- Don't define _GNU_SOURCE on Windows, no need
- Define WIN32_LEAN_AND_MEAN to strip some unneded headers from
windows.h
- Define NOMINMAX and _USE_MATH_DEFINES as they are common for Windows
headers
We prefer to fail fast rather than degrade in unpredictable ways.
The example in sub/ is particularly egregious because the code just
skips the work it's meant to do when an allocation fails.
I'd like some names to be more descriptive, but to work with 15 chars
limit we have to make some sacrifice.
Also because of the limit, remove the `mpv/` prefix and prioritize
actuall thread name.
It was found that this causes issues with at least ao_coreaudio,
essentially revealing a way bigger issue:
Some AOs don't check for 0 and/or have no way to deal with short writes.
Someone will have to figure out a fix later but get rid of the direct
cause for now.
This reverts commit ae908a70ce.
ao_read_data() is used by pull AOs potentially from threads managed by
external libraries. These threads can be sensitive to blocking.
For example the pipewire ao is using a realtime thread for the
callbacks.
since i was going to fix the include order of stdatomic, might as well
sort the surrouding includes in accordance with the project's coding
style.
some headers can sometime require specific include order. standard
library headers usually don't. but mpv might "hack into" the standard
headers (e.g pthreads) so that complicates things a bit more.
hopefully nothing breaks. if it does, the style guide is to blame.
replace it with <stdatomic.h> and replace the mp_atomic_* typedefs with
explicit _Atomic qualified types.
also add missing config.h includes on some files.
Pull AOs work off of a callback that relies on mpv's internal timer. So
like with the related video changes, convert all of these to nanoseconds
instead. In many cases, the underlying audio API does actually provide
nanosecond resolution as well.
There's a lot of wild 1e6, 1000, etc. lying around in the code. A macro
is much easier to read and understand at a glance. Add some helpers for
this. We don't need to convert everything now but there's some simple
things that can be done so they are included in this commit.
Why a bigger search-interval is required:
scaletempo2 doesn't do a good job when the signal contains frequencies
less then 1/search_interval. With a search interval of 30ms that means
anything below 33.333Hz sounds bad.
Depending on the genre it's very for music to contain frequencies down
to 30Hz, and sometimes even a little bit below that. Therefore a higher
default value is needed to handle such cases.
Based on that an argument can be made for a value of 50, as that should
work down to 20Hz, or something even higher because movies sometimes
have some infrasonic content.
However the downside of big search intervals is increased CPU usage and
intelligibility at higher speeds, as it effectively leads to parts of
the audio being skipped.
A value of 40 can handle frequencies down to 25Hz, enough for all music
except very rare edge cases, while still providing decent
intelligibility.
Why a smaller window-size is required:
Large values reduce intelligibility at high speeds and therefore small
values are preferred.
However when values get too small it starts to sound weird
(similar to librubberband).
In my testing a value of 10 already works well, but adding a small
safety margin seems like a good idea, especially since it made no
noticeable difference to intelligibility, which is why 12 was chosen.
Linux and macOS already use nanosecond resolution for their sleep
functions. It was just being converted from microseconds before. Since
we have mp_time_ns now, go ahead and bump the precision here. The timer
for windows uses some timeBeginPeriod thing which I'm not sure what it
does really but whatever just convert the units to ms like they were
doing before. There's really no reason to keep the mp_sleep_us helper
around. A multiplication by 1000 is trivial and underlying OS clocks
have nanosecond precision.
This is the most supported in standard layout, if we request more it
tends to fallback to stereo instead. Also channels mask is 32-bit and it
can get truncated.
A bit different from the OPT_REPLACED/OPT_REMOVED ones in that the
options still possibly do something but they have a deprecation
message. Most of these are old and have no real usage. The only
potentially controversial ones are the removal of --oaffset and
--ovoffset which were deprecated years ago and seemingly have no real
replacement. There's a cryptic message about --audio-delay but who
knows. The less encoding mode code we have, the better so just chuck
it.
