These are the latest version of the patches I already sent to the list. There should be nothing controversial in them, as functionality has not been changed. The correctness of the first one is not trivial to see from the patch. So I propose that the reviewer just trusts me when I say that when looking at sample values, the patch fixes the peaks resampler. If you want to avoid thinking the whole algorithm through, just check whether the pavucontrol meters still make sense. Maarten
