Disable Format_RGB24 only in case of QOpenGLContext::LibGLES by
moving the no-opengl endif after the check against Format_RGB24.
Change-Id: I389ae2bb32dc8b2c55ec801f52d820620630b1c0
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Add support for changing resolution and frame rate with mfw_v4lsrc.
This is essential for embedded applications that are not happy with
the default VGA@30.
This makes constructs like the following functional on devices like
i.MX6 with MIPI cameras:
Camera { viewfinder { resolution: "320x240"; maximumFrameRate: 15 } }
Change-Id: Ia297afdb5ca51c6e55ad45dce37fdab7da3a5cfb
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
Enable qt.multimedia.video to get the logs. Also enhance the printing
when creating the video node implementation. It is essential to have
an easy way to figure out what handle and formats the node in use
supports.
Change-Id: Idf3a9f076ba03b5e613c19f2347204c841850b45
Reviewed-by: Andy Nichols <andy.nichols@theqtcompany.com>
For QVideoFrame purposes they are equivalent.
Change-Id: I776c60aa8ca848e58499bbd0c1fc25c457d46ef6
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
Camera input is often in this format. Until now the lack of GL_VIV_I420 (which
was presumably missing in older versions of the extension?) in the format list
resulted in falling back to the built-in i420 node instead of the zero-copy imx6
one. This is now corrected by adding the correct format mapping.
Change-Id: I6e891bb3bb07d64ccd5b2d170b052f677c8bd19c
Reviewed-by: aavit <eirik.aavitsland@theqtcompany.com>
Fixed multiply defined symbols when linking statically with both
PulseAudio and Alsa plugins enabled:
The private classes In/OutputPrivate had identical names and have been
renamed.
Change-Id: I9415beeeed9fb0e14ead3f0ab906f343b3934341
Task-number: QTBUG-43514
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
This was preventing the surface from restarting with a different format.
Change-Id: I1f86ddb1b16618f167183c7e2fcb32658df578f3
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
These should match the capture mode. Additionally, there was a semantic
error preventing the encoding properties from being properly selected.
This fixes a bug in which the viewfinder was receiving frames too large
for display as an OpenGL texture.
Task-number: QTBUG-41065
Change-Id: Ia82c8f44bba1692a219edc5f9d78fc76c3d8a4ba
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@theqtcompany.com>
There is no Windows Runtime API to find the camera sensor rotation, so
assume that phones always have a camera mounting of 270 degrees. Tablet
and webcams remain mounted at the default (0 degrees). As the frame is not
flipped automatically by the system, the scan line direction is set to
BottomToTop for front-facing cameras to achieve compatibility with
other platforms.
Task-number: QTBUG-41066
Change-Id: Icf17ecd4aca9fa9d5b24d94e5b21b63ee6f21f28
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
The QPA header isn't needed anymore, as nativeOrientation was added as
a QScreen property in 5.2.
Change-Id: I7cd00feae769175fd0c4be65b503e74ee910814a
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
Use the device ID for QCameraInfo::description() when the driver
doesn't provide any display name.
Change-Id: Iff1f17187ecb52262412f85db04d7108fae71717
Reviewed-by: Andrew den Exter <andrew.den.exter@qinetic.com.au>
The OpenSL volume interface is not available for audio inputs on
Android so we apply the volume ourselves on the PCM data.
Task-number: QTBUG-42159
Change-Id: If43d8aa576bc70a925681f0db1ca8b40e71f7b29
Reviewed-by: Christian Stromme <christian.stromme@theqtcompany.com>
The VideoOutput's camera info was not updated when switching cameras.
Change-Id: I23537ce98b08009898eaa26ef14d5b9a746ab5f7
Reviewed-by: Andrew Knight <andrew.knight@theqtcompany.com>
One is not used, the others are not even initialized, so don't use the
memory.
qsoundeffect_pulse_p.h:172:20: warning: private field 'm_resourceStatus' is not used [-Wunused-private-field]
qvideowidget_p.h:211:25: warning: private field 'm_aspectRatioMode' is not used [-Wunused-private-field]
qalsaaudioinput.h:156:22: warning: private field 'timestamp' is not used [-Wunused-private-field]
qalsaaudiooutput.h:135:22: warning: private field 'timestamp' is not used [-Wunused-private-field]
Change-Id: I7c9f50d57c4d29ee0dfd7dc086771d721cdb5b05
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
Commit 4d35f66acf changed from -1 to ~0 because
ICC complained, but Clang still sees ~0 as int. So force to unsigned.
qgstappsrc.cpp:144:31: warning: comparison of integers of different signs: 'unsigned int' and 'int' [-Wsign-compare]
Change-Id: I6b0fdf8dc5df157d8e184f22dba44660e60ca3b2
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
Each plugin must provide its own unique key. Otherwise we will only ever
see one single plugin.
Right now running on i.MX6 is often broken because the imx6 videonode plugin is
not picked up since only the egl one is seen by the system. With the fix both plugins
provide their own unique key so both become visible.
Additionally, introduce a QT_VIDEONODE environment variable. This is useful to specify
which plugin to use. This is necessary in case multiple custom videonode plugins support
the same formats.
Change-Id: Iaa1988f8436dcb938cb9a95e2e0d68a4e92e113c
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
Header inclusion is not needed, hence remove it for WinRT.
