Commit 251f0534b7 ("qcam: viewfinder_gl: Take color space into account
for YUV rendering") introduced maybe-uninitialized warnings with gcc 11
and 12 when compiling with -O3. Both compilers warn that
../../src/qcam/viewfinder_gl.cpp: In member function ‘void ViewFinderGL::selectColorSpace(const libcamera::ColorSpace&)’:
../../src/qcam/viewfinder_gl.cpp:392:21: error: ‘offset’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
391 | fragmentShaderDefines_.append(QString("#define YUV2RGB_Y_OFFSET %1")
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
392 | .arg(offset, 0, 'f', 1));
| ~~~~^~~~~~~~~~~~~~~~~~~
Additionally, gcc 12 warns that
../../src/qcam/viewfinder_gl.cpp: In member function ‘void ViewFinderGL::selectColorSpace(const libcamera::ColorSpace&)’:
../../src/qcam/viewfinder_gl.cpp:379:36: error: ‘yuv2rgb’ may be used uninitialized [-Werror=maybe-uninitialized]
379 | yuv2rgb[i] *= 255.0 / 219.0;
../../src/qcam/viewfinder_gl.cpp:330:31: note: ‘yuv2rgb’ declared here
330 | std::array<double, 9> yuv2rgb;
|
While this should never happen here, the compiler isn't necessarily
wrong, as C++17 allows initializing a scoped enum from an integer using
direct-list-initialization, even if the integer value doesn't match any
of the enumerators for the scoped enum ([1]). Whether this is valid or
borderline paranoia from gcc may be debatable, but in any case it can't
be classified as blatantly wrong. Fix the warnings by adding default
cases to the switch statements in ViewFinderGL::selectColorSpace().
Which case is selected as the default doesn't matter, as this is not
meant to happen.
[1] https://en.cppreference.com/w/cpp/language/enum#enum_relaxed_init_cpp17
Bug: https://bugs.libcamera.org/show_bug.cgi?id=143
Fixes: 251f0534b7 ("qcam: viewfinder_gl: Take color space into account for YUV rendering")
Signed-off-by: Marco Felsch <m.felsch@pengutronix.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Rewrote commit message, added a default case for the encoding switch.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Currently to request a frame, we operate the camera directly.
This approach is also scattered in two places,
MainWindow::startCapture() and MainWindow::queueRequest().
This makes it difficult to account for requests.
Centralize all the queuing to a single function queueRequest()
Rename the current queueRequest() to renderComplete().
This makes more sense as this slot is triggered when
the render is complete and we want to queue another
request.
Signed-off-by: Utkarsh Tiwari <utkarsh02t@gmail.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Replace the cameraCombo_ on the toolbar with a QPushButton which
displays the CameraSelectorDialog. This would allow the user to view
information about the camera when switching.
The QPushButton text is set to the camera Id currently in use.
Signed-off-by: Utkarsh Tiwari <utkarsh02t@gmail.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Currently if there is HotPlug event when the user is on the Camera
selection dialog, the QComboBox doesn't update to reflect the change.
Add support for hotplugging / unplugging cameras.
Signed-off-by: Utkarsh Tiwari <utkarsh02t@gmail.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Currently we use QInputDialog convenience dialogs to allow the user to
select a camera. This doesn't allow adding of more information (such as
camera location, model etc).
Create a QDialog with a QFormLayout that shows a QComboBox with camera
Ids. Use a QDialogButtonBox to provide buttons for accepting and
cancelling the action.
The CameraSelectorDialog is only initialized the first time when the
MainWindow is created.
From this commit we cease to auto select the camera if only a single
camera is available to libcamera. We would always display the selection
dialog with the exception being that being if the camera is supplied on
the command line.
Signed-off-by: Utkarsh Tiwari <utkarsh02t@gmail.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Update the YUV shaders and the viewfinder_gl to correctly take the
Y'CbCr encoding and the quantization range into account when rendering
YUV formats to RGB. Support for the primaries and transfer function will
be added in a subsequent step.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Reviewed-by: Kunal Agarwal <kunalagarwal1072002@gmail.com>
Define Span types explicitly as either variable- or fixed-sized. This
introduces a new convention for defining Span dimensions in the property
and control value definitions and generates Span types as variable-sized
Span<T> or as fixed-sized Span<T,N>.
Signed-off-by: Christian Rauch <Rauch.Christian@gmx.de>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The OpenGL ES shading language has no default precision declared
implicitly for floats in fragment shaders. The lack of an explicit
default precision results in shader compilation errors.
Specify a default precision of mediump for floats. This matches the
other fragment shaders, and is guaranteed by the OpenGL ES shader
language specification to be supported by all devices, while the higher
precision highp is optional.