Avoid generating too much audio after EOF.
Note: This often has no effect, because less audio is produced than
required.
Usually this comes to effect with the userspeed filter at high speed
(4x) and going back to 1x speed to remove the filter.
After the final input packet, the filter padded with silence to allow
one more iteration. That was not enough to process the final frames.
Continue padding the end of `input_buffer` with silence until the final
frames have been processed.
Implementation: Instead of padding when adding final samples, pad before
running WSOLA iteration. Count number of added silent frames and
remaining input frames for time keeping.
This changes the emitted pts values from the start of the search block
to the center of the search block. Change initial `output_time`
accordingly. Initial `search_block_index` is irrelevant, because it's
overwritten before the first iteration.
Using the `output_time` removes the rounding of `search_block_index`,
which also fixes the <20 microsecond gaps in timestamps between output
packets.
Rationale:
The variance in audio position was in the range `0..search-interval`.
With this change, the range is
(- search-interval / 2)..(search-interval / 2)`
which ensures lower maximum offset.
Target block can be anywhere in the previous search-block, varying by
`search-interval` while the filter is active. This resulted in constant
audio offset when returning to 1x playback speed.
- Move the search block to the target block to sync up exactly.
- Drop old frames to minimize input_buffer usage.
The internal time update function involved multiple problems:
- Time was updated after WSOLA iteration. The means speed was updated
one iteration later than it could be.
- The update functions caused spikes of too many or too few samples
advanced, leading to audio glitches on speed changes.
- The inconsistent updates made it very difficult to produce gapless
audio packets.
- The `output_time` update function involved complicated feedback:
`search_block_index` influenced how many frames from `input_buffer`
are retained, which influenced how much `output_time` is changed,
which influenced `search_block_index`.
With these changes:
- Time is updated before WSOLA iterations. Speed changes are effective
instantly.
- There are no spikes in playback speed during speed changes.
- No significant gaps are introduced in output packets.
- The time update function becomes (function calls omitted for brevity)
output_time += ola_hop_size * playback_rate
Functions received a `playback_rate` parameter to check how many samples
are needed before iteration. Internal state is only updated when the
iteration is actually run, so the speed is allowed to change until
enough data is received.
The first WSOLA iteration overlapped audio with whatever was in the
`wsola_output` buffer. This was either silence (if not run before), or
old frames (if switching to 1x and back to a different speed).
Track the state of the output buffer and memcpy the whole window for the
first iteration instead.
`output_time` is used to set the center of the search block. Init of
both `search_block_index` and `output_time` with 0 caused inconsistent
search block movement for the first iterations.
Initialize with `search_block_center_offset` for consistency with initial
`search_block_index`.
Fixes#12028
There was an additional issue that audio was always delayed by half the
configured search-interval. This was caused by the `out` buffer length
not being included in the delay calculation.
Notes:
- Every WSOLA iteration advances the input buffer by _some amount_, and
produces data in the output buffer always of size `ola_hop_size`.
- `mp_scaletempo2_fill_buffer` is always called with `ola_hop_size`
- Thus, the rendered frames are always cleared immediately after
processing, and `num_complete_frames` is 0 in the delay calculation.
- The factors contributing to delay are:
- the pending samples in the input buffer according to the search
block position, and
- the pending rendered samples in the output buffer (always empty in
practice).
The frame_delay code looked like that of the rubberband filter, which
might not work for scaletempo2. Sometimes a different amount of input
audio was consumed by scaletempo2 than expected. It may have been caused
by speed changes being a more dynamic process in scaletempo2. This can
be seen by where `playback_rate` is used in `run_one_wsola_iteration`:
`playback_rate` is only referenced after the iteration, when updating
the time and removing old data from buffers.
In scaletempo2, the playback speed is applied by changing the amount the
search block is moved. That apparently averages out correctly at
constant playback speed, but when the speed changes, the error in this
assumption probably spikes. This error accumulated across all speed
changes because of the persistent `frame_delay` value.
