Le dim. 3 févr. 2019 à 03:24, John Warburton <j...@johnwarburton.net> a écrit : > > On Sat, Feb 2, 2019 at 9:32 PM Romain Beauxis <romain.beau...@gmail.com> > wrote: > > > There shouldn't be any chance on that part of the code. I'm happy to > > investigate anything that looks wrong if you can share some of your > > script/a way to reproduce. > > Certainly. Happy to do so. > > The difficulty appears when using crossfades. In Liquidsoap 1.3.6, the > "liq_amplify" annotation works as expected: the track level is > increased by exactly the amount given, whether or not the playlist is > passed through the crossfade function. > > But in Liquidsoap 1.4.0-alpha1, the "liq_amplify" annotation, if > present, appears to result in much more gain than what is specified. > If it is absent, or set to 0, there is no change in gain. > > Here is my test playlist of three short tracks. Each is recorded > quietly, so each requires an amount of gain. These are three extreme > examples, but all are measured properly. > > annotate:liq_cue_in="0.000",liq_start_next="5.560",duration="312.060",liq_amplify="8.000dB":/home/john/src/radio/mez/204-zoot_sims-blue_room.PbTVdK.mka > annotate:liq_cue_in="0.800",liq_start_next="9.140",duration="286.840",liq_amplify="8.200dB":/home/john/src/radio/mez/02 > Cha Cha Loco.FKMSrM.mka > annotate:liq_cue_in="0.100",liq_start_next="5.280",duration="162.080",liq_amplify="8.100dB":/home/john/src/radio/mez/12.Song > to Woody.wOgkma.mka > > These two lines under the release version produce correct results: > > myplaylist = amplify(override="liq_amplify", 1.0, > cue_cut(playlist(length=60.0, mode="normal", > "/home/john/src/radio/cha.m3u8"))) > myplaylist = crossfade(fade_out=0.01, fade_in=0.01, conservative=true, > myplaylist) > > But under 1.4.0-alpha1, the addition of the second line produces way > too much gain. Under 1.3.6, it produces the correct amount of gain. > Is, perhaps, liq_amplifiy being applied twice, somewhere, in the > crossfade process in addition to the amplify process? > > If you'd like to see full files, etc., please let me know. That will > be no problem.
Thanks for these details. I'm gonna try to reproduce. Meanwhile, do you see logs of the form: Overriding amplification: 2.5674 This would allow to track down the amplification levels in both version.. Romain _______________________________________________ Savonet-users mailing list Savonet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/savonet-users
