I'll take goofy as a substitute for silly. The original goofy point is you could make an approximate pitchshifter with nearly all analog parts. No, it's not a good idea, it's like outsider art for signal processing. But imagining it has always tickled me a bit. I guess the humor of it didn't land here.... oh well.
-Russ On Fri, Dec 8, 2023 at 11:17 AM robert bristow-johnson < [email protected]> wrote: > > > > On 12/07/2023 11:45 AM EST Russell Wedelich <[email protected]> wrote: > > > > > > No splicing required: > https://www.semanticscholar.org/paper/Low-latency-audio-pitch-shifting-in-the-time-domain-Juillerat-Schubiger-Banz/9fa827c688688157cc93ec9b296f24a45c6dfdf6 > > > > It's all IIRs, HIlberts, oscillators, e.g. a large bank of narrow band > frequency shifters. It could be done with all analog right? Key word I used > is "silly" b/c that's what it would be. > > > > Hi Russ, > > There are so many goofy things about this. In 1992, before I did my first > pitch-shifter for the DSP4000, I did a Hilbert-transformer-based frequency > shifter. Gil was so excited about our new "glitch-free pitch shifter" and > we had to remind him "*frequency* shifter" and the response was "what's the > difference?" Richard, being a ham operator, was sorta into it. > > Ken and I discussed this multiband frequency shifter but this thing leaves > holes in the output spectrum if you're upshifting and, for downshifting, it > either has holes in the input spectrum or there is overlap in the output > spectrum. If there is overlap, then the same sinusoid gets shifted by two > different frequency-shifting bands and will get shifted to different > frequencies. > > I just don't think it's a good idea. > > For me, the purpose of a frequency-shifter is to slightly detune the > harmonic partials to slightly non-harmonic frequencies to sorta fatten up > the sound. It might help for chorusing. Or just for another, sorta > surrealistic spacey effect that sounds like Single Sideband (SSB) that's > not perfectly tuned. > > I just think, for pitch shifting, that it's important that the frequencies > of every sinusoid are *multiplied* by exactly the same ratio. And then > detune the harmonic partials afterward, if you want that effect. > > - robert > > > > > On Tue, Dec 5, 2023 at 9:22 PM robert bristow-johnson < > [email protected]> wrote: > > > Still I wonder how you would make good splices. Even with lotsa > overlapping grains, some are gonna be way outa phase and makes a glitch. > > > > > > It just seems to me that some kinda analysis needs to be done to > decide where to splice to and when to start. > > > > > > I knew about the spinning heads on tape but never saw any of them. > > > > > > Powered by Cricket Wireless > > > > > > ------ Original message------ > > > From:Russell Wedelich > > > Date:Tue, Dec 5, 2023 8:15 PM > > > To:[email protected]; > > > Cc: > > > Subject:Re: MUSIC-DSP Digest - 3 Dec 2023 to 4 Dec 2023 (#2023-67) > > > > > > “(Like in pitch shifting. But I'd like to see you do decent pitch > shifting with a purely analog device.) “ > > > I once had a silly idea for an all analog (well digitally controlled) > Rollers type Pitchshifter, frequency shifting approximating pitch shifting > in many bands. If would be sooooo expensive! So many opamps. > > > Long time reader, first time replied (maybe). > > > -Russ > > > > On Dec 5, 2023, at 3:23 PM, robert bristow-johnson wrote: > > > > > > > > (Like in pitch shifting. But I'd like to see you do decent pitch > shifting with a purely analog device.) > > > > > -- > > r b-j . _ . _ . _ . _ [email protected] > > "Imagination is more important than knowledge." > > . > . > . >