With the removal of the persistent `frame_delay`, there should be no way
for the audio to drift off. By deriving the delay from filter buffer
positions, and the buffers are filled only as much as needed, the delay
always stays within buffer bounds.
The idea behind period_size is that it's the minimum amount of data
that your audio subsystem wants to fetch.
For PulseAudio, this is given by the minreq buffer attribute, which
is in bytes for all channels. Hence the divisions.
This change sets the device_buffer member of the ao struct for
the JACK ao to whatever value we read during init.
While JACK allows changing the audio buffer size on-the-fly
(you can do this for example through DBus), having it somehow
reconfigure the entire audio buffer logic of mpv that depends
on device_buffer in some way when that happens only leads to
audio dropout and weird code. device_buffer's role is mostly for
prebuffer from what I understand, which means that simply setting
it to its current value in the init function is fine.
In addition to the patch, appropriate additions to the mpv.conf file
will of course be needed for this to work. For example on my system:
audio-device=oss//dev/dsp4
audio-spdif=ac3,dts,dts-hd,eac3,truehd
This has been tested using recent FreeBSD-13.1-stable, using external
surround processors (both a Trinnov Altitude 16 and an LG OLED that
supports Dolby Atmos, and connected with HDMI to an NVidia RTX 2070).
Author and tester: David G Lawrence <dg1007@dglawrence.com>
There's an edge cause with gapless audio and pausing. Since, gapless
audio works by sending an EOF immediately, it's possible to pause on the
next file before audio actually finishes playing and thus the sound gets
cut off. The fix is to simply just always do an ao_drain if the ao is
about to set a pause on EOF and we still have audio playing.
Fixes#8898.
The difference this makes is that the OS API will notice
when we underrun (as opposed to being fed silence).
Other AOs mostly seem to not do this because they've committed
to filling a buffer of a certain size no matter what, but I have
not observed any ill effects for AudioTrack in my testing.
This looks like a pretty bad bug but only became a problem
with the last commit that allows rates like 22.5kHz to pass through
directly instead of being resampled.
wireplumber uses the media role when the node is first created.
To have the property available at this point reliably we need to set it
directly when creating the stream/node.
As the role property is interpreted by wireplumber it can only be
evaluated when creating the stream. The current, dynamic mechanism is
racy so revert it.
See: #11253Fixes: #12041
This reverts commit 535cd6f313.
The avpkt was created once on decoder init but destroyed each time the
filter was destroyed, this obviously can't work. Move the packet alloc
to the filter init function instead.
fixes: 4574dd5dc6
Doing a pw_thread_loop_wait() without checking conditions is invalid.
The thread loop could be signalled for other reasons and in this case
the wait needs to continue.
PipeWire added such additional signaling in
commit 33be898130f0 ("thread-loop: signal when started").
This meant that for_each_sink would return before the callbacks have
fired and session_has_sink() would incorrectly return "false", failing
the initialization of ao_pipewire.
Fixes#11995
Now that Debian 12 is release bump the minium required version to what
is provided in Ubuntu Jammy (22.04).
The same as has been done for the wayland dependencies.
Signed-off-by: Thomas Weißschuh <thomas@t-8ch.de>
Instead of brute forcing the name until it is set, without any error
checking and expecting it would start to work, fallback to client name
if initial request fails.
Fixes player going into infinite loop with very long title names. The
API rejects unreasonably long names, which make sense.
As for alleged "weird race condition in the IAudioSessionControl itself"
I cannot comment. It works on my end and even if it fails, it is not a
critical error or even something that we should care about... and
obviously not hang the whole player for that.
Fixes: #11803
By making the data thread realtime it is able to serve requests faster
and more reliable reducing crackling in certain situations.
As the mpv callbacks that are running on the data thread are all
non-blocking and very short this should be safe.
The same mechanism is also used by pw-cat and the alsa plugin shipped by
pipewire.
c784820454 introduced a bool option type
as a replacement for the flag type, but didn't actually transition and
remove the flag type because it would have been too much mundane work.
