One shameful question…
I do not understand why the output data rate does not change when changing fftsettings.
Shouldn’t it be the case ?
Because the stuff has to bounce from the audio thread back down to the scheduler, there’s not really any way of doing it for all the various relationships between the signal vector size and the hop size that are possible. If the hop is < sigvs, then we can’t keep up, and if it’s some weird ratio, there will be vectors where not enough stuff happens. So, to keep it simple, the output rate is once per signal vector.
OK. That might explain why you often have a “zl change” after an analysis object.
Thanks !
1 Like