[FFmpeg-user] Measurement differences between loudnorm and ebur128 filter

[Paul B Mahol]

On Mon, Feb 14, 2022 at 1:50 PM Remko Tronçon <remko at el-tramo.be> wrote:

> Hi,
>
> When I run an audio file through the 'loudnorm' filter, I get the
> following:
>     > ffmpeg -i <file> -af loudnorm=print_format=json -f null -
>     ...
>     "input_i" : "-13.87",
>     "input_tp" : "0.06",
>     "input_lra" : "8.50",
>     "input_thresh" : "-24.45",
>     ...
>
> When I run the same audio file through the ebur128 filter, I get the
> following:
>
>     > ffmpeg -i <file>  -af ebur128 -f null -
>     Integrated loudness:
>        I:         -13.8 LUFS
>        Threshold: -24.4 LUFS
>     Loudness range:
>        LRA:         8.7 LU
>        Threshold: -34.3 LUFS
>        LRA low:   -20.1 LUFS
>        LRA high:  -11.4 LUFS
>
> Why are these measurements different? Even if rounding would be towards 0,
> that would not explain the LRA difference?
>

Run ebur128=peak=true, than it will use same sample rate as loudnorm
filter, 192000

Also do note that ebur128 output is rounded to 1 decimal and loudnorm to 2
decimals
The metadata output of ebur128 is rounded to 3 decimals.


>
> thanks,
> Remko
> _______________________________________________
> ffmpeg-user mailing list
> ffmpeg-user at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-user
>
> To unsubscribe, visit link above, or email
> ffmpeg-user-request at ffmpeg.org with subject "unsubscribe".
>