise, we reset the
dropped_frames count only when scheduling a new frame for display as
well.
Some inspiration was taken from earlier work by xnor (see issue #620),
although the implementation turned out quite different.
This still uses the demuxer-reported (possibly broken) FPS value. It
also doesn't account for filters changing FPS. We can't do much about
this, because without decoding _and_ filtering, we just can't know how
long a frame is. In theory, you could derive that from the raw packet
timestamps and the filter chain contents, but actually doing this is
too involved. Fortunately, the main thing the FPS affects is actually
the displayed framedrop count.
Rename video_decode_and_filter to video_filter, and add a new
video_decode_and_filter function. This function now calls the decoder.
This is done so that we can check filters a second time after decoding,
which avoids a useless playloop iteration.
(This and the previous commits are really just microoptimizations, which
simply reduce the number of times the playloop has to recheck
everything.)
Move the check to a function. Run the check a second time after
decoding/filtering. This second check is strictly speaking redundant
(which is why it wasn't done until now), but it avoids a useless
playloop iteration.
Move this code below the code that "shifts" the newly filtered frame.
This allows us to skip a useless playloop iteration later, because
obviously we need to filter a new frame after the previous frame has
been "shifted", and not before that.
Until now, you could override only level 3 with --osd-status-msg. Extend
this, add add --osd-msg1 to --osd-msg3 (one for each OSD level). OSD
level 0 always means disable OSD, so that isn't included.
--osd-msg3 corresponds to --osd-status-msg, but they're not exactly the
same. To allow more customization, --osd-msgN do not include the OSD
symbol. The symbol can be manually added with "${osd-sym-cc}". We keep
the "old" option for some short-term compatibility.
--osd-msg1 should be particularly useful; for example you could do:
--osd-msg1='${?pause==yes:${osd-sym-cc}}'
to display a "paused" symbol when paused, and nothing during normal
playback. (Although admittedly, the syntax is quite a bit of work.)
We don't allow this by default, because it would be silly if random
external data (like filenames or file tags) could accidentally trigger
them.
Add a property that magically disables this ASS tag escaping.
Note that malicious input could still disable ASS tag escaping by
itself. This would be annoying but harmless.
If --write-filename-in-watch-later-config is used, and the filename
contains newline characters (as generally allowed on Unix), then the
newline will be written to the resume file literally, and the parts
after the newline character are interpreted as options.
This is possibly security relevant.
Change newline characters (and in fact any other special characters)
to '_'.
Reported as #1099 (this commit is a reimplementation of the proposed
pull request).
CC: @mpv-player/stable
Merges pull request #1094, with some minor changes. mpv expects IEEE,
and IEEE allows divisions by 0 for floats, so these shouldn't actually
be a problem, but do it anyway for the sake of clang.
Signed-off-by: wm4 <wm4@nowhere>
If you send the "quit" or "stop" command with the client API, it will
now attempt to kill network I/O immediately (same as normal input in the
previous commits).
This mechanism originates from MPlayer's way of dealing with blocking
network, but it's still useful. On opening and closing, mpv waits for
network synchronously, and also some obscure commands and use-cases can
lead to such blocking. In these situations, the stream is asynchronously
forced to stop by "interrupting" it.
The old design interrupting I/O was a bit broken: polling with a
callback, instead of actively interrupting it. Change the direction of
this. There is no callback anymore, and the player calls
mp_cancel_trigger() to force the stream to return.
libavformat (via stream_lavf.c) has the old broken design, and fixing it
would require fixing libavformat, which won't happen so quickly. So we
have to keep that part. But everything above the stream layer is
prepared for a better design, and more sophisticated methods than
mp_cancel_test() could be easily introduced.
There's still one problem: commands are still run in the central
playback loop, which we assume can block on I/O in the worst case.
That's not a problem yet, because we simply mark some commands as being
able to stop playback of the current file ("quit" etc.), so input.c
could abort playback as soon as such a command is queued. But there are
also commands abort playback only conditionally, and the logic for that
is in the playback core and thus "unreachable". For example,
"playlist_next" aborts playback only if there's a next file. We don't
want it to always abort playback.
As a quite ugly hack, abort playback only if at least 2 abort commands
are queued - this pretty much happens only if the core is frozen and
doesn't react to input.
Idle mode went to sleep too early, e.g. just pressing "ESC" did nothing,
until the next event happened. This was because it directly went to
sleep after processing commands. What we should do instead is rechecking
all state after processing commands, redraw OSD, and then go to sleep.
This also fixes some strange OSD-related behavior.
Also move some other code around to separate idle mode initialization
from the normal run loop.
