Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
 help / color / mirror / Atom feed
From: "Carotti, Elias via ffmpeg-devel" <ffmpeg-devel@ffmpeg.org>
To: "ffmpeg-devel@ffmpeg.org" <ffmpeg-devel@ffmpeg.org>
Cc: "Carotti, Elias" <eliascrt@amazon.it>
Subject: Re: [FFmpeg-devel] libavc/libx264: add support to propagate SSE values through encoder stats
Date: Fri, 13 Oct 2023 16:35:14 +0000
Message-ID: <083eb9e5b4bd3f68ec49ee63556376f8917598ee.camel@amazon.it> (raw)
In-Reply-To: <169720660959.32606.9316543522345619707@lain.khirnov.net>

[-- Attachment #1: Type: text/plain, Size: 2119 bytes --]

On Fri, 2023-10-13 at 16:16 +0200, Anton Khirnov wrote:
> CAUTION: This email originated from outside of the organization. Do
> not click links or open attachments unless you can confirm the sender
> and know the content is safe.
> 
> 
> 
> Quoting Carotti, Elias via ffmpeg-devel (2023-10-11 12:54:21)
> > diff --git a/libavcodec/libx264.c b/libavcodec/libx264.c
> > index 77a9f173b4..85bd870f5d 100644
> > --- a/libavcodec/libx264.c
> > +++ b/libavcodec/libx264.c
> > @@ -726,7 +726,39 @@ FF_ENABLE_DEPRECATION_WARNINGS
> > 
> >      pkt->flags |= AV_PKT_FLAG_KEY*pic_out.b_keyframe;
> >      if (ret) {
> > -        ff_side_data_set_encoder_stats(pkt, (pic_out.i_qpplus1 -
> > 1) * FF_QP2LAMBDA, NULL, 0, pict_type);
> > +        const AVPixFmtDescriptor *pix_desc =
> > av_pix_fmt_desc_get(csp_to_pixfmt(pic_out.img.i_csp));
> 
> There's a problem here - we do not handle all values of i_csp.
> E.g. we have no equivalent of X264_CSP_NV12 | X264_CSP_HIGH_DEPTH,
> which
> x264 will use for YUV420P10 input.
> 
> The best solution is probably to use AVCodecContext.pix_fmt and
> assume
> that x264 doesn't do any nontrivial (i.e. other than interleaving and
> such) pixel format transformations internally.
> 


I see. Wouldn't not outputting the SSE values when csp_to_pixfmt()
returns AV_PIX_FMT_NONE work better? 
That wouldn't be worse than it is today (meaning that right now we
don't get those values for any pix_fmt).
Anyway, I did as you suggested and used AVCodecContext.pix_fmt.


> > <snip>
> > +            av_log(ctx, AV_LOG_DEBUG, "PSNR values from libx264:
> > %.3f %.3f %.3f.\n",
> > +                   pic_out.prop.f_psnr[0], pic_out.prop.f_psnr[1],
> > pic_out.prop.f_psnr[2]);
> 
> In my tests libx264 prints these values by itself, so this seems
> redundant.

removed.




NICE SRL, viale Monte Grappa 3/5, 20124 Milano, Italia, Registro delle Imprese di Milano Monza Brianza Lodi REA n. 2096882, Capitale Sociale: 10.329,14 EUR i.v., Cod. Fisc. e P.IVA 01133050052, Societa con Socio Unico



[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-avcodec-libx264-Add-the-SSE-computation-for-libx264.patch --]
[-- Type: text/x-patch; name="0001-avcodec-libx264-Add-the-SSE-computation-for-libx264.patch", Size: 2210 bytes --]

From b702fcd76cf0626f75a941875f31add50d08894d Mon Sep 17 00:00:00 2001
From: Elias Carotti <eliascrt _at_ amazon _dot_ .it>
Date: Fri, 15 Sep 2023 20:05:43 +0200
Subject: [PATCH] avcodec/libx264: Add the SSE computation for libx264.

Since libx264 only provides a per-frame per-channel PSNR, this is inverted to get back the SSE.
---
 libavcodec/libx264.c | 31 ++++++++++++++++++++++++++++++-
 1 file changed, 30 insertions(+), 1 deletion(-)

diff --git a/libavcodec/libx264.c b/libavcodec/libx264.c
index 77a9f173b4..6ebe210039 100644
--- a/libavcodec/libx264.c
+++ b/libavcodec/libx264.c
@@ -726,7 +726,36 @@ FF_ENABLE_DEPRECATION_WARNINGS

     pkt->flags |= AV_PKT_FLAG_KEY*pic_out.b_keyframe;
     if (ret) {
-        ff_side_data_set_encoder_stats(pkt, (pic_out.i_qpplus1 - 1) * FF_QP2LAMBDA, NULL, 0, pict_type);
+        const AVPixFmtDescriptor *pix_desc = av_pix_fmt_desc_get(ctx->pix_fmt);
+        int error_count = 0;
+        int64_t *errors = NULL;
+        int64_t sse[3] = {0};
+
+        if (ctx->flags & AV_CODEC_FLAG_PSNR) {
+            double scale[3] = { 1,
+                (double)(1 << pix_desc->log2_chroma_h) * (1 << pix_desc->log2_chroma_w),
+                (double)(1 << pix_desc->log2_chroma_h) * (1 << pix_desc->log2_chroma_w),
+            };
+
+            error_count = pix_desc->nb_components;
+
+            for (int i = 0; i < pix_desc->nb_components; ++i) {
+                double max_value = (double)(1 << pix_desc->comp[i].depth) - 1.0;
+                double plane_size = ctx->width * (double)ctx->height / scale[i];
+
+                /* psnr = 10 * log10(max_value * max_value / mse) */
+                double mse = (max_value * max_value) / pow(10, pic_out.prop.f_psnr[i] / 10.0);
+
+                /* SSE = MSE * width * height / scale -> because of possible chroma downsampling */
+                sse[i] = (int64_t)floor(mse * plane_size + .5);
+            };
+
+            errors = sse;
+        }
+
+        ff_side_data_set_encoder_stats(pkt, (pic_out.i_qpplus1 - 1) * FF_QP2LAMBDA,
+                                       errors, error_count, pict_type);
+
         if (wallclock)
             ff_side_data_set_prft(pkt, wallclock);
     }
--
2.34.1


[-- Attachment #3: Type: text/plain, Size: 251 bytes --]

_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

  reply	other threads:[~2023-10-13 16:35 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-23 10:04 Carotti, Elias via ffmpeg-devel
2023-10-02 17:35 ` Carotti, Elias via ffmpeg-devel
2023-10-10 10:54   ` Anton Khirnov
2023-10-11 10:54     ` Carotti, Elias via ffmpeg-devel
2023-10-13 14:16       ` Anton Khirnov
2023-10-13 16:35         ` Carotti, Elias via ffmpeg-devel [this message]
2023-10-19 11:50           ` Anton Khirnov
2023-10-19 11:48 ` Anton Khirnov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=083eb9e5b4bd3f68ec49ee63556376f8917598ee.camel@amazon.it \
    --to=ffmpeg-devel@ffmpeg.org \
    --cc=eliascrt@amazon.it \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
		ffmpegdev@gitmailbox.com
	public-inbox-index ffmpegdev

Example config snippet for mirrors.


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git