<d1b2>
<Mughees> Hi @azonenberg, Was looking at the peak-peak measurement filter. Any reason why we did not use 1st derivative to detect peaks? It uses an averaged mid-point to find cycles and then finds maximum and minimum peaks in each cycle. I think this will miss peaks in signals that lie below the midpoint.
<azonenberg>
@mughees I'm all for improving it. Note that we now have two outputs in the filter, a vector with cycle-by-cycle output and a scalar for the global peak to peak of the entire waveform
<azonenberg>
the FFT peak detection can probably also be improved
<d1b2>
<Mughees> Got it. Thanks
<azonenberg>
@mughees in general we are moving towards this for future filters (and gradually retrofitting into existing ones)
<azonenberg>
having both vector and scalar outputs where it makes sense
bvernoux has quit [Read error: Connection reset by peer]
<azonenberg>
And then down the road as an optimization, i want to have filters only calculate outputs that are being used
<azonenberg>
there is currently no way for a filter to ask "is anybody using this output"?
<azonenberg>
so to support that we have to provide some way to back-propagate that connectivity information