Do terminal input with a thread, instead of using the central select()
loop. This also changes some details how SIGTERM is handled.
Part of my crusade against mp_input_add_fd().
The purpose is making accessing the current playlist entry saner when
commands are executed during initialization, termination, or after
playlist navigation commands.
For example, the "playlist_remove current" command will invalidate
playlist->current - but some things still access the playlist entry even
on uninit. Until now, checking stop_play implicitly took care of it, so
it worked, but it was still messy.
Introduce the mpctx->playing field, which points to the current playlist
entry, even if the entry was removed and/or the playlist's current entry
was moved (e.g. due to playlist navigation).
Continues commit 348dfd93. Replace other places where input was manually
fetched with common code.
demux_was_interrupted() was a weird function; I'm not entirely sure
about its original purpose, but now we can just replace it with simpler
code as well. One difference is that we always look at the command
queue, rather than just when cache initialization failed. Also, instead
of discarding all but quit/playlist commands (aka abort command), run
all commands. This could possibly lead to unwanted side-effects, like
just ignoring commands that have no effect (consider pressing 'f' for
fullscreen right on start: since the window is not created yet, it would
get discarded). But playlist navigation still works as intended, and
some if not all these problems already existed before that in some
forms, so it should be ok.
This makes the player wait until each script is loaded. Do this to give
the script a chance to setup all its event handlers. It might also be
useful to allow a script to change options that matter for playback.
While waiting for a script to be loaded, the player actually accepts
input. This is needed because the scripts can execute player commands
anyway while they are being "loaded". The player won't react to most
commands though: it can't quit or navigate the playlist in this state.
For deciding whether a script is finally loaded, we use a cheap hack: if
mpv_wait_event() is called, it's considered loaded. Let's hope this is
good enough. I think it's better than introducing explicit API for this.
Although I'm sure this will turn out as too simplistic some time in the
future, the same would probably happen with a more explicit API.
Expose the central event handling functions explicitly, so that other
parts of the player can use them.
No functional changes. Preparation for the next commit.
With e.g --start=-3 --audio-buffer=10 the decoder entered EOF state
before the initial sync was finished, entered STATUS_EOF, and just
started playing audio from a random position.
This doesn't handle seeking outside of the file, which is a different
case. E.g. --start=30:00 with audio and video enabled in a file shorter
than 30:00 will play a random last part of audio. This could perhaps be
fixed by using the hr-seek target for cutting audio, instead of the
video PTS, but that would be kind of intrusive, so don't do it for now.
The simpler solution, assuming audio EOF on video EOF, wouldn't work,
because we allow audio to start before video, or to last after video.
This is a deadlock caused by a lock order issue: sub/osd.c locks the OSD
first, then the subtitle decoder lock. player/sub.c does the reverse.
Fix this by discussing away the requirement for locking (see below),
which allows us to drop the broken sub lock. sub_get_text() still
acquires and releases the sub decoder lock, but it's not held at the
same time as the OSD lock anymore, so it should be fine.
Originally, the sub lock was acquired because sub_get_text() returns a
pointer to a mutable string. We simply declare that it's ok to call it
unlocked, as long as only 1 thread accesses it, which works out fine in
this case.
Somehow, there was a larger misunderstanding in the code: ao_buffer
does not need to be preserved over audio reinit for proper support of
gapless audio. The actual AO internal buffer takes care of this.
In fact, preserving ao_buffer just breaks audio resync. In the ordered
chapter case, end_pts is used, which means not all audio data in the
buffer is played, thus some data is left over when audio decoding
resumes on the next segment. This triggers some code that aborts resync
if there's "audio decoded" (ao_buffer contains something), but no PTS
is known (nothing was actually decoded yet).
Simplify, and always bind the output buffer to the decoder.
CC: @mpv-player/stable (maybe)
--hls-bitrate=min/max lets you select the min or max bitrate. That's it.
Something more sophisticated might be possible, but is probably not even
worth the effort.
Because that might be a bad idea.
Note that remote playlists still can use any protocol marked with
is_safe and is_network, because the case of http-hosted playlists
containing URLs using other streaming protocols is not unusual.
The event was copied early, and wasn't released if it was rejected
instead of being added to the event queue. Fix by copying the event at a
point when it's certainly added to the event queue.
The dup_event_data() function is merely moved.
Until now, you had to use --load-unsafe-playlists or --playlist to get
playlists loaded. Change this and always load playlists by default.
This still attempts to reject unsafe URLs. For example, trying to invoke
libavdevice pseudo-demuxer is explicitly prevented. Local paths and any
http links (and some more) are always allowed.
