[FFmpeg-devel] Encoding image
I have a PNG image converted to YUV 420P and I need to encode it to a specific RTMPS URI, with fps 10. The encoding process is quite heavy for the CPU. In this specific case the frame is always the same, re-encoded at a rate basing on fps. Is it necessary to call "avcodec_send_frame / avcodec_receive_packet" for each frame or there is a better way given that the frame to encode does not change? Thank you ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-devel] Performances improvement in "image_copy_plane"
On Wednesday, July 13, 2022 at 06:16:15 PM GMT+2, James Almer
wrote:
On 7/13/2022 12:54 PM, Marco Vianini wrote:
> Sorry, my mail client was using html format.
> I hope now the mail will be sent correctly.
>
>
> You can get a very big improvement of performances in the special (but very
> likely) case of: "(dst_linesize == bytewidth && src_linesize == bytewidth)"
>
> In this case in fact We can "Coalesce rows", that is using ONLY ONE MEMCPY,
> instead of a smaller memcpy for every row (that is looping for height times).
>
> Code:
> "
> static void image_copy_plane(uint8_t *dst, ptrdiff_t dst_linesize,
> const uint8_t *src, ptrdiff_t src_linesize,
> ptrdiff_t bytewidth, int height)
> {
> if (!dst || !src)
> return;
> av_assert0(abs(src_linesize) >= bytewidth);
> av_assert0(abs(dst_linesize) >= bytewidth);
>
> /// MY PATCH START
> /// Coalesce rows.
> if (dst_linesize == bytewidth && src_linesize == bytewidth) {
> bytewidth *= height;
> height = 1;
> src_linesize = dst_linesize = 0;
> }
> /// MY PATCH STOP
>
> for (;height > 0; height--) {
> memcpy(dst, src, bytewidth);
> dst += dst_linesize;
> src += src_linesize;
> }
> }
> "
>
>
> I did following tests on Windows 10 64bit.
> I compiled code in Release.
> I copied my pc camera frames 1000 times (resolution 1920x1080):
>
> With Coalesce:
> copy_cnt=100 size=1920x1080 tot_time_copy(us)=36574 (average=365.74)
> copy_cnt=200 size=1920x1080 tot_time_copy(us)=78207 (average=391.035)
> copy_cnt=300 size=1920x1080 tot_time_copy(us)=122170(average=407.233)
> copy_cnt=400 size=1920x1080 tot_time_copy(us)=163678(average=409.195)
> copy_cnt=500 size=1920x1080 tot_time_copy(us)=201872(average=403.744)
> copy_cnt=600 size=1920x1080 tot_time_copy(us)=246174(average=410.29)
> copy_cnt=700 size=1920x1080 tot_time_copy(us)=287043(average=410.061)
> copy_cnt=800 size=1920x1080 tot_time_copy(us)=326462(average=408.077)
> copy_cnt=900 size=1920x1080 tot_time_copy(us)=356882(average=396.536)
> copy_cnt=1000 size=1920x1080 tot_time_copy(us)=394566(average=394.566)
>
> Without Coalesce:
> copy_cnt=100 size=1920x1080 tot_time_copy(us)=44303 (average=443.03)
> copy_cnt=200 size=1920x1080 tot_time_copy(us)=100501(average=502.505)
> copy_cnt=300 size=1920x1080 tot_time_copy(us)=150097(average=500.323)
> copy_cnt=400 size=1920x1080 tot_time_copy(us)=201010(average=502.525)
> copy_cnt=500 size=1920x1080 tot_time_copy(us)=256818(average=513.636)
> copy_cnt=600 size=1920x1080 tot_time_copy(us)=303273(average=505.455)
> copy_cnt=700 size=1920x1080 tot_time_copy(us)=359152(average=513.074)
> copy_cnt=800 size=1920x1080 tot_time_copy(us)=414413(average=518.016)
> copy_cnt=900 size=1920x1080 tot_time_copy(us)=465315(average=517.017)
> copy_cnt=1000 size=1920x1080 tot_time_copy(us)=520381(average=520.381)
>
>
> I think the results are very good.
> What do you think about?
It looks like a good speed up, but we need a patch created with git
format-patch that can be applied to the source tree to properly review
this. Can you send that?
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
I generated the eml file with "git format-patch" (see attachment).
Is it ok for You?
