[FFmpeg-trac] #7813(avfilter:new): ebur128 filter reports sample peak > 0 dBFS

[FFmpeg]

#7813: ebur128 filter reports sample peak > 0 dBFS
--------------------------------------+------------------------------------
             Reporter:  ahthovaikied  |                    Owner:
                 Type:  defect        |                   Status:  new
             Priority:  normal        |                Component:  avfilter
              Version:  unspecified   |               Resolution:
             Keywords:                |               Blocked By:
             Blocking:                |  Reproduced by developer:  0
Analyzed by developer:  0             |
--------------------------------------+------------------------------------

Comment (by ahthovaikied):

 @cus Thanks for your answer

 Using `-loglevel debug` I see:

 {{{
 [auto_resampler_0 @ 0x5646abd68300] [SWR @ 0x5646abd68780] Using fltp
 internally between filters
 [auto_resampler_0 @ 0x5646abd68300] ch:2 chl:stereo fmt:s16 r:44100Hz ->
 ch:2 chl:stereo fmt:dbl r:48000Hz
 [auto_resampler_1 @ 0x5646abd7d8c0] [SWR @ 0x5646abd7ddc0] Using dblp
 internally between filters
 [auto_resampler_1 @ 0x5646abd7d8c0] ch:2 chl:stereo fmt:dbl r:48000Hz ->
 ch:2 chl:stereo fmt:s16 r:48000Hz
 }}}

 I understand that resampling can change the '''true''' peak, but the
 internal change of sample format makes the reported '''sample''' peak
 information completely useless IMO.

 The reported sample peak is unrelated to the input sample scale, and in
 practice is identical to the true peak, because it is not clipped by the
 input bit depth.

--
Ticket URL: <https://trac.ffmpeg.org/ticket/7813#comment:5>
FFmpeg <https://ffmpeg.org>
FFmpeg issue tracker