[Bug 797143] vaapih265dec: support 422 chroma format hevc decode.
GStreamer (GNOME Bugzilla)
bugzilla at gnome.org
Wed Sep 19 05:41:35 UTC 2018
https://bugzilla.gnome.org/show_bug.cgi?id=797143
--- Comment #11 from Fei <fei.w.wang at intel.com> ---
(In reply to Víctor Manuel Jáquez Leal from comment #10)
> Review of attachment 373678 [details] [review]:
>
> Also this change would need changes here:
>
> https://cgit.freedesktop.org/gstreamer/gstreamer-vaapi/tree/gst/vaapi/
> gstvaapipluginbase.c#n1257
>
> if I understand correctly
>
> ::: gst-libs/gst/vaapi/gstvaapicontext.c
> @@ +140,3 @@
> + else if (cip->chroma_type == GST_VAAPI_CHROMA_TYPE_YUV420)
> + gst_video_info_set_format (&vi, GST_VIDEO_FORMAT_NV12, cip->width,
> + cip->height);
>
> This snippet has a couple problems, for example, if libva<0.34, vi is not
> used, so this computation is useless. Also, not all the possible chroma
> types are considered, it assumes, by hard-code, that the only possible
> context's chroma types are YUV422 and YUV420.
Will fix this.
>
> @@ +150,3 @@
> cip->chroma_type, cip->width, cip->height);
> +#endif
> +
>
> This code also is a bit problematic: we would have to change all the calls
> to gst_vaapi_surface_new() with this. Isn't practical, in my opinion.
>
> I guess this feature request needs deeper changes in gstvaapisurface, a
> refactoring of gst_vaapi_surface_new() and gst_vaapi_surface_new_full(), in
> order to use the newer version depending on the used libva.
Yes, it seems better to refactor gst_vaapi_surface_new, so that don't need to
modify all positions which calls for gst_vaapi_surface_new.
--
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
More information about the gstreamer-bugs
mailing list