Using --start with files that use DTS only, or which simply have broken
PTS timestamps, would incorrectly drop frames and possibly not execute
the seek correctly.
Add yet another heuristic to detect this. The intent is that --start and
hr-seeks in general should work correctly, but in order to keep things
fast, we still want to allow frame dropping during hr-seek if there are
no problems doing so. Do this by disabling frame dropping by default,
but re-enabling it if there are no problems found for a while. As a
consequence, --start might be somewhat slower, but normal user
interaction should remain as fast as before.
Note that there's something subtle about the added code: the
has_broken_packet_pts field is checked even before the first packet is
fed to dec_video.c, so the field must not be set to 0 right on start.
It's not initially set to 0 anyway, because the heuristic requires
decoding some images before enabling frame drop anyway.
Note 2: it's not clear whether frame dropping during hr-seek really
helps; I didn't benchmark it.
The d_video->pts field was a bit strange. The code overwrote it multiple
times (on decoding, on filtering, then once again...), and it wasn't
really clear what purpose this field had exactly. Replace it with the
mpctx->video_next_pts field, which is relatively unambiguous.
Move the decreasing PTS check to dec_video.c. This means it acts on
decoder output, not on filter output. (Just like in the previous commit,
assume the filter chain is sane.) Drop the jitter vs. reset semantics;
the dec_video.c determined PTS never goes backwards, and demuxer
timestamps don't "jitter".
Also get rid of the PTS check _after_ filters. This means if there's a
video filter which unsets PTS, no warning will be printed. But we assume
that all filters are well-behaved enough by now.
These used the suffix _resync_stream, which is a bit misleading. Nothing
gets "resynchronized", they really just reset state.
(Some audio decoders actually used to "resync" by reading packets for
resuming playback, but that's not the case anymore.)
Also move the function in dec_video.c to the top of the file.
These packets have to be explicitly dropped, because usually libavcodec
uses 0-sized packets to flush delayed frames, meaning just passing
through these packets would have bad consequences.
Normally, libavformat doesn't output 0-sized packets anyway. But I don't
want to take any chances, so don't delete it, and just move it out of
the way to demux.c.
If the value for --cache-on-pause is larger than --cache-min, and the
cache runs below --cache-on-pause, but above --cache-min, the logic
would demand to pause the player and then unpause immediately again.
This doesn't make much sense, and alternating the pause state in each
playloop iteration has negative consequences. Add an explicit check to
avoid this situation.
This means the code that tries to figure out the timestamp from
demuxer and decoder output is now all in dec_video.c. We set the
final timestamp on the returned image (mp_image.pts), as well as
the d_video->pts field.
The way the player uses d_video->pts field is still a bit messy. Maybe
this could be cleaned up later.
Having the DTS directly can be useful for restoring PTS values.
The avi file format doesn't actually store PTS values, just DTS. An
older hack explicitly exported the DTS as PTS (ignoring the [I assume]
genpts generated non-sense PTS), which is not necessary anymore due to
this change.
Now the --no-correct-pts mode is like the normal mode, just with
different timestamp calculations. The semantics should be about the
same as before this commit.
Before this commit, this mode estimated the frame time by subtracting
successive packet PTS values. This is complete non-sense for video
codecs which use reordering. The code compensated frame times for these
non-sense using the FPS value, but confused the rest of the player with
non-sense jumping around timestamps. So, all in all this mode is not
very useful.
Repurpose this mode for fixed frame rate playback. This gives almost the
same behavior as the old mode with forced framerate (--fps option). The
result is simpler and often more robust.
Instead of passing the PTS as separate field, pass it as part of the
usual data structures. Basically, this removes strange artifacts from
the API. (It's not finished, though: the final decoded PTS goes through
strange paths, and filter_video() finally overwrites the decoded
mp_image's pts field with it.)
We also stop using libavcodec's reordered_opaque fields, and use
AVPacket.pts and AVFrame.pkt_pts. This is slightly unorthodox, because
these pts fields are not "really" opaque anymore, yet we treat them as
such. But the end result should be the same, and reordered_opaque is
marked as partially deprecated (it's not clear whether it's really
deprecated).
This normally shouldn't happen. It does happen with VfW-muxed mkv files,
and would normally happen with avi files (except that we force the
correct association mode using an explicit hack for avi).
