Frequently I want to apply different silence thresholds to the beginning and end. I often find music tracks with a trailing hiss. If I set the threshold high enough to detect it as silence, it's possible to unintentionally remove quieter elements from the beginning (like a fade-in). I would like to use different parameters for the beginning and end when trimming silence to avoid this issue.
Currently it's not possible to work around this by specifying split points because there's no reliable way to determine if and how much any of the beginning of the track was removed. If the new logfile were also written out when using -r, then we could use split points, but the logfile is only written out for -s and that doesn't work eith -r.
