-
-
Notifications
You must be signed in to change notification settings - Fork 88
HW accelerated video decode matrix
Consti10 edited this page Jan 29, 2023
·
4 revisions
RPI:
- h264: hw decode & hw compose, low latency. needs "fkms" enabled on rpi 4
- h265: hw decode & gpu compose (Pi 4 only), medium latency
- mjpeg: sw decode & gpu compose, medium latency
X86:
- h264: sw decode & gpu compose, low latency (on devices with good cpu compute power)
- h265: sw decode & gpu compose, low latency (on devices with good cpu compute power)
- mjpeg: sw decode & gpu compose, low latency (on devices with good cpu compute power)
Other platforms: sw decode & gpu compose should always be available as fallback.
You can also always use QOpenHD with a transparent background and start your manual video decode pipeline in a window behind. (requires X server)