[FFmpeg-devel] af_silencedetect.c: silencedetect reporting format change to follow blackdetect reporting format

[Billy Shambrook]

On Sun, May 18, 2014 at 11:39:57PM +0100, Billy Shambrook wrote:
>> https://github.com/FFmpeg/FFmpeg/pull/68
>>
>> Reports silence using the same format as the vf blackdetect. This is to
>> simplify parsing the output of each filter.
>>

>This delays the silence_start printing. It can be an issue for real time.
>The main motivation behind this filter was originally to detect a silence
>issue in live stream. This is why it is important to print the
>silence_start ASAP.

>BTW, you're not supposed to parse that output, metadata injections are
>here for that (use ffprobe). But since some people might be doing this
>anyway, you should probably bump micro version in libavfilter/version.h.

Thanks for the info. I have updated my pull request with the version bumped.

I have read into metadata injections and looks extremely promising since I
use ffprobe as well to query my assets. I have seen that blackdetect
currently does use metadata injections, but will try to get a patch out for
that to share.

One question is that I have managed to get the two commands below to work
to utilise metadata injection (blackdetect doesn't yet), however, is there
a way to combine these into one command?

ffprobe -f lavfi -i movie=1080p24.mxf,b
​lackdetect​
-show_frames -of json
ffprobe -f lavfi -i amovie=1080p24.mxf,silencedetect -show_frames -of json

​Thanks​

-- 
*Billy Shambrook*