Signed-off-by: Kunal Agarwal <kunalagarwal1072002@gmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Now that the ControlList::get() function returns an instance of
std::optional<>, we can replace the ControlList::contains() calls with a
nullopt check on the return value of get(). This avoids double lookups
of controls through the code base.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
Previously, ControlList::get<T>() would use default constructed objects to
indicate that a ControlList does not have the requested Control. This has
several disadvantages: 1) It requires types to be default constructible,
2) it does not differentiate between a default constructed object and an
object that happens to have the same state as a default constructed object.
std::optional<T> additionally stores the information if the object is valid
or not, and therefore is more expressive than a default constructed object.
Signed-off-by: Christian Rauch <Rauch.Christian@gmx.de>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Support the X variants of the RGB pixel formats alongside the
equivalent Alpha component based versions.
The QImage::Format_RGB32 and QImage::Format_RGBX8888 formats only
specify that the alpha component must be 0xff. While the Qt
documentation doesn't guarantee that the alpha value will be ignored by
consumers, this seems to be the implemented behaviour, at least when
rendering with QPainter::drawImage().
The existing use of QImage::Format_RGBA8888 is updated to use the
QImage::Format_RGBX8888 variant to ensure that the image data is
visible.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Remove the verbose #ifndef/#define/#endif pattern for maintaining
header idempotency, and replace it with a simple #pragma once.
This simplifies the headers, and prevents redundant changes when
header files get moved.
Signed-off-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
"using namespace" in a header file propagates the namespace to
the files including the header file. So it should be avoided.
This removes "using namespace" in header files in qcam.
Signed-off-by: Hirokazu Honda <hiroh@chromium.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The RGB and YUV conversion doesn't take the stride into account, neither
when creating the textures, nor when sampling them. Fix it by using the
stride as the texture width, and multiplying the x coordinate in the
vertex shaders by a factor to only sample the active portion of the
image.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Add support for the YUV420, YVU420 and YUV422 formats supported by
libcamera. YUV420 can be produced by the Raspberry Pi pipeline handler,
being able to display it is useful for testing.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
qcam currently assumes that no padding is used at end of lines, and uses
the image width as the stride. This leads to rendering failures with
some formats on some platforms. To prepare for stride support, add a
stride parameter to the ViewFinder::setFormat() function to pass the
stride from the stream configuration to the viewfinder.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Replace the manual implementation of frame buffer mapping with the Image
class to improve code sharing. The ViewFinder API is updated to take an
Image pointer in the render() function to prepare for multi-planar
buffer support.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The number of metadata planes should always match the number of frame
buffer planes. Enforce this by making the vector private and providing
accessor functions.
As this changes the public API, update all in-tree users accordingly.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Jean-Michel Hautbois <jeanmichel.hautbois@ideasonboard.com>
Reviewed-by: Hirokazu Honda <hiroh@chromium.org>
When disconnecting a signal from a receiver, it is usually not necessary
to specify the receiver's slot function explicitly, as the signal is
often connected to a single slot for a given receiver. We can thus use a
simpler version of Signal::disconnect() that takes a pointer to the
receiver object only. This reduces code size, as the disconnect()
function is a template function.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Umang Jain <umang.jain@ideasonboard.com>
The GL_RG and GL_RED texture formats are not supported in OpenGL ES
prior to 3.0. In order to be compatible with OpenGL ES 2.0, use
GL_LUMINANCE_ALPHA and GL_LUMINANCE instead. The shader code needs to be
updated accordingly for GL_RG, as the second component is now stored in
the alpha component instead of the green component. Usage of the red
component is fine, the luminance value is stored in the red, green and
blue components.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Reviewed-by: Andrey Konovalov <andrey.konovalov@linaro.org>
The shader supports all 4 packed RAW10 variants.
Simple bi-linear Bayer interpolation of nearest pixels is implemented.
The 2 LS bits of the 10-bit colour values are dropped as the RGBA
format we convert into has only 8 bits per colour.
The texture coordinates passed to the fragment shader are adjusted
to point to the nearest pixel in the image. This prevents artifacts
when the image is scaled from the frame resolution to the window size.
Signed-off-by: Andrey Konovalov <andrey.konovalov@linaro.org>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reduce the indentation in the gstreamer and qcam meson.build with
subdir_done(). This follows similar patterns in cam and v4l2.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
The meson.build files mix array declarations with and without a space
after the opening and before the closing square bracket. The vast
majority of cases don't use spaces, so standardize on that.
While it it, fix indentation in a few places.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund@ragnatech.se>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Utkarsh Tiwari
Marco Felsch
Laurent Pinchart
Christian Rauch
Naushir Patuck
Kunal Agarwal
Kieran Bingham
Hirokazu Honda
Andrey Konovalov
Fabrice Fontaine