Thanks--- Begin Message ---
Signed-off-by: Marco Vianini
---
libavutil/imgutils.c | 8
1 file changed, 8 insertions(+)
diff --git a/libavutil/imgutils.c b/libavutil/imgutils.c
index 9ab5757cf6..9ccb398a3b 100644
--- a/libavutil/imgutils.c
+++ b/libavutil/imgutils.c
@@ -349,6 +349,14 @@ static void image_copy_plane(uint8_t *dst, ptrdiff_t
dst_linesize,
return;
av_assert0(FFABS(src_linesize) >= bytewidth);
av_assert0(FFABS(dst_linesize) >= bytewidth);
+
+if (dst_linesize == bytewidth && src_linesize == bytewidth) {
+/** Coalesce rows in this specific case, for perfomances improvement */
+bytewidth *= height;
+height = 1;
+src_linesize = dst_linesize = 0;
+}
+
for (;height > 0; height--) {
memcpy(dst, src, bytewidth);
dst += dst_linesize;
--
2.30.0.windows.2
--- End Message ---
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-devel] Request: make framepool visible externally
On Wednesday, July 13, 2022 at 02:02:18 PM GMT+2, James Almer wrote: On 7/13/2022 5:03 AM, Marco Vianini wrote: > Actually I was speaking about framepool, and not bufferpool. > framepool is intended to get an "AVFrame *" from a pool, by > "ff_frame_pool_get". > At the moment it is available only internally to "libavfilter". > It permits an important improvement on performances, by using a pool. So it > should be very nice if I could use it in my own code. > To be possible framepool.h should become a public header. > Thank You Please, do not top post in this list. I know you want AVFrames. What i mean is that you should use the existing buffer pool API to achieve that. Write your own function that returns a freshly allocated AVFrame that fills the buf[] and data[] fields with buffers returned by av_buffer_pool_get() instead of av_buffer_alloc(). ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe". Sorry, I misunderstood what you mean with previous answer. However Yes, surely I can write my own code (replicating what already done by framepool). Thank You ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-devel] Performances improvement in "image_copy_plane"
On Wednesday, July 13, 2022 at 05:08:27 PM GMT+2, Paul B Mahol
wrote:
On Wed, Jul 13, 2022 at 5:02 PM Marco Vianini <
marco_vianini-at-yahoo...@ffmpeg.org> wrote:
> I did following tests on Windows 10 64bit.I compiled code in Release.
> I copied my pc camera frames 1000 times (resolution 1920x1080):
> With Coalesce (MY PATCH):copy_cnt=100 size=1920x1080
> tot_time_copy(us)=36574 (average=365.74)copy_cnt=200 size=1920x1080
> tot_time_copy(us)=78207 (average=391.035)copy_cnt=300 size=1920x1080
> tot_time_copy(us)=122170(average=407.233)copy_cnt=400 size=1920x1080
> tot_time_copy(us)=163678(average=409.195)copy_cnt=500 size=1920x1080
> tot_time_copy(us)=201872(average=403.744)copy_cnt=600 size=1920x1080
> tot_time_copy(us)=246174(average=410.29)copy_cnt=700 size=1920x1080
> tot_time_copy(us)=287043(average=410.061)copy_cnt=800 size=1920x1080
> tot_time_copy(us)=326462(average=408.077)copy_cnt=900 size=1920x1080
> tot_time_copy(us)=356882(average=396.536)copy_cnt=1000 size=1920x1080
> tot_time_copy(us)=394566(average=394.566)
> Without Coalesce:copy_cnt=100 size=1920x1080 tot_time_copy(us)=44303
> (average=443.03)copy_cnt=200 size=1920x1080
> tot_time_copy(us)=100501(average=502.505)copy_cnt=300 size=1920x1080
> tot_time_copy(us)=150097(average=500.323)copy_cnt=400 size=1920x1080
> tot_time_copy(us)=201010(average=502.525)copy_cnt=500 size=1920x1080
> tot_time_copy(us)=256818(average=513.636)copy_cnt=600 size=1920x1080
> tot_time_copy(us)=303273(average=505.455)copy_cnt=700 size=1920x1080
> tot_time_copy(us)=359152(average=513.074)copy_cnt=800 size=1920x1080
> tot_time_copy(us)=414413(average=518.016)copy_cnt=900 size=1920x1080
> tot_time_copy(us)=465315(average=517.017)copy_cnt=1000 size=1920x1080
> tot_time_copy(us)=520381(average=520.381)
> I think the results are very good.What do you think about?
> Thank You
>
>
First stop top posting.
Where is patch?
>
> Il mercoledì 13 luglio 2022 11:52:23 CEST, Paul B Mahol <
> one...@gmail.com> ha scritto:
>
> On Wed, Jul 13, 2022 at 11:38 AM Marco Vianini <
> marco_vianini-at-yahoo...@ffmpeg.org> wrote:
>
> > You can get a very big improvement of performances in the special (but
> > very likely) case of: "(dst_linesize == bytewidth && src_linesize ==
> > bytewidth)"
> >
> > In this case in fact We can "Coalesce rows", that is using ONLY ONE
> > MEMCPY, instead of a smaller memcpy for every row (that is looping for
> > height times).