Probably no observable effect, but it's more correct. Setting audio to
EOF could have bad effects otherwise (anywhere the player logic for
example decides whether EOF was reached, and such).
This inserts an automatic conversion filter if a Matroska file is marked
as 3D (StereoMode element). The basic idea is similar to video rotation
and colorspace handling: the 3D mode is added as a property to the video
params. Depending on this property, a video filter can be inserted.
As of this commit, extending mp_image_params is actually completely
unnecessary - but the idea is that it will make it easier to integrate
with VOs supporting stereo 3D mogrification. Although vo_opengl does
support some stereo rendering, it didn't support the mode my sample file
used, so I'll leave that part for later.
Not that most mappings from Matroska mode to vf_stereo3d mode are
probably wrong, and some are missing.
Assuming that Matroska modes, and vf_stereo3d in modes, and out modes
are all the same might be an oversimplification - we'll see.
See issue #1045.
The player didn't quit when seeking past EOF in audio-only mode while
paused. The only case when we don't want to quit is when the last video
frame is displayed while paused.
This logic was probably broken a while ago, but I'm not exactly sure.
CC: @mpv-player/stable
bstr.c doesn't really deserve its own directory, and compat had just
a few files, most of which may as well be in osdep. There isn't really
any justification for these extra directories, so get rid of them.
The compat/libav.h was empty - just delete it. We changed our approach
to API compatibility, and will likely not need it anymore.
Add a mechanism to the client API code, which allows the player core to
query whether a client API event is needed at all. Use it for the cache
update.
In this case, this is probably a pure microoptimization; but the
mechanism will be useful for other things too.
Don't attempt to resync after speed changes. Note that most other cases
of audio reinit (like switching tracks etc.) still resync, but other
code paths take care of setting the audio_status accordingly.
This restores the old behavior of not trying to fix audio desync, which
was probably changed with commit 261506e3.
Note that the code as of now wasn't even entirely correct, since the A/V
sync values are slightly shifted. The dsync depends on the audio buffer
size, so a larger buffer size will show more extreme desync. Also see
mplayer2 commit 213a224e, which should fixed this - it was not merged
into mpv, because it disabled audio for too long, resulting in a worse
user experience. This is similar to the issue this commit attempts to
fix.
Fixes: #1042 (probably)
CC: @mpv-player-stable
Remove the hardcoded wait time of 2 seconds. Instead, adjust the wait
time each time we unpause: if downloading the data took longer than its
estimated playback time, increase the amount of data we wait for. If
it's shorter, decrease it.
The +/- is supposed to avoid oscillating between two values if the
elapsed time and the wait time are similar. It's not sure if this
actually helps with anything, but it can't harm.
Use the "native" underrun detection, instead of guessing by a low cache
duration. The new underrun detection (which was added with the original
commit) might have the problem that it's easy for the playloop to miss
the underrun event. The underrun is actually not stored as state, so if
the demuxer thread adds a new packet before the playloop happens to see
the state, it's as if it never happened. On the other hand, this means
that network was fast enough, so it should be just fine.
Also, should it happen that we don't know the cached range (the
ts_duration < 0 case), just wait until the demuxer goes idle (i.e.
read_packet() decides to stop). This pretty much should affect broken or
unusual files only, and there might be various things that could go
wrong. But it's more robust in the normal case: this situation also
happens when no packets have been read yet, and we don't want to
consider this as reason to resume playback.
The cache percentage was useless. It showed how much of the total stream
cache was in use, but since the cache size is something huge and
unrelated to the bitrate or network speed, the information content of
the percentage was rather low.
Replace this with printing the duration of the demuxer-cached data, and
the size of the stream cache in KB.
I'm not completely sure about the formatting; suggestions are welcome.
Note that it's not easy to know how much playback time the stream cache
covers, so it's always in bytes.
The "buffering" logic was active even if the stream cache was disabled.
This is contrary to what the manpage says. It also breaks playback
because of another bug: the demuxer cache is smaller than 2 seconds,
and thus the resume condition never becomes true.
Explicitly run this code only if the stream cache is enabled. Also, fix
the underlying problem of the breakage, and resume when the demuxer
thread stops reading in any case, not just on EOF.
Broken by previous commit. Unbreaks playback of local files.
Add the --cache-secs option, which literally overrides the value of
--demuxer-readahead-secs if the stream cache is active. The default
value is very high (10 seconds), which means it can act as network
cache.
Remove the old behavior of trying to pause once the byte cache runs
low. Instead, do something similar wit the demuxer cache. The nice
thing is that we can guess how many seconds of video it has cached,
and we can make better decisions. But for now, apply a relatively
naive heuristic: if the cache is below 0.5 secs, pause, and wait
until at least 2 secs are available.
