Romain Vimont authored 3 years ago and Romain Vimont committed 3 years ago

The option --deinterlace-mode selects the deinterlace algorithm to use,
but the same algorithm might be implemented by several modules, for
example a CPU version and an OpenGL version.

For that purpose, add a separate option to select the module(s) to use.
The default value is "deinterlace".

To select the blend deinterlace algorithm:

    --deinterlace-mode=blend

or equivalently ("deinterlace" is the default filter):

    --deinterlace-filter=deinterlace --deinterlace-mode=blend

To run the OpenGL version (the mode is "auto"):

    --deinterlace-filter=glblend

The module glblend only implements the "blend" algorithm, it can be
passed explicitly:

    --deinterlace-filter=glblend --deinterlace-mode=blend

To pass a list a modules, executed in order until one works:

    --deinterlace-filter=glblend,deinterlace --deinterlace-mode=blend

Romain Vimont authored 4 years ago and Romain Vimont committed 3 years ago

This will allow to try an alternate deinterlace filter (for example
an OpenGL filter).

Thomas Guillem authored 3 years ago and Jean-Baptiste Kempf committed 3 years ago

Quoting mpg123.de website:

"Fix misguided handling of part2_3_length checks in
III_get_scale_factors_1() and III_get_scale_factors_2() which
invalidated decoding of a mono source encoded as ms+i-stereo.
This was a regression introduced with version 1.25.7."

Fixes #25798

Jean-Baptiste Kempf authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

Close #23602

1. Avoid dynamic memory allocation for small samples_per_chunk
(less or equal to 2 sample per chunk). Depends on the test file
and OS, the time taking by MP4_TrackClean can be reduced from a
few seconds to dozens of milliseconds.

2. Merge some allocation to single memory chunk. For mp4 files
with a lot of chunks, the time taking by MP4_TrackClean can be
reduced by half.

Romain Vimont authored 3 years ago and Romain Vimont committed 3 years ago

The config "advanced" flag was unused and has been removed by
6a7a137f.

It has been removed from many add_*() macros, but not all. Remove it
from the remaining macros.

Alexandre Janniaux authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

The UIKit functions must be called from the main thread, whereas the
functions from the inhibiter can be called from the vout thread. It
didn't seem to have any consequence on iPad, or the testing I did was
not triggering the crash, but it's leading to a crash on tvOS when
running with VLCKit.

Brad Smith authored 3 years ago and Thomas Guillem committed 3 years ago

Set volume/mute function pointers when handle is freed, to ensure
sio_setvol() is not called on a freed handle. Fixes crashes when volume
is adjusted while playback is stopped.

Fixes #25796

Refs #25694

Move it to DecoderWaitUnblock(). It is now possible to do this
simplification because of the previous commit: DecoderWaitUnblock() is
now called only when has_data should be changed to true.

Video decoders always display their first picture, in a forced state,
bypassing the DecoderWaitUnblock() check. Re-arrange the code to make it
more obvious.

We generally see cond_wait surrounded by a while loop.

It is not needed if the picture is forced or if we don't care about
frame lateness (frame_next).

No functional changes.

Not needed anymore since this variable is always locked by the owner->lock.

In order to merge the usage of the same lock and avoid locking it twice
in a row. This new status struct be used to get the status of sub ES
tracks created by decoder modules.

Prince Gupta authored 3 years ago and Pierre Lamot committed 3 years ago

parentId property was removed in cd89555e

  both the intf and the vout window reference the qt interface structure
  it will only be released once not used by both modules, this way the windows
  continues to exists if the interface is closed before.

  When the interface module is released before the vout window, the qml view and
  the associated resources are unloaded from the compositor.

  fix: #25600

  this allows to unload the GUI part of the interface without destroying the
  windows that may still be used by the vout window part.

  the variable was no longer initialized properly as it was stored
  in a "C" structure and there was no reason to store this variable in
  the module

  this will allow to unload the interface module without unloading Qt in a
  further patch.

  the new structure is still a vlc object as we use it notably for logging
  capacity and to interact with vlc variables

  This is a preparatory commit

Niklas Haas authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

It's finally time to use RAVU/FSRCNN(X)/NNEDI3/Anime4K in VLC.

Currently limited to only one shader, which isn't a huge limitation
seeing as you can just `cat` together as many shaders as you want to use
into one big file. But patches welcome(tm).

Niklas Haas authored 3 years ago and Hugo Beauzée-Luyssen committed 3 years ago

Unfortunately the UI for loading these is a bit clunky as they're hidden
behind so many layers of menu and advanced options, plus the fact that
the new LUTs still only get applied on vout reinit.

But it's still a feature that a lot of users have requested heavily, so
supporting them at all is a good thing. The implementation is rather
straightforward even though the amount of #ifdef's are really getting on
my nerves.

Can somebody just time travel me into a future universe in which I can
make libplacebo v4 the minimum dependency?