hamidi, there are definitely some problems getting it to use the hardware
decoder. For us we have such a problem on iOS builds.
We have to use code something like this to be sure it finds a hardware decoder
...
#include <libavcodec/videotoolbox.h>
static enum AVPixelFormat find_pixel_format(struct AVCodecContext *s, const
enum AVPixelFormat *fmt) {
while (*fmt != AV_PIX_FMT_NONE) {
if (*fmt == AV_PIX_FMT_VIDEOTOOLBOX) {
if (s->hwaccel_context == NULL) {
int result = av_videotoolbox_default_init(s);
if (result < 0) {
printf(" the 'SPS/PPS problem' has happened.\n");
printf(" you are NOT getting hardware decoding on
this stream.\n");
return s->pix_fmt;
}
else {
printf(" the test was OK....\n");
}
}
return *fmt;
}
++fmt;
}
return s->pix_fmt;
}
the problem seems to relate to finding the SPS/PPS information on a variety of
sources
see,
https://stackoverflow.com/q/25787347/294884
<https://stackoverflow.com/q/25787347/294884>
a couple yrs ago there was a patch on the 264 problem (which I think was not
incorporated..)
https://ffmpeg.org/pipermail/ffmpeg-devel/2016-June/194842.html
<https://ffmpeg.org/pipermail/ffmpeg-devel/2016-June/194842.html>
we really have not been able to solve the issue in the 265 case.
If this info is relevant to your case, I hope it helps …
> On Jan 30, 2019, at 7:58 AM, hamidi <[email protected]> wrote:
>
> I got the latest version binaries of ffmpeg from here
> <https://ffmpeg.zeranoe.com/builds/>. When I examine CPU and GPU usages when
> I play a video by its ffplay, I see that GPU is used during play. Not much
> using of CPU also indicates it. But when I get the latest version sources
> from the original site, I can't use GPU. ….
_______________________________________________
Libav-user mailing list
[email protected]
http://ffmpeg.org/mailman/listinfo/libav-user
To unsubscribe, visit link above, or email
[email protected] with subject "unsubscribe".
