QAbstractVideoSurface::supportedFormats() returns the formats in
descending order of preference while IMFMediaTypeHandler expects the
list to be in ascending order.
Change-Id: I0ecb5a3b228c0cad78c1a84091e3c3cf272f7ebc
Reviewed-by: Christian Stromme <christian.stromme@theqtcompany.com>
MFVideoFormat_RGB24 has a BGR layout and should therefore be matched to
QVideoFrame::Format_BGR24, not Format_RGB24.
Task-number: QTBUG-42323
Change-Id: I1749b575b1738168e3d3c0d055dee10f6d870c0c
Reviewed-by: Christian Stromme <christian.stromme@theqtcompany.com>
Add a warning in that case.
Task-number: QTBUG-41966
Change-Id: I491ee9e13c9504040ab789951656983a736d9203
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
Checking the sender of the mediaSourceReady signal to prevent accessing
the incorrect source resolver.
When the source resolver has finished the asynchronous operation and
the source resolver gets recreated in the player at the same time in a
different thread the signal mediaSourceReady still gets emitted from the
old source resolver.
The player assumes that the signal was emitted from the current source
resolver and accesses the unresolved media source in the
handleMediaSourceReady slot.
Task-number: QTBUG-39980
Change-Id: Ic52f6918995aac250048d91f89c520cfea111bd0
Reviewed-by: Yoann Lopes <yoann.lopes@digia.com>
A race condition could cause a frame to be presented even after the
QAbstractVideoSurface was deleted. We now check that the surface is
valid before presenting a frame.
Task-number: QTBUG-41158
Change-Id: If593469a8267583e499e781336af38d3fbf318fd
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
The format of the offscreen surface used for video rendering should be
compatible with the EGL config.
Change-Id: Ic016245ce80c2483771e620c3eed345262d03c44
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
According to the docs, MESessionTopologyStatus with status == MF_TOPOSTATUS_READY should be the correct place for the GetService call.
Change-Id: I7fdbedbe43b2191b35b95c7fd9c86940f58daff7
Reviewed-by: Wouter Huysentruit <wouter_huysentruit@hotmail.com>
Reviewed-by: Yoann Lopes <yoann.lopes@digia.com>
- Correctly initialize and clear PROPVARIANT structures
- Return coherent data even when the information is not available
Change-Id: I22b46f95f255cbb740a154c6296a5c3a91e64f67
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
It was not taken into account at all.
Change-Id: I4ce85aba214cb4d89dcd018b1616a2a38094b5a6
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
When going back to normal rate after playing in fast forward (greater
than 2x), playback seemed frozen for some amount of time (up to 8
seconds).
When playing in fast forward, only key frames are shown, ignoring all
the others. When returning to a normal rate, the source reader will
usually be pointing to a key frame in the future compared to the
player clock position, meaning that all the frames in between won't be
shown until the player clock catches up with the latest key frame that
was read.
When leaving fast-forward, we now reset the position on the player to
force the source reader to point back to the frame at the current clock
position and avoid the seamingly frozen playback.
Also, emit playbackRateChanged() signal when changing the playback
rate.
Change-Id: I4f04f0f250083378e94fb4a47f9f917abeaaf24e
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
When seeking, the presentation clock can enter an undefined state until
it is started again from the new position. Wait for the clock to be
restarted before scheduling the prerolled frames, otherwise these
frames might get a wrong presentation time.
Change-Id: I02cb3338239775b7ef5d206ec5aa1b26719ac978
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
The function was returning a member variable after the object was
deleted. Return a local copy instead.
Change-Id: I077e7e1c184ed0cfd4899a80c9147a2939c93cd7
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
Add virtual destructor to SampleGrabberCallback to let it properly
destruct its subclass when calling 'detete this'
Change-Id: I5f74f82314e22225a0a153c7f4bca531a79dd8f9
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Yoann Lopes <yoann.lopes@digia.com>
Initialize m_cRef to 0 in constructor of SourceResolver::State, otherwise
it will be leaked
Change-Id: I703ef8e91ad88112102ede7e8321d0eca795b840
Reviewed-by: Yoann Lopes <yoann.lopes@digia.com>
- When scrubbing, request frames only one at a time.
- Discard frames when too late or too much in advance
- Fix integer overflow causing undefined behavior
[ChangeLog][QtMultimedia][Windows] Fixed video playback playing at
twice the normal rate after reaching 3:34.
[ChangeLog][QtMultimedia][Windows] Fixed video playback that could
freeze after seeking to a different position.
Task-number: QTBUG-31800
Change-Id: Ie620c684c58ee790537969ffc40f01610b6745ea
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
Instead of setting the volume on the audio session, which is shared by
all QMediaPlayers, we now set the volume on the media player's own audio
stream. This results in all QMediaPlayers correctly having independent
volumes.
[ChangeLog][QtMultimedia][Windows] QMediaPlayer::setVolume() does not
affect the volume of other QMediaPlayers anymore.
