Hi, I'm trying to make sense of how delread/delwrite work in a fft subpatch.

So, I have attached a test patch, why does it sound so bad? All I'm trying to do here is delay the sound input before it gets analyzed/ressynthesized. This is not my ultimate goal so I don't really need to do what I'm doing here, but I am exploring delays in spectral patches so I need to figure a few things out.

thanks

del-fft.pd.zip

@porres The signal isn't returned from the re-blocked delay line until the next cycle of DSP computation, one block later; so your signal would normally be delayed by one block......1.45125 msec by default at a block size of 64.
But with a re-blocked size of 2048 the minimum delay for [delread] will be 46.44ms (32 x 1.45), and then as you have upsampled x4 it will be 11.61ms before that is returned to the main patch.
At delay times of 11.61, 23.22, 46.44, 69.6, 92.8, 116 etc. the signal returned from the sub patch happens to align with the blocks in the main patch, and in between it doesn't because the delay is no longer a power of 2.
Not a great explanation, I know..... a simplistic view.
The truth is here...... https://puredata.info/docs/developer/PdMemoryModel/
David.