- Update to version 21.3.5:
* New features:
+ VP9 Encode: Added WebRTC mode.
+ Samples:
. Added "VuiTC" option to set VUI TransferCharacteristics in
sample_multi_transcode.
. Added the verification of input params before used in Init
for sample_encode
+ Misc:
. Added support of DRM_FORMAT_NV12 for console mode
rendering.
. Added runtimes support matrix for Media SDK and oneVPL GPU
Runtime.
+ Software requirements:
. Libdrm 2.4.84 or later
. Kernel 4.14 or later (5.4 recommended, consult kernel
support matrix wiki page for details)
+ Known issues: Kernel 5.0 have known issue with endurance on
Skylake see
https://bugs.freedesktop.org/show_bug.cgi?id=110285 for
details.
* Limited support on certain platforms:
+ MPEG-2 encode is not supported on Apollo Lake
+ H.264 Flexible Encode Infrastructure only supported on
Broadwell and Skylake
+ Multi Frame Encode (MFE), HEVC Flexible Encode Infrastructure
only supported on Skylake
+ VP9 decoder is supported starting from Kabylake platform
+ VP9 encoder is supported starting from Icelake platform
+ SW fallback is unsupported for all components but MJPEG
+ Keem Bay requires a VPU runtime library
+ The following features are supported by Keem Bay runtime and
are not supported by Gen graphics runtime:
. mfxExtInsertHeaders
. mfxExtEncoderIPCMArea
OBS-URL: https://build.opensuse.org/request/show/934596
OBS-URL: https://build.opensuse.org/package/show/X11:XOrg/libmfx?expand=0&rev=37
- update to 20.3.0:
* Added Keem Bay platform support
* Added AV1 decode support
* Added implicit Motion-Compensated Temporal Filter (MCTF) to improve coding quality on certain scenarios
* Improved CodecLevel initialization from rate control parameters
* Added support of 4k streams with look-ahead bitrate control mode
* Enabled QPOffset by default
* Improved efficiency of FPQ
* Enabled QP modulation (Icelake+)
* Added 16k support
* Added A2RGB10 input in case of passthru copy
* Added -dec::sys flag for setting of the output memory to system type in sample_multi_transcode
* Added p016/y216 support in sample_encode
* Added i420/nv12 support in sample_multi_transcode
* Fixed chroma size for Y210/Y216 formats in sample_deocde
* Added handle for iterpolation methods for resize in sample_vpp
* Added support for discrete Intel Xe graphics codenamed DG1 and SG1
* Added support for Rocketlake platform
* Actualized Elkhartlake device id support list
OBS-URL: https://build.opensuse.org/request/show/840908
OBS-URL: https://build.opensuse.org/package/show/X11:XOrg/libmfx?expand=0&rev=28
* Samples: Fixed#2190. rotate_opencl plugin switched to recent
version of OpenCL C++ headers, validated with OpenCL-CLHPP
v2.0.12 and OpenCL-Headers v2020.06.16 / Intel® SDK for OpenCL™
Applications for Windows* 2020.2.487.
- Update to version 20.2.0
* bugfixes and improvements
* includes: Build system does not enforce specific instruction
set (issue#2047, boo#1167346)
- supersedes u_buildfix-for-gcc10.patch
OBS-URL: https://build.opensuse.org/package/show/X11:XOrg/libmfx?expand=0&rev=24
* X11_DRI3
* WAYLAND
* TEXTLOG
* STAT
* TESTS
* TOOLS
- ITT and KERNELS features remain disabled for now since libITT
and cm compiler is not available on (open)SUSE right now
- gcc7 needed for building tests on sle12-sp5
- moved documentation and binaries to main package; let require
the library subpackage the main package
OBS-URL: https://build.opensuse.org/package/show/X11:XOrg/libmfx?expand=0&rev=11
