@nishithaa I have tried to do this for years to make a VR video player using Qt and I'm finally close to achieving it but without gstreamer. I don't know if gstreamer is a requirement to you but I use libvlc 4, based on this example : https://github.com/videolan/vlc/tree/master/doc/libvlc/QtGL . It compiles and runs well but I don't want to use QOpenGLWidget (I use QWindows with manually-managed QOpenGLContexts) and somehow I can't make it to work in this setup.

If it's a library that uses CMake, it's usually pretty straightforward. For glfw, which is one you appear to be using, see its CMake documentation

@HimanshuKaushik said in error: opengl32.lib: No such file or directory: Please suggest how do I build it in RHEL environment Don't link against opengl32.lib then - it's a windows library.

@Chris-Kawa That would work for multiple QOpenGLWidgets in the same application. My use case involves sharing the context of one QOpenGLWidget with that of a QOffscreenSurface

It's been solved by create a process and launch the QOpenGLWidget's .exe file.

Hi, Which version of Qt are you using ? Do you have the same issue if you run Qt's OpenGL examples ?

Issue is resolved. I had updated my GPU drivers and after restart, it began to work.

@JoeCFD I was working on other stuffs but now I remember why I can't override QQuickItem update() method... it is a non-virtual slot, so it can't solve my rendering issues. I am still looking for a solution ...

@Rich-v Closing this. Qt support confirmed they do not support windows hosts building 6.2 for QNX.

@harlamer Without a minimal compilable example that demonstrates the problem there's not much we can do. We do not know what platform, what font, what texture and texture setup, how the painter has been set up, what "displayed correctly" means, etc.

It should be supported indeed. Did you try with a more recent release of Qt ? Like 5.15 or even 6.2 ?

@SGaist Ok, thank you. I'll give it a go

I don't have an exact answer, but do you need any of that PBO stuff? You have a QImage and you want a QOpenGLTexture, so you can just do that directly: https://doc.qt.io/qt-5/qopengltexture.html#QOpenGLTexture-1 The texture object can be made directly from a QImage, with no extra steps. @Carlos-Roberto said in How to implement QOpenGLBuffer ::PixelPackBuffer: Is this possible to keep a Offscreen render in memory and pass it as texture on a future shader call? Sure. Render to your QOpenGLFramebufferObject, release it, and it has a texture() method so you can use it as a texture in a render somewhere else.

I resolve it!

@wrosecrans Yes, I'm rendering using a QOffscreenSurface on a separate thread then displaying the frame on a QWindow on the GUI thread, but it makes sense to have all of the video rendering on the same thread so we're not interfering with the GUI thread. I've successfully tested this concept with GLFW-based OpenGL windows and QT UI. I doubt Vulkan will be any different since from what I understand the render window will still need to sit on the GUI thread.

Ok, I figured out how to use updateData properly. Leaving this here for posterity. For my use case, I needed to upload a grayscale 3D texture but this should be the same for any texture type: using namespace Qt3DRender; QTexture * tex = new QTexture3D; tex->setSize(vol.width(), vol.height(), vol.depth()); tex->setFormat(QTexture3D::TextureFormat::R8_UNorm); tex->setLayers(1); tex->setGenerateMipMaps(false); QTextureWrapMode wm(QTextureWrapMode::ClampToBorder); tex->setWrapMode(wm); data->setFormat(QOpenGLTexture::TextureFormat::R8_UNorm); data->setWidth(vol.width()); data->setHeight(vol.height()); data->setDepth(vol.depth()); data->setMipLevels(1); data->setLayers(1); data->setPixelFormat(QOpenGLTexture::PixelFormat::Red); data->setPixelType(QOpenGLTexture::PixelType::UInt8); QByteArray qba; qba.setRawData((const char *)vol.dataUint8(), static_cast<int>(vol.byteSize())); data->setData(qba, 1); QTextureDataUpdate update; update.setX(0); update.setY(0); update.setData(data); tex->updateData(update); Qt3DRender::QParameter p(); p.setName("volume"); // the uniform name in the shader p.setValue(QVariant::fromValue(tex)); QMaterial * mat = ...; // queried from the qml mat->addParameter(p); The Qt3D test for texture update was pretty helpful in figuring this out.

Are you thinking of the new RHI abstraction layer ?

