Incoherence between recording times and sorting/analyzer times

[JulienBrn]

Recording segments seem to be aware of non zero start times. However, when in an analyzer or sorting, times always assume the recording starts at time 0. This can be checked in the GUI for example.

I believe this does not create any main issue as the analyzer seems consistent with itself. However, this is very unintuitive and maybe not expected ?

[h-mayorquin]

Probably a bug, have you used the generate_recording and generate_sorting functionality? If you can create a minimal example this will go a long way for us to move forward quicker with the fix.

[alejoe91]

@JulienBrn I think this issue is actually on the GUI side. Could you open an issue there?

Any other instances where so see a timestamp mismatch in the SpikeInterface API?

[samuelgarcia]

THe GUI do not handle t_start at the moment.