Libyami is core video library to accelerate decoding/encoding based on libva.
Libyami can be treated as development kit for VAAPI. It is developed by C++
language, with simple/efficient APIs, without external dependencies.
libyami targets two type of usage:
- Hardware accelerated codec is the major gap (there is existing media
framework ), for example: Chromeos/Android, maybe ffmpeg.
- Usage is simple enough that codec is the major interest, for example:
remote desktop client/server, video surveillance.
So far H.264 decode/encode, VP8 and JPEG decode are supported. VP8/JPEG encode
support is under development.
There are rich tests/examples in yami for a try. The tests can run without
Window system (X11/Wayland etc),
some modes of video output are supported: including raw data dump and some
modes of texture (TFP, drm flink name, dma_buf).
More details see: https://github.com/01org/libyami/wiki
The future release plan are:
Nov: v0.11:
V4L2 VDA/VEA
Dec: v0.2
Vp8 and jpeg encoder(depends on driver release)
unified input/output buffer of yami
Feb'15: v0.3
Vp9 decoder
transcoding support with unified input/output buffer
camera dma_buf support, camera with jpeg input
use yami in ffmpeg for hw codec
Future:
h265 decoder
legacy codec: mpeg2 decoder/encoder, vc1 decoder, mpeg4/h263
decoder/encoder
Known issues:
some jpeg file similarity <0.99 (~0.98) after decoding
https://github.com/01org/libyami/issues/108
playback performance through gst-omx is not good:
https://github.com/01org/libyami/issues/34
Encodeing quality through gst-omx is not good:
https://github.com/01org/libyami/issues/36
_______________________________________________
Libva mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/libva