> >
> > Code:"static void image_copy_plane(uint8_t *dst, ptrdiff_t
> > dst_linesize, const uint8_t *src, ptrdiff_t
> > src_linesize, ptrdiff_t bytewidth, int
> > height){ if (!dst || !src) return;
> > av_assert0(abs(src_linesize) >= bytewidth);
> av_assert0(abs(dst_linesize)
> > >= bytewidth); // MY PATCH START // Coalesce rows. if (dst_linesize
> > == bytewidth && src_linesize == bytewidth) { bytewidth *= height;
> > height = 1; src_linesize = dst_linesize = 0; }// MY PATCH STOP
> > for (;height > 0; height--) { memcpy(dst, src, bytewidth);
> > dst += dst_linesize; src += src_linesize; }}"
> > What do You think about?Thank You
> >
>
> Show the benchmark numbers.
>
>
> > Marco Vianini
> > ___
> > ffmpeg-devel mailing list
> > ffmpeg-devel@ffmpeg.org
> > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> >
> > To unsubscribe, visit link above, or email
> > ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
> >
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
>
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
>
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
Sorry, my mail client was using html format.
I hope now the mail will be sent correctly.
You can get a very big improvement of performances in the special (but very
likely) case of: "(dst_line
Re: [FFmpeg-devel] Performances improvement in "image_copy_plane"
I did following tests on Windows 10 64bit.I compiled code in Release.
I copied my pc camera frames 1000 times (resolution 1920x1080):
With Coalesce (MY PATCH):copy_cnt=100 size=1920x1080 tot_time_copy(us)=36574
(average=365.74)copy_cnt=200 size=1920x1080 tot_time_copy(us)=78207
(average=391.035)copy_cnt=300 size=1920x1080
tot_time_copy(us)=122170(average=407.233)copy_cnt=400 size=1920x1080
tot_time_copy(us)=163678(average=409.195)copy_cnt=500 size=1920x1080
tot_time_copy(us)=201872(average=403.744)copy_cnt=600 size=1920x1080
tot_time_copy(us)=246174(average=410.29)copy_cnt=700 size=1920x1080
tot_time_copy(us)=287043(average=410.061)copy_cnt=800 size=1920x1080
tot_time_copy(us)=326462(average=408.077)copy_cnt=900 size=1920x1080
tot_time_copy(us)=356882(average=396.536)copy_cnt=1000 size=1920x1080
tot_time_copy(us)=394566(average=394.566)
Without Coalesce:copy_cnt=100 size=1920x1080 tot_time_copy(us)=44303
(average=443.03)copy_cnt=200 size=1920x1080
tot_time_copy(us)=100501(average=502.505)copy_cnt=300 size=1920x1080
tot_time_copy(us)=150097(average=500.323)copy_cnt=400 size=1920x1080
tot_time_copy(us)=201010(average=502.525)copy_cnt=500 size=1920x1080
tot_time_copy(us)=256818(average=513.636)copy_cnt=600 size=1920x1080
tot_time_copy(us)=303273(average=505.455)copy_cnt=700 size=1920x1080
tot_time_copy(us)=359152(average=513.074)copy_cnt=800 size=1920x1080
tot_time_copy(us)=414413(average=518.016)copy_cnt=900 size=1920x1080
tot_time_copy(us)=465315(average=517.017)copy_cnt=1000 size=1920x1080
tot_time_copy(us)=520381(average=520.381)
I think the results are very good.What do you think about?
Thank You
Il mercoledì 13 luglio 2022 11:52:23 CEST, Paul B Mahol
ha scritto:
On Wed, Jul 13, 2022 at 11:38 AM Marco Vianini <
marco_vianini-at-yahoo...@ffmpeg.org> wrote:
> You can get a very big improvement of performances in the special (but
> very likely) case of: "(dst_linesize == bytewidth && src_linesize ==
> bytewidth)"
>
> In this case in fact We can "Coalesce rows", that is using ONLY ONE
> MEMCPY, instead of a smaller memcpy for every row (that is looping for
> height times).
>
> Code:"static void image_copy_plane(uint8_t *dst, ptrdiff_t
> dst_linesize, const uint8_t *src, ptrdiff_t
> src_linesize, ptrdiff_t bytewidth, int
> height){ if (!dst || !src) return;
> av_assert0(abs(src_linesize) >= bytewidth); av_assert0(abs(dst_linesize)
> >= bytewidth); // MY PATCH START // Coalesce rows. if (dst_linesize
> == bytewidth && src_linesize == bytewidth) { bytewidth *= height;
> height = 1; src_linesize = dst_linesize = 0; }// MY PATCH STOP
> for (;height > 0; height--) { memcpy(dst, src, bytewidth);
> dst += dst_linesize; src += src_linesize; }}"
> What do You think about?Thank You
>
Show the benchmark numbers.