This is usually prepended by warnings like:
Decreasing video pts: 0.125000 < 0.167000
and after the switch, there should be no warnings anymore.
Background: avi likes to use DTS for timestamps instead of PTS.
Basically, there are no proper timestamps in the file, only frame
numbers. And Matroska is insane and stores the made-up DTS instead
of a proper PTS. This will be handled properly later.
This is not needed anymore, because we decided that the PAR of the
decoded video matters, and not the PAR of the filtered video that
arrives at the VO.
In theory, we can't really do this, because we don't know when a spdif
frame ends. Spdif transports compressed audio through audio setups that
were originally designed for PCM only (which includes the audio filter
chain, the AO API, most audio output APIs, etc.), and to reach this
goal, spdif pretends to be PCM. Compressed data frames are padded with
zeros, until a certain data rate is reached, which corresponds to a
pseudo-PCM format with 2 bytes per sample and 2 channels at 48000 Hz.
Of course an actual spdif frame is significantly larger than a frame
of the PCM format it pretends to be, so cutting audio data on frame
boundaries (as according to the pseudo-PCM format) merely yields an
incomplete and broken frame, not audio that plays for the desired
duration.
However, sending an incomplete frame might still be much better than the
current behavior, which simply ignores --end/--length (but still lets
the video end at the exact end time).
Should this result in trouble with real spdif receivers, this commit
probably has to be reverted.
If the --fps option was given (MPOpts->force_fps), the demuxer FPS value
was overwritten with the forced value. This was fine, since the demuxer
value wasn't needed anymore. But with the recent changes not to write to
the demuxer stream headers, we don't want to do this anymore. So
maintain the (forced/updated) FPS value in dec_video->fps.
The removed code in loadfile.c is probably redundant, and an artifact
from past refactorings.
Note that sub.c will now always use the demuxer FPS value, instead of
the user override value. I think this is fine, because it used the
demuxer's video size values too. (And it's rare that these values are
used at all.)
The only reason why these structs were dynamically allocated was to
avoid recursive includes in stheader.h, which is (or was) a very central
file included by almost all other files. (If a struct is referenced via
a pointer type only, it can be forward referenced, and the definition of
the struct is not needed.) Now that they're out of stheader.h, this
difference doesn't matter anymore, and the code can be simplified.
Also sneak in some sanity checks.
This used to be needed to access the generic stream header from the
specific headers, which in turn was needed because the decoders had
access only to the specific headers. This is not the case anymore, so
this can finally be removed again.
Also move the "format" field from the specific headers to sh_stream.
Use sh_stream over sh_sub. Use dec_sub (and mpctx->d_sub) instead of the
stream header. This aligns the subtitle code with the recent audio and
video refactoring.
sh_sub still has the decoder context, though. This is because we want to
avoid reinit when switching segments with ordered chapters. (Reinit is
fast, except for creating the ASS_Renderer, which in turn triggers
fontconfig.) Not sure how much this matters, though, because the initial
segment switch will lazily initialize the decoder anyway.
This is similar to the sh_audio commit.
This is mostly cosmetic in nature, except that it also adds automatical
freeing of the decoder driver's state struct (which was in
sh_video->context, now in dec_video->priv).
Also remove all the stheader.h fields that are not needed anymore.
This means most code accessing this struct must now include hwdec.h
instead of dec_video.h. I just put it into dec_video.h at first because
I thought a separate file would be a waste, but it's more proper to do
it this way, as there are too many files which include dec_video.h only
to get the mp_hwdec_info definition.
This was printed before something was decoded, and thus was not really
correct. Also, this code is hilariously broken:
/* Assume FOURCC if all bytes >= 0x20 (' ') */
if (sh_audio->format >= 0x20202020)
mp_msg(MSGT_IDENTIFY, MSGL_INFO,
"ID_AUDIO_FORMAT=%.4s\n", (char *)&sh_audio->format);
Time to kill it.
This information can be accessed through properties instead.
sh_audio is supposed to contain file headers, not whatever was decoded.
Fix this, and write the decoded format to separate fields in the decoder
context, the dec_audio.decoded field. (Note that this field is really
only needed to communicate the audio format from decoder driver to the
generic code, so no other code accesses it.)
Move all state that basically changes during decoding or is needed in
order to manage decoding itself into a new struct (dec_audio).
sh_audio (defined in stheader.h) is supposed to be the audio stream
header. This should reflect the file headers for the stream. Putting the
decoder context there is strange design, to say the least.
