Recording notes

Tried an experiment 2023-09-26 using headsets with laptop at the BL. Bryan's Sennheiser USB headset is noticeably worse than the Yeti, and shouldn't be used. Possibly using the portable Zoom H4n as an audio device (with normal headphones) could fix this. The gaming headset is fine but possibly not as good as the Yeti.

To mix the two files, panning 60/40 with quality = 5.

FILE1=1.wav
FILE2=2.wav
QUALITY=5
LEFT=0.6
RIGHT=0.4
ffmpeg -i $FILE1 -i $FILE2 -filter_complex "[0:a]amerge=inputs=2,pan=stereo|FL<$LEFT*c0+$RIGHT*c1|FR<$RIGHT*c0+$LEFT*c1[a]" -map "[a]" -codec:a libmp3lame -qscale:a $QUALITY output.mp3

• Use Chrome on both sides and better interfaces
• Try Zoom H4n as an audio interface

Another idea from here. Basic idea is to duplicate the mono track into two channels, offset one by a slight time difference, then run different filters. (He does a high-pass on one channel and a low-pass on the other, then different reverbs on each channel.) Now one option is to do this manually, another would be to run an automatic "pseudo-stereo" plugin.

But in the case of Squadcast, what we have is two mono tracks, each containing one person's voice. The above ffmpeg command just pans them so that they're still simultaneous, but they are each panned a bit to one side. From here, I could try offsets from 10 microseconds to 700 microseconds. In other words, duplicate each track, then attenuate one of each and offset it by a similar duration, then mix it all down.

I got good effects with adding the minimum: 1 or 2 milliseconds delay, and then adding a tiny bit of reverb to the quieter track.