> Marco Vianini
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
>
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-devel] Performances improvement in "image_copy_plane"
You can get a very big improvement of performances in the special (but very
likely) case of: "(dst_linesize == bytewidth && src_linesize == bytewidth)"
In this case in fact We can "Coalesce rows", that is using ONLY ONE MEMCPY,
instead of a smaller memcpy for every row (that is looping for height times).
Code:"static void image_copy_plane(uint8_t *dst, ptrdiff_t dst_linesize,
const uint8_t *src, ptrdiff_t src_linesize,
ptrdiff_t bytewidth, int height){ if (!dst || !src)
return; av_assert0(abs(src_linesize) >= bytewidth);
av_assert0(abs(dst_linesize) >= bytewidth); // MY PATCH START // Coalesce
rows. if (dst_linesize == bytewidth && src_linesize == bytewidth) {
bytewidth *= height; height = 1; src_linesize = dst_linesize = 0;
}// MY PATCH STOP
for (;height > 0; height--) { memcpy(dst, src, bytewidth);
dst += dst_linesize; src += src_linesize; }}"
What do You think about?Thank You
Marco Vianini
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-devel] Request: make framepool visible externally
Yes, I know the version 4.1.6 is very old. But I stuck in 4.1.x versions because I need NDI support. In the file "framepool.h" there are only some very high level functions to deal with AVFrame pool: "FFFramePool *ff_frame_pool_video_init(...);FFFramePool *ff_frame_pool_audio_init(...);void ff_frame_pool_uninit(FFFramePool **pool);int ff_frame_pool_get_video_config();int ff_frame_pool_get_audio_config();AVFrame *ff_frame_pool_get(FFFramePool *pool);" In my debug tests I tried to make "framepool.h" public and to use it, and it worked like a charm!With an important improvement on performances (like a pool does). Probably I don't understand what You mean with "It's also not in any way stable, so can break its API at any random point.".Can You please clarify? I think that it is very stable, and contains only high level functions (see above).Then my request to make it public. Thank You Il martedì 12 luglio 2022 19:16:02 CEST, Timo Rothenpieler ha scritto: On 12.07.2022 18:28, Marco Vianini wrote: > Hi all > I'm using Libav libraries (version 4.1.6) to make operations on audio/video > AVFrame: conversions, decoding, encoding, etc. That is a really old version, and you should desperately update. > To improve performances I'd like to use framepool. > So I need to include "libavfilter/framepool.h", but I cannot, because this > file is not exported. > Should be possible to add "framepool.h" to HEADERS in "libavfilter\Makefile" ? > Code:"NAME = avfilterDESC = FFmpeg audio/video filtering library > HEADERS = avfilter.h \ > buffersink.h \ > buffersrc.h \ > version.h \ framepool.h > framepool.h is not a public header, and contains no publicly exported functions. Even if you got the header, you couldn't use the functions since they're not exported from the library. So no, you can't use any of the stuff in there outside of libavfilter. It's also not in any way stable, so can break its API at any random point. ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe". ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
Re: [FFmpeg-devel] Request: make framepool visible externally
Actually I was speaking about framepool, and not bufferpool. framepool is intended to get an "AVFrame *" from a pool, by "ff_frame_pool_get". At the moment it is available only internally to "libavfilter". It permits an important improvement on performances, by using a pool. So it should be very nice if I could use it in my own code. To be possible framepool.h should become a public header. Thank You Il martedì 12 luglio 2022 19:19:34 CEST, James Almer ha scritto: On 7/12/2022 1:28 PM, Marco Vianini wrote: > Hi all > I'm using Libav libraries (version 4.1.6) to make operations on audio/video > AVFrame: conversions, decoding, encoding, etc. > > To improve performances I'd like to use framepool. > So I need to include "libavfilter/framepool.h", but I cannot, because this > file is not exported. > Should be possible to add "framepool.h" to HEADERS in "libavfilter\Makefile" ? > Code:"NAME = avfilterDESC = FFmpeg audio/video filtering library > HEADERS = avfilter.h \ > buffersink.h \ > buffersrc.h \ > version.h \ framepool.h > \ ... " > Thank YouMarco Vianini What you want is to use the bufferpool API in libavutil for your frames. ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe". ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
[FFmpeg-devel] Request: make framepool visible externally
Hi all I'm using Libav libraries (version 4.1.6) to make operations on audio/video AVFrame: conversions, decoding, encoding, etc. To improve performances I'd like to use framepool. So I need to include "libavfilter/framepool.h", but I cannot, because this file is not exported. Should be possible to add "framepool.h" to HEADERS in "libavfilter\Makefile" ? Code:"NAME = avfilterDESC = FFmpeg audio/video filtering library HEADERS = avfilter.h \ buffersink.h \ buffersrc.h \ version.h \ framepool.h \ ... " Thank YouMarco Vianini ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