Note that due to timestamp reordering, the estimated cached duration
of video might be inaccurate, depending on the file format. If the
file format has DTS, it's easy, otherwise the duration will seemingly
jump back and forth.
When video format changes, the frame before the frame with the new
format sets video_status briefly to STATUS_DRAINING. This caused the
code to handle the EOF case to kick in, which just pauses the player
when trying to step past the last frame. As a result, trying to
framestep over format changes resulted in pausing the player.
Fix by testing against the correct status.
This shouldn't change anything functionally.
Change the A/V desync message. --framedrop is enabled by default now, so
the text must be changed a little. I've never heard of audio outputs
messing up A/V sync recently, so remove that part.
Remove the unused ao_pts field.
Reorder 2 A/V sync related expressions so that they look the same.
Commit 846257da introduced an accidental feature: if you kept seeking
(so playback never really resumes), the audio would never be played.
This was nice, but commit 4c25b000 accidentally removed it again (due
to the video_next_pts being earlier available than it used to be, so
audio could be played before the player executed the next queued seek).
Implicitly reintroduce the old behavior again by not decoding a second
video frame immediately. Usually, the second frame is used to compute
the frame duration needed to for accurate framedropping, but since the
first frame after a seek is never dropped, we don't need this.
Now the video code will queue the new frame to the VO immediately, and
since fill_audio_out_buffers() is called in the playloop before
write_video() and execute_queued_seek(), it never gets the chance to
enter STATUS_READY, and seeks will be silent.
This also has a nice side-effect: since the second frame is not decoded
and filtered, seeking becomes slightly faster (back to the same level
as with framedrop disabled).
It seems this still sometimes plays a period of audio when keeping a
seek key down. In my tests, this appeared to happen because the seek
finished before the next key repeat was sent.
In theory, timestamps can be negative, so we shouldn't just return -1
as special value.
Remove the separate code for clearing decode buffers; use the same code
that is used for normal seek reset.
Commit 5afc025c broke this. The reason is that mpctx->delay is updated
when a new video frame is added. This value is also needed to resync
audio, but it will be for the wrong PTS. They must be consistent with
each other, and if they aren't, initial sync will be off by N video
frames, which results at least in worse user experience.
This can be reproduced by for example heavily switching between normal
and 2x speed, or similar.
Fix by readding the video_next_pts field (keeping its use minimal,
instead of reverting the commit that removed it).
If video reaches EOF, and audio is also EOF (or is otherwise not
meaningful, like audio disabled), then the playback position was briefly
set to 0. Fix this by not trying to use a bogus audio PTS.
CC: @mpv-player/stable (maybe)
This simplifies the code, and fixes an odd bug: the second-last frame
was displayed for a very short duration if framedrop was enabled. The
reason was that basically the time difference between second-last and
last frame were skipped, because at this point EOF was already
signaled. Also see commit b0959488 for a similar issue in the
same code.
This removes the messiness of the next_frame 2-frame queue, and
strictly runs the "new frame" code when a frame is moved to the first
position of the queue, instead of somehow messing with return codes.
This also merges update_video() into video_output_image().
No functional changes. init_vo() is now needed a bit further down, and
moving it keeps definition and use close. adjust_sync() will be used by
a function further up in one of the following commits.
After a new file is loaded, playback never starts instantly. Rather, it
takes some playloop iterations until initial audio and video have been
decoded, and the outputs have been (lazily) initialized. This means you
will get status line updates between the messages that inform of the
initialized outputs. This is a bit annoying and clutters the terminal
output needlessly.
Fix this by never printing the status line before playback isn't fully
initialized. Do this by reusing the --term-playing-msg code (which
prints a message once playback is initialized). This also makes sure the
status line _is_ shown during playback restart when doing seeks.
It's possible that the change will make the output more confusing if for
some reason is stuck forever initializing playback, but that seems like
an obscure corner case that never happens, so forget about it.
print_status() is called at a later point anyway (and before sleeping),
so this code has little effect. This code was added in commit a4f7a3df5,
and I can't observe any problems with idle mode anymore.
Now print_status() is called from a single place only, within osd.c.
This mostly uses the same idea as with vo_vdpau.c, but much simplified.
On X11, it tries to get the display framerate with XF86VM, and limits
the frequency of new video frames against it. Note that this is an old
extension, and is confirmed not to work correctly with multi-monitor
setups. But we're using it because it was already around (it is also
used by vo_vdpau).
This attempts to predict the next vsync event by using the time of the
last frame and the display FPS. Even if that goes completely wrong,
the results are still relatively good.
On other systems, or if the X11 code doesn't return a display FPS, a
framerate of 1000 is assumed. This is infinite for all practical
purposes, and means that only frames which are definitely too late are
dropped. This probably has worse results, but is still useful.
"--framedrop=yes" is basically replaced with "--framedrop=decoder". The
old framedropping mode is kept around, and should perhaps be improved.
Dropping on the decoder level is still useful if decoding itself is too
slow.
Apparently users prefer this behavior.
It was used for subtitles too, so move the code to calculate the video
offset into a separate function. Seeking also needs to be fixed.
Fixes#1018.
The OSD is marked as changed, but the core isn't notified and this
doesn't immediately wakeup. (Possibly the OSD code should wakeup the
core instead, but maybe that woudl be overkill.)
Observed when using "mp.use_suspend = false" in the OSC, and pausing,
and moving the mouse pointer out of the window. The last part of the
fade remained visible for longer than intended.
Code reorganized to make layouts exchangeable
alternative test layout can be tested with
layout=slimbox
in the OSC config
timers are now used to properly animate the fade out when the
player is paused
duplicate seeks are discarded again
sub_reset() was called on cycling subtitle tracks and on seeking. Since
we don't want that subtitles disppear on cycling, sd_lavc.c didn't clear
its internal subtitle queue on reset, which meant that seeking with PGS
subtitles could leave the subtitle on screen (PGS subtitles usually
don't have a duration set).
Call it only on seeking, so we can also strictly clear the subtitle
queue in sd_lavc.
(This still can go very wrong if you disable a subtitle, seek, and
enable it again - for example, if used with libavformat that uses "SSA"
style demuxed ASS subtitle packets. That shouldn't happen with newer
libavformat versions, and the user can "correct" it anyway by executing
a seek while the subtitle is selected.)
Until recently, vo_opengl could be accessed from a single thread only,
due to the OpenGL API context being thread-specific. This issue doesn't
exist anymore, because VOs run on their own thread. This means we can
simply lock/unlock the playloop instead of doing something complicated
to get the playloop thread to execute our code.
This ran adjust_sync() on every playloop iteration, instead of every
newly decoded frame. It seems this was idempotent in the common case,
but the code was originally designed to be run once only, so restore
that.
The previous commit broke these things, and fixing them is separate in
this commit in order to reduce the volume of changes.
Move the image queue from the VO to the playback core. The image queue
is a remnant of the old way how vdpau was implemented, and increasingly
became more and more an artifact. In the end, it did only one thing:
computing the duration of the current frame. This was done by taking the
PTS difference between the current and the future frame. We keep this,
but by moving it out of the VO, we don't have to special-case format
changes anymore. This simplifies the code a lot.
Since we need the queue to compute the duration only, a queue size
larger than 2 makes no sense, and we can hardcode that.
Also change how the last frame is handled. The last frame is a bit of a
problem, because video timing works by showing one frame after another,
which makes it a special case. Make the VO provide a function to notify
us when the frame is done, instead. The frame duration is used for that.
This is not perfect. For example, changing playback speed during the
last frame doesn't update the end time. Pausing will not stop the clock
that times the last frame. But I don't think this matters for such a
corner case.
The VO is run inside its own thread. It also does most of video timing.
The playloop hands the image data and a realtime timestamp to the VO,
and the VO does the rest.
In particular, this allows the playloop to do other things, instead of
blocking for video redraw. But if anything accesses the VO during video
timing, it will block.
This also fixes vo_sdl.c event handling; but that is only a side-effect,
since reimplementing the broken way would require more effort.
Also drop --softsleep. In theory, this option helps if the kernel's
sleeping mechanism is too inaccurate for video timing. In practice, I
haven't ever encountered a situation where it helps, and it just burns
CPU cycles. On the other hand it's probably actively harmful, because
it prevents the libavcodec decoder threads from doing real work.
Side note:
Originally, I intended that multiple frames can be queued to the VO. But
this is not done, due to problems with OSD and other certain features.
OSD in particular is simply designed in a way that it can be neither
timed nor copied, so you do have to render it into the video frame
before you can draw the next frame. (Subtitles have no such restriction.
sd_lavc was even updated to fix this.) It seems the right solution to
queuing multiple VO frames is rendering on VO-backed framebuffers, like
vo_vdpau.c does. This requires VO driver support, and is out of scope
of this commit.
As consequence, the VO has a queue size of 1. The existing video queue
is just needed to compute frame duration, and will be moved out in the
next commit.
