[Development] [QML][EGLFS] Bits depth

[Laszlo Agocs]

Hi Denis,


QSurfaceFormat::setDefaultFormat() with a format of red/blue/green set to 8, or alternatively setting QT_QPA_EGLFS_FORCE888 is indeed what you need. The default format with unspecified sizes results in choosing 16 over 32 bit configs when both are offered, that's normal.


Check with QSG_INFO what you get. It may either be 8880 or 8888, depending on what is offered by the EGL implementation. As for the 24 bit depth, regardless of what X or EGL says, under the hood it is quite likely that you have 32 bit framebuffers anyway, true packed 24 bit ones are probably rare.


As for the "other bad issues", the flickering, that is the real issue here, but that's likely caused by something else outside of Qt's scope.


Best regards,

Laszlo

________________________________
From: Denis Shienkov <denis.shienkov at gmail.com>
Sent: Thursday, August 4, 2016 4:48:40 PM
To: development at qt-project.org; Laszlo Agocs
Subject: [QML][EGLFS] Bits depth

Hi guys,

I have use Qt 5.7 on Linux Embedded board (Toradex Apalis T30) with the EGLFS via X11 platform (NVidia drivers).

My QML application uses 24-bit PNG images. But I see that these images is looks like 16-bit (I see the gradients borders, and it looks ugly),
howewer my X11 uses DefaultDepth == 24!!!

Also this command:

{quote}
root at apalis-t30:~/deploy-apps# xdpyinfo | grep 'depth of root'
  depth of root window:    24 planes
{quote}

shows 24-bit mode too.

Next, I have added this debug env variable:

{quote}
 export QSG_INFO=1
{quote}

to see what happens when my QML app starts:

{quote}
...
qt.scenegraph.general: threaded render loop
qt.scenegraph.general: Using sg animation driver
qt.scenegraph.general: Animation Driver: using vsync: 16.67 ms
qt.scenegraph.general: Animation Driver: using vsync: 16.67 ms
qt.scenegraph.general: texture atlas dimensions: 2048x2048
qt.scenegraph.general: R/G/B/A Buffers:    5 6 5 0
qt.scenegraph.general: Depth Buffer:       16
qt.scenegraph.general: Stencil Buffer:     8
qt.scenegraph.general: Samples:            0
qt.scenegraph.general: GL_VENDOR:          NVIDIA Corporation
qt.scenegraph.general: GL_RENDERER:        NVIDIA Tegra 3
...
{quote}

I see that OpenGL (yes?) initializes with the "R/G/B/A Buffers:    5 6 5 0" mode,
that is 16-bit depth!!! O_O

Next, I have tried to change the FB depth via:

{quote}
  fbset -fb /dev/fb1 -depth 24
{quote}

I see that FB changes own settings, but nothing changes when my app starts,
it still shows images ugly, as 16-bit depth.

Also I have tried to use QT_QPA_EGLFS_DEPTH variable (with 16, 24 or 32),
but nothing changes, not helps.

Helps only the QT_QPA_EGLFS_FORCE888 variable, but it introduces others bad issues,
like the whole/half HDMI display flickering.. So, I can not to use QT_QPA_EGLFS_FORCE888..

So, is there are any tricks to use 24-bit depth?

BR,
Denis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.qt-project.org/pipermail/development/attachments/20160804/76385eef/attachment.html>