This was forgotten when the parser for mplayer2 EDL files was removed.
Change the header of the mpv EDL format to include a '#', so a naive
parser could skip the header as comment. (Maybe this is questionable;
on the other hand, if it can be simpler, why not.)
Also, strip the header in demux_edl.c before passing on the data, so the
header check doesn't need to be duplicated in tl_mpv_edl.c.
Edit Decision Lists (EDL) allow combining parts from multiple source
files into one virtual file. MPlayer had an EDL format (which sucked),
which mplayer2 tried to improve with its own format (which sucked). As
logic demands, mpv introduces its very own format (which sucks).
The new format should actually be much simpler and easier to use, and
its implementation is simpler and smaller too.
The intention of the existing code was trying to match demuxer-reported
stream IDs, instead of using possibly arbitrary ordering of the frontend
track list. But EDL files can consist of quite different files, for
which trying to match the stream IDs doesn't always make sense.
This didn't have any consequences, other than suddenly reinitializing
video when it works again (such as with EDL timeline mixing video and
audio-only files).
The AF control commands used an elaborate and unnecessary organization
for the command constants. Get rid of all that and convert the
definitions to a simple enum. Also remove the control commands that
were not really needed, because they were not used outside of the
filters that implemented them.
Some decoders used to read packets and decode data when calling
resync_audio_stream(). This required a special case in mp_seek() for
audio. (A comment mentions liba52, which is long gone; but until
recently ad_mpg123.c actually exposed this behavior.)
No decoder does this anymore, and resync_audio_stream() works similar
as resync_video_stream(). Remove the special case.
When the decoder detects a format change, it overwrites the values
stored in sh_audio (this affects the members sample_format, samplerate,
channels). In the case when the old audio data still needs to be
played/filtered, the audio format as identified by sh_audio and the
format used for the decoder buffer can mismatch. In particular, they
will mismatch in the very unlikely but possible case the audio chain is
reinitialized while old data is draining during a format change.
Or in other words, sh_audio might contain the new format, while the
audio chain is still configured to use the old format.
Currently, the audio code (player/audio.c and init_audio_filters) access
sh_audio to get the current format. This is in theory incorrect for the
reasons mentioned above. Use the decoder buffer's format instead, which
should be correct at any point.
demuxer->filepos contains the byte offset of the last read packet. This
is so that the player can estimate the current playback position, if no
proper timestamps are available. Simplify it to use demux_packet->pos in
the generic demuxer code, instead of bothering every demuxer
implementation about it.
(Note that this is still a bit incorrect: it relfects the position of
the last packet read by the demuxer, not that returned to the user. But
that was already broken, and is not that trivial to fix.)
Most libavcodec decoders output non-interleaved audio. Add direct
support for this, and remove the hack that repacked non-interleaved
audio back to packed audio.
Remove the minlen argument from the decoder callback. Instead of
forcing every decoder to have its own decode loop to fill the buffer
until minlen is reached, leave this to the caller. So if a decoder
doesn't return enough data, it's simply called again. (In future, I
even want to change it so that decoders don't read packets directly,
but instead the caller has to pass packets to the decoders. This fits
well with this change, because now the decoder callback typically
decodes at most one packet.)
ad_mpg123.c receives some heavy refactoring. The main problem is that
it wanted to handle format changes when there was no data in the decode
output buffer yet. This sounds reasonable, but actually it would write
data into a buffer prepared for old data, since the caller doesn't know
about the format change yet. (I.e. the best place for a format change
would be _after_ writing the last sample to the output buffer.) It's
possible that this code was not perfectly sane before this commit,
and perhaps lost one frame of data after a format change, but I didn't
confirm this. Trying to fix this, I ended up rewriting the decoding
and also the probing.
Before this commit, the af_instance->mul/delay values were in bytes.
Using bytes is confusing for non-interleaved audio, so switch mul to
samples, and delay to seconds. For delay, seconds are more intuitive
than bytes or samples, because it's used for the latency calculation.
We also might want to replace the delay mechanism with real PTS
tracking inside the filter chain some time in the future, and PTS
will also require time-adjustments to be done in seconds.
For most filters, we just remove the redundant mul=1 initialization.
(Setting this used to be required, but not anymore.)
wm4
Stefano Pigozzi