Task-number: QTBUG-30317
Change-Id: I8ea8ec47fc86127da01dc5c8247fb6f72c834630
Reviewed-by: Wouter Huysentruit <wouter_huysentruit@hotmail.com>
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
The class allows to get the list of available cameras on the system as
well as getting some static information about them such as their unique
ID, display name, physical position and sensor orientation.
This makes QCamera::availableDevices() and QCamera::deviceDescription()
obsolete.
This patch contains the API, documentation and auto-tests but not the
actual implementation by each backend (except for retrieving the default
camera device).
[ChangeLog][QtMultimedia] Added new QCameraInfo class
[ChangeLog][QtMultimedia] QCamera: availableDevices() and
deviceDescription() are deprecated, use QCameraInfo instead
Change-Id: I64fd65729ab26a789468979ed5444ee90bb82cd0
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
When -no-opengl was used then it was not automatically pulling in the
needed Windows libraries as it was not loading opengl.prf to get them
from.
Change-Id: I9b73e5ee01da964a64c71b6261ff4ed98ce6f1a9
Reviewed-by: Yoann Lopes <yoann.lopes@digia.com>
- Switch to BufferedMedia only once playback actually started, not when
requesting to start.
- Report the position to have changed when seeking in stopped state.
Change-Id: I930b3e6977cebe5935ed033d0a4d4e1eb899ad2c
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
This reverts commit d599f7319a.
This was not the correct logic...
According to the documentation, the BufferedMedia status should be
set only when in the PlayingState.
Change-Id: I36053ebc09c0517fcd2a1a7f2b091fbe8f04f3d0
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
Do not add the AudioEndpoint to the topology if no Audio output device
is available. Fixes video not playing if you deactivate your soundcard
or have no headphones/speakers plugged in.
Change-Id: I9fc2486198a299b3e75af648f69475270968c6f7
Reviewed-by: Yoann Lopes <yoann.lopes@digia.com>
This is necessary for QML MediaPlayer to report the correct position at
the end of a media.
Change-Id: Ifac2a721b850c726305d1a98e360da638b1fa87a
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
The EVR sink allocates its buffers and expect the node connected to it
to use them. Our video probe MFTransform should therefore copy the input
buffers into the EVR buffers rather than just passing on the input
buffers.
Task-number: QTBUG-30435
Change-Id: I978479ced341b96ce627c682f99662dec1dfdb3f
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
This also allows to play streams without providing a MIME type.
Task-number: QTBUG-33631
Task-number: QTBUG-33518
Change-Id: Ia106592cb5ed0a0f7ede6308dc8e2c12cf4e63b1
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
We were using one function which is available only on Windows 7 (and
later). Replace it with Vista-compatible calls.
Task-number: QTBUG-32864
Change-Id: I77492a407330c3689dfbf8dc1180894cf7ca5f8d
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
The QVideoFrame documentation explicitly says that the time is in
microseconds, however the GStreamer backend was setting the time in
milliseconds and the WMF backend in 100-nanosecond units.
With WMF, the time was missing from the QVideoFrame when presenting it to
the video surface.
Task-number: QTBUG-31731
Change-Id: I0638d2abf8eed25b3a531db67c19a18703e5b630
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
It was changing to EndOfMedia status and explicitly stopping playback
when receiving the MEEndOfPresentation event from the WMF session.
However, this event means that all data has bean read from the source but
not necessarily played yet. According to the documentation, playback is
done when the MESessionEnded event is sent. It now reports the EndOfMedia
status at that moment instead. stop() is not explicitly called anymore since
MESessionEnded also implies the session has stopped.
Task-number: QTBUG-30825
Change-Id: I6c6c09e736fe33f7cf17c75038ea7be1b5701a1c
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
To have a consistent behavior with other backends, the WMF plugin now
starts the session after loading a media in order to start buffering some
data and correctly notify when the media is buffered.
It was previously reporting a BufferedMedia status only (and wrongly)
after explicitly starting the media player.
Not all source readers (usually a source reader is specific to a file
format) implement the service needed to query buffering progress. In that
case just report the media to be buffered immediately after loading.
Change-Id: I6e6332ae08e96fc789556761e5169b88c36c5e37
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
They previously had the multimedia services they provided as keys. This
was a problem when several plugins were available on the same platform,
providing the same multimedia service, but with different features or
capabilities. Since they had the same key, only the first plugin in the
directory was loaded.
Nevertheless, it was actually working until commit 732dcfe7 in qtbase,
as all plugins were loaded even when sharing the same key.
The services a multimedia plugin provides are now declared in the
'Services' property. To preserve compatibility with third-party plugins,
if that new property doesn't exist it falls back to the 'Keys' property.
Task-number: QTBUG-31868
Task-number: QTBUG-31476
Change-Id: Ic3ba32eeef21b69b922bd3d4feb111101559d132
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
