Sorry to go off-topic here: I wonder if there is also a "musical" approach for handling linearity or thermal drift.@guest
I applaud your path and goal. Image may be NSFW.
Clik here to view.
With that said Image may be NSFW.
Clik here to view.
I'm of the belief that as digital gets faster, smaller and cheaper we will be headed more and more towards 'modern' versions of 'autotune' instead