Hi this might be due to the new Qt RHI, an abstraction over OpenGL, Vulkan, Metal, DirectX (might have forgotten some or mentioned some that shouldn't be). Qt Quick will no longer be speaking directly to OpenGL but there will be an abstraction on top of it to ensure compatility with other graphical stacks. You will get more info about this by googling around or watching this webinar : https://www.youtube.com/watch?v=eInD6H4XrpM

@ebatsin is your QQuickItem in a QQuickWindow? I guess it is the case if you are not using a QQuickWidget. Calling QQuickWindow::openglContext() on your item's window should give you what you need.

@oria66 see https://github.com/qt/qtquick3d/blob/14b434d12786ab1b6b90dabb38b0fe4001e54c6a/tests/auto/quick3d/qquick3dgeometry/tst_qquick3dgeometry.cpp

You should not call GUI function from a different thread. Signal and slots is the way to go.

Yup, looks like you've spotted it -- in particular, the specular term is materialSpecularColor * lightStrength * pow(cosAlpha, 10) That's the part that will contribute to the shininess of the material. And, materialSpecularColor comes from highp vec3 materialSpecularColor = lightColor.rgb; which is to say that the materialSpecularColor in that shader is not really a property of the material since the value is just from the lighting. Not really the clearest name if you are trying to pick it apart. This type of look is traditionally used for a plastic type material which has white highlights for white lights regardless of the color of the material. If you want something a bit more metallic looking, you can multiply the lightcolor.rgb by some actual material color. This lighting model is a bit of an old school aphysical hack, so you may want to look into to more modern physically based lighting models if you are looking for more realism. But the classic specular + diffuse model is way simpler to understand even if it isn't super accurate to the real world.

@sierdzio Ok, i see, I'm using open source. Maybe I need to change a deviece or try to use qwidget. Thanks for your patience.

@Julius-Aetherwing Sorry, I'm not familiar with OpenGL, too. But you can find some example code in Qt install dir, something like "C:\Qt\Examples\Qt-5.12.6\opengl" The docs of the examples can be found in All Qt Examples

@fcarney said in Poor quality colored emojis, how to fix that?: font.hintingPreference: Font.PreferNoHinting Wow excellent! This resolved my issue. Thank you very much

That's becoming even more weird: GLFunctions::gl30()->glGenTextures(1, &texHandle); returns 2 in texHandle. but for using it in the shader it works for a value of 1 in the bindTexture GLFunctions::gl30()->glBindTexture(GL_TEXTURE_2D, texHandle-1); This is driving me nuts.

What processing are you doing with OpenCV ?

For those who end up looking here : It seems that this is a Qt 5.6 bug, it works fine in Qt 5.14 Edit : The bug itself may be related to the use of QSurfaceFormat::setDefaultFormat after application startup : Qt 5.14 warned me about a piece of code I was not aware saying that it could cause context sharing issues. Having it removed seems to have solved the bug when i got back to Qt 5.6

@Chris-Kawa ah ok that worked and learned something new thanks

Thanks, so that is what "OpenGL-based composing" does under the hood! Is there any way to synchronize/batch my QGraphicsView updates, so that there is less of a performance impact? Or would it be preferable to switch the QGraphicsView-based widgets over to render their scene via OpenGL?

@mcleary @Patterson Its been a long time since this thread was started but I think it makes sense to post the updates here. I put together a basic project with native OpenGL and also using Qt's QOffscreenSurface and QWindow classes. The github repository can be found here: https://github.com/rtavakko/VSync In this setup there is a single renderer that render to an FBO texture using its own thread and 3 display windows that share the GUI thread with all of them sharing OpenGL contexts together. On my machine I see a clear lag with VSync on when there are 2 or more display windows. Turning VSync off clears the lag. The calcualted FPS values do not indicate any issues in either case which doesn't help but I see a clear difference when I start the application with say 2 displays and then close one of them. I have my internal laptop display and an RCA RT1970 (both running at 1920x1080 - 60Hz) and both connected to my integrated Intel HD 530 GPU. My discrete GPU is an NVIDIA GeForce GTX 950M and I'm running Qt 5.12.0 MSVC2017 64 bit. I'll dig in further myself also with other machines but if anyone has any ideas on what could be causing this, it would really help debug further. Cheers!

@yijia Maybe with different QLabel and QPixmap or QPicture https://doc.qt.io/qt-5/qlabel.html#pixmap-prop https://doc.qt.io/qt-5/qlabel.html#setPicture