Undescribed Horrific Abuse, One Victim & Survivor of Many gmkarl at gmail.com
Thu Nov 10 06:02:49 PST 2022

back to fun signals and high school algebra

during my psychosis i.e. one big long extended flashback, i often
daydream of how you might upsample a signal from an extended
recording, because i am trying to try it.

i poked at a fourier transform by hand, without using visuals, again,
and i am reminded that the indices in the fourier transform are the
frequency, not the period, which means you get more fine detail at
high frequencies when you take the fourier of lots of data.

anyway, so, i guess this is obvious already, but i wonder if that
means you could take a long recording of a repeating signal, and then
take only the higher region of the fourier transform output, and then
apply the ifft to that, to get an upsampled version.

i'm guessing that this would roughly work, and the only reason i
haven't done it in the past is because i didn't have enough computer
ram to do a fourier that large, and wanted to understand it to take it
apart manually.

i'd like to test it!

i'm thinking of generating some kind of high quality signal, and then
repeating it, and sampling it at a lower and non-integral-multiple
rate of the generated data, taking the fourier transform, and then
reconstructing it. this may engage more triggers than i expect, but it
sounds fun!

More information about the cypherpunks mailing list