Older versions of pipewire segfault when calling spa_hook_remove() on
hooks that are zeroed.
Add a backfill for the logic added by pipewire 0.3.57.
Being able to remove zeroed hooks makes errorhandling much easier.
See #11309
PipeWire supports a global volume control for streams that works on top
of the per-channel volumes.
As mpv only supports a single volume with ao-volume it can make sense to
use the single global volume from PipeWire for it.
This allows the user to also specify per-channel volumes and not have
mpv trample over them.
This mode is not the default as pulseaudio does not support this
global volume control and all tooling controlling PipeWire via
pipewire-pulse (like pavucontrol) will not be able to see this channel.
As ao_pipewire is probed first if a user does not have PipeWire running
they will see a scary warning message even if another AO afterwards is
probed fine.
Tone down the error message so as not to confuse users.
ao-volume is represented in the code with a `struct ao_control_vol_t`
which contains volumes for two channels, left and right.
However the code implementing this property in command.c never treats
these values individually. They are always averaged together.
On the other hand the code in the AOs handling these values also has to
handle the case where *not* exactly two channels are handled.
So let's remove the `struct ao_control_vol_t` and replace it with a
simple float.
This makes the semantics clear to AO authors and allows us to drop some code from the AOs and command.c.
In debug mode the macro causes an assertion failure.
In release mode it works differently and tells the compiler that it can
assume the codepath will never execute. For this reason I was conversative
in replacing it, e.g. in mpv-internal code that exhausts all valid values
of an enum or when a condition is clear from directly preceding code.
[motivation]
Seeking on MacOS appears to be lagged when users connect
to wireless audio output (airpods for example).
This commit attempts to fixmpv-player/mpv#10270
[observation]
1. When using other media player (VLC to be exact) simultaneously,
the lagging on seek disappear. We could guess that the AudioDevice
is on some sort of "warm-up" state.
See mpv-player/mpv#9243 for detailed description.
2. `AudioOutputUnitStart` takes significant longer time after each seek
or pause/play when using wireless output devices compares to wired devices.
[rationale]
After investigate codes in ao_coreaudio.c, it appears that the the `stop`
function was used as `ao_driver.reset` function. Therefore every seek
and pause would call `AudioOutputUnitStop`.
It turns out that `ao_driver.reset` function is used in `ao_reset`.
And `ao_reset` function is used to clean up the state of current `ao`
so I think `AudioUnitReset` is more proper than `AudioOutputUnitStop`
under this semantics.
Since ao_coreaudio use pull base mechanism, audio playback behaviors
upon pause/seek could be handled by callback function
(streaming silence when paused) so there is no need to stop AudioUnit when resetting.
Therefore using `AudioUnitReset` as `ao_driver.reset` looks proper.
Additionally, after using proper reset, the AudioUnit that represents
hardware I/O devices doesn't need to be restart everytime seek/pause actions happen.
Restarting wireless devices simply takes longer in MacOS which is
the root cause of lagging observed by users when they seek or pause/play media.
[method]
Use `AudioUnitReset` for ao_driver.reset.
When a pull AO reaches reaches EOF then ao_read_data() will set
p->playing = false.
Because the ao is marked as not playing ao_set_pause(true) will not
reset the AO.
This keeps the output stream unintentionally open.
Fixes#9835
This has been a long standing annoyance - ffmpeg is removing
sizeof(AVPacket) from the API which means you cannot stack-allocate
AVPacket anymore. However, that is something we take advantage of
because we use short-lived AVPackets to bridge from native mpv packets
in our main decoding paths.
We don't think that switching these to `av_packet_alloc` is desirable,
given the cost of heap allocation, so this change takes a different
approach - allocating a single packet in the relevant context and
reusing it over and over.
That's fairly straight-forward, with the main caveat being that
re-initialising the packet is unintuitive. There is no function that
does exactly what we need (what `av_init_packet` did). The closest is
`av_packet_unref`, which additionally frees buffers and side-data.
