Export to audio does not respect Synthesizer settings unless saved to score
Steps to reproduce; try to make the zita1 effect as dry as possible and then render.
1. View -> Synthesizer
2. Under Master effects you should see the default Zita1. Set Delay to 20 and Output to dry. You should now hear that there are little or no reverb when playing in Musescore.
3. Export to a sound format, I used wav. There shouldn't be any reverb?
Win8.1, 64bit. Musescore 2.0.2
Comments
Just tried to turn off all effects to only export the dry sound, the mp3 had reverb in it no doubt, and way to much which is the problem.
Oh, when you click "save to score" it takes it into account. Maybe that is the case for the rest of the settings too?
Yes, it is. That's just the way it works for now, but I'm leaving this active as a reminder that it would be nice if that wasn't necessary.
I experienced this too with musescore 2.0.2, I loaded my song in mscz format, then exported to audio (wav or mp3). It didn't use my mixer settings, it only used (it seems) the first soundfont in the synth list. When playing in musescore, the mixing was correct. Then I clicked on "save to score" in the synthesizer, and after the audio export was correct this time.
I'm pretty sure this is by design, although I'm not at all sure why it was designed this way.
See this?: #21592: Exported audio doesn't play loaded soundfont
Fixed in branch 2.1, commit 09df048818
fix #70791: Export to audio does not respect Synthesizer settings unless saved to score
Fixed in branch master, commit cdb3a48fe4
fix #70791: Export to audio does not respect Synthesizer settings unless saved to score
Does this fix save the synthesizer volume setting as well?
@geetar, I was just about to complain about export audio not obeying the synthesizer's volume, however upon inspection of the MuseScore::saveAudio() code, I discovered that musescore is automatically normalizing the audio. It does two passes through the audio, and in the first pass it figures out what is the song's peak amplitude: https://github.com/musescore/MuseScore/blob/09df048818e934a82522374c1ad…
Then after the first pass completes, it calculates the gain to nomalize according to that peak amplitude: https://github.com/musescore/MuseScore/blob/09df048818e934a82522374c1ad…
And then in the second pass it normalizes the audio according to that calculated gain: https://github.com/musescore/MuseScore/blob/09df048818e934a82522374c1ad…
So I guess MuseScore doesn't obey the synthesizer volume "by design".
I've updated the manual to explicitly state that the audio is normalized.
Automatically closed -- issue fixed for 2 weeks with no activity.