Change-Id: I8117439849143975cad3dc14e36118b8da4621de
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
Add a warning in that case.
Task-number: QTBUG-41966
Change-Id: I491ee9e13c9504040ab789951656983a736d9203
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
Don't use relaxed load and stores, we need to be stricter to avoid
problems with high frequency re-fills of the buffer. If we don't enforce
ordering we might end-up spending more time trying to acquire an open
slot in the buffer. Updating processes bytes is also moved off the
"OpenSL" thread.
Added some comments for improved readability.
Change-Id: Ie27965fc6bf4b8394081ae6419f4933522ada98e
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
According to our own docs we should always return 1.0 (MAX) volume level
if the device doesn't support setting the volume, which is the case
for OpenSL ES on Android.
Task-number: QTBUG-42159
Change-Id: I9144e9e2cd84cc80633c68c5d6fa32b8dbd0fcd5
Reviewed-by: Yoann Lopes <yoann.lopes@theqtcompany.com>
It can be used to set which video source element should be used by
the camerabin.
Change-Id: I8d1cd8c4ba6fe5a89817699f645b0997e713aaca
Reviewed-by: Samuli Piippo <samuli.piippo@digia.com>
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
To get the recording duration, we were using the camerabin's position,
which represents the time since it was started, not the time it's been
recording to a file.
We now retrieve the camerabin's filesink position.
Change-Id: I68eeb25d1718666288655d22deea23e25de73b90
Reviewed-by: Andrew den Exter <andrew.den.exter@qinetic.com.au>
The AVPlayerLayer was set on the QVideoRendererControl before its
geometry was updated, causing the renderer control to display frames
with an invalid size.
Change-Id: I90e18dce69d4b48a3d7932d44a7eab4fd443f1fb
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
If video surface object is set to 0, it must not be accessed to set a
property on it.
Task-number: QTBUG-40746
Change-Id: I1de0e5495918d3ea06706412fab15bf1af012f36
Reviewed-by: Fabian Bumberger <fbumberger@rim.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>
QMediaMetaData::AlbumArtist now maps to GST_TAG_ALBUM_ARTIST and
QMediaMetaData::ContributingArtist to GST_TAG_ARTIST.
Change-Id: Ifa2cb90ee8ae09beaee572ad113c05776e699432
Reviewed-by: Andrew den Exter <andrew.den.exter@qinetic.com.au>
- metaDataAvailableChanged was never emitted
- metaDataChanged was not emitted when clearing the current media
Change-Id: Ide05056450171a87aeb018be7e0bdea136341946
Reviewed-by: Andrew den Exter <andrew.den.exter@qinetic.com.au>
- Correctly emit positionChanged signal. One of the problems with
this was that QMediaPlayer automatically sends periodic
position updates while playing a media. There's no need to have
the same logic in the backend.
- Seeking after reaching the end of the media now correctly works
Auto tests included.
Change-Id: I6d5ecbae6e05f94a8aac1a0834cf57427adf219b
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
The backend was kept only for Mac OS X 10.6 support. AVFoundation is
used in 10.7 and later. Following the 10.6 deprecation plan, the code
is now being removed for Qt 5.4.
Change-Id: I513d00adda35f6012dda548b8d824d158efffb16
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Emit signals only after both properties are written to avoid having
incoherent values in signal handlers.
Task-number: QTBUG-40314
Change-Id: I6c8445e61cccf1a9803647329c4fa1f0e452f56d
Reviewed-by: Christian Stromme <christian.stromme@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>
Important warnings such as missing services are still printed though.
Task-number: QTBUG-31080
Change-Id: I96fd3837c4edc58c61b97bf950b7cab05ceed014
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
This reverts commit 0ed18d846c.
The same patch was submitted in both 5.3 and 5.4 branches. It should
have been pushed only to 5.3.
Change-Id: I9e2974886190dfb2e5def5bc325c58f6e7783f55
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Perform a seek before transitioning from the stopped state to paused or
playing to force the pipeline to resupply the video sink with any
pre-roll buffer it may have previously ignored during loading. And
don't assume showPrerollFrames to be true if the current state is not
stopped as the policy handling may have prevented an effectual state
change.
Change-Id: I288a70bc4da32f3534eab4b14702ca8f8fdb4222
Reviewed-by: Yoann Lopes <yoann.lopes@digia.com>
We where very strict about tearing down the audio device. While this is
a good strategy to avoid unnecessary resource usage, it also causes
excessive re-allocations, e.g., when transiting from start to stop and
back again. This can increase latency, especially in case where a short
clip is re-played at a high frequency.
This change also decrease the chance of the player ending up in some
unknown state where it drops audio clips without any warning.
Task-number: QTBUG-40864
Change-Id: I1afad4af0622983f0f0c221d91cf794585d8cad2
Reviewed-by: Yoann Lopes <yoann.lopes@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>
This is not comprehensive since different encoders have different names
and representations for many settings, but it should cover bit rate for
most encoders, and quality and encodingMode for a number of common
encoders.
Change-Id: I0ba4e70c2f234e0deaaa02bdecc0f5198122c1e9
Reviewed-by: Yoann Lopes <yoann.lopes@digia.com>