However, we don't copy those things - we just assign them in from our
own packet, so we have to explicitly clear the pointers before calling
`av_packet_unref`. But at least we can make a wrapper function for
that.
The weirdest part of the change is the handling of the vtt subtitle
conversion. This requires two packets, so I had to pre-allocate two in
the context struct. That sounds excessive, but if allocating the
primary packet is too expensive, then allocating the secondary one for
vtt subtitles must also be too expensive.
This change is not conditional as heap allocated AVPackets were
available for years and years before the deprecation.
This allows us to more easily see the datapath from mpv to pipewire.
We know how often the callbacks are triggered, how big the buffers are
and how much data mpv provides to pipewire.
This allows the core of mpv to know about issues in the AO.
Otherwise playback will just freeze as no more data callbacks are sent
by PipeWire.
Also it allows mpv to try to reconnect the AO or find another, working
AO.
We want to add more logic to the stream event handler.
This logic should not be triggered during normal stream shutdown, so we
remove the listener beforehand.
The AO is feature-complete now.
As PipeWire also provides compatibility with PulseAudio, ALSA and Jack
we should put it before those for the autodetection to work.
The pure presence of PipeWire does not mean that it is actually driving
the audio session. For example it could only be meant for video.
Currently there is no proper API to detect this (see [0]), so we check
for the presence of audio sinks.
As soon as a proper API exists, we should use that.
[0] https://gitlab.freedesktop.org/pipewire/pipewire/-/issues/1835
Specifying the id of the target node during stream connect is
deprecated. Instead the property target.object should be used to link
by target serial or name. Using the name allows us to drop a bunch of
custom code.
When af_scaletempo2.c:process() detects a format change, it goes back
through mp_scaletempo2_init() to reinitialize everything. However,
mp_scaletempo2.input_buffer is not necessarily reallocated due to a
check in af_scaletempo2_internals.c:resize_input_buffer(). This is a
problem if the number of audio channels increases, since without
reallocating, the buffer for the new channel(s) will at best point to
NULL, and at worst uninitialized memory.
Since resize_input_buffer() is only called from two places, pull size
check out into mp_scaletempo2_fill_input_buffer(). This allows each
caller to decide whether they want to resize or not. We could be
smarter about when to reallocate, but that would add a lot of machinery
for a case I don't expect to be hit often in practice.
Previously we would only call list_devs() on available AOs if an AO
*did not* have a hotplug_init() callback or for the first one that *did*
have it.
This is problematic when multiple fully functional hotplug-capable AOs
are available.
The second one would not be able to contribute discovered devices.
This problem prevents ao_pipewire from introducing full hotplug support
with hotplug_init().
When a platform has multiple valid AOs that can provide hotplug events
we should try to use the one that also provides playback.
Concretely this will help when introducing hotplug support for
ao_pipewire.
Currently ao_pulse is probed by ao_hotplug_get_device_list() before
ao_pipewire and on the common setups where both AOs could work pulse
will be selected for hotplug handling.
This means that hotplug_init() of ao_pipewire will never be called and
list_devs() has to do its own initialization.
But if ao_pulse is non-functional or not compiled-in suddenly
ao_pipewire *must* implement hotplug_init() for hotplugging events to
work for all.
Also if the hotplug ao_pulse connects to a PulseAudio instance that is
not emulated by the same PipeWire instance as the playback ao_pipewire
the hotplug events are useless.
uau did some investigation and noticed that we do not send a wakeup
event when we encounter end-of-stream in ao_read_data(), in contrast to
the equivalent logic for push AOs in ao_play_data().
Inserting that wakeup fixes the original problem of lack of
reinitialization on a format change without the problems we saw with
the previous attempted fix.
Fixes#10566
The error description in #10545 could indicate that we are overflowing
we are corrupting the buffer metadata ourselves through out-of-bound
writes.
This check is also present in pw-cat so it seems to be expected for
b->requested to exceed the actual available buffer space.
Potential fix for #10545
