[FFmpeg-devel] [PATCH] avcodec/libsvtav1: use larger of bit rate and max rate for buffer size

[FFmpeg-devel] [PATCH] avcodec/libsvtav1: use larger of bit rate and max rate for buffer size

[Jan Ekström]

Generally if maxrate is set, the calculation should be maxrate over
bufsize. This additionally enables CRF + maxrate & bufsize usage.

In order to keep negative values from enabling zero to be treated
as larger and causing a division by zero, check that one of the
variables is larger than zero.
---
 libavcodec/libsvtav1.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/libavcodec/libsvtav1.c b/libavcodec/libsvtav1.c
index 56e1e22b51..9174e2753c 100644
--- a/libavcodec/libsvtav1.c
+++ b/libavcodec/libsvtav1.c
@@ -184,8 +184,10 @@ static int config_enc_params(EbSvtAv1EncConfiguration *param,
         param->min_qp_allowed       = avctx->qmin;
     }
     param->max_bit_rate             = avctx->rc_max_rate;
-    if (avctx->bit_rate && avctx->rc_buffer_size)
-        param->maximum_buffer_size_ms = avctx->rc_buffer_size * 1000LL / avctx->bit_rate;
+    if ((avctx->bit_rate > 0 || avctx->rc_max_rate > 0) && avctx->rc_buffer_size)
+        param->maximum_buffer_size_ms =
+            avctx->rc_buffer_size * 1000LL /
+            FFMAX(avctx->bit_rate, avctx->rc_max_rate);
 
     if (svt_enc->crf > 0) {
         param->qp                   = svt_enc->crf;
@@ -302,7 +304,8 @@ static int config_enc_params(EbSvtAv1EncConfiguration *param,
     avctx->bit_rate       = param->rate_control_mode > 0 ?
                             param->target_bit_rate : 0;
     avctx->rc_max_rate    = param->max_bit_rate;
-    avctx->rc_buffer_size = param->maximum_buffer_size_ms * avctx->bit_rate / 1000LL;
+    avctx->rc_buffer_size = param->maximum_buffer_size_ms *
+                            FFMAX(avctx->bit_rate, avctx->rc_max_rate) / 1000LL;
 
     if (avctx->bit_rate || avctx->rc_max_rate || avctx->rc_buffer_size) {
         AVCPBProperties *cpb_props = ff_add_cpb_side_data(avctx);
-- 
2.39.2

_______________________________________________
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".

Re: [FFmpeg-devel] [PATCH] avcodec/libsvtav1: use larger of bit rate and max rate for buffer size

[Jan Ekström]

On Thu, Mar 9, 2023 at 8:57 PM Jan Ekström <jeebjp@gmail.com> wrote:
>
> Generally if maxrate is set, the calculation should be maxrate over
> bufsize. This additionally enables CRF + maxrate & bufsize usage.
>
> In order to keep negative values from enabling zero to be treated
> as larger and causing a division by zero, check that one of the
> variables is larger than zero.
> ---
>  libavcodec/libsvtav1.c | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)

The idea and patch was OK'd by Christopher from SVT-AV1 and James
noted that it seemed OK on IRC.

Thus will be pulling this in in the evening unless there are concerns raised.

Jan
_______________________________________________
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".

Re: [FFmpeg-devel] [PATCH] avcodec/libsvtav1: use larger of bit rate and max rate for buffer size

[Jan Ekström]

On Mon, Mar 13, 2023 at 11:57 AM Jan Ekström <jeebjp@gmail.com> wrote:
>
> On Thu, Mar 9, 2023 at 8:57 PM Jan Ekström <jeebjp@gmail.com> wrote:
> >
> > Generally if maxrate is set, the calculation should be maxrate over
> > bufsize. This additionally enables CRF + maxrate & bufsize usage.
> >
> > In order to keep negative values from enabling zero to be treated
> > as larger and causing a division by zero, check that one of the
> > variables is larger than zero.
> > ---
> >  libavcodec/libsvtav1.c | 9 ++++++---
> >  1 file changed, 6 insertions(+), 3 deletions(-)
>
> The idea and patch was OK'd by Christopher from SVT-AV1 and James
> noted that it seemed OK on IRC.
>
> Thus will be pulling this in in the evening unless there are concerns raised.

Applied as fba9d9609fe00c8c4d04b76f9ecdd274aad318dd .

Jan
_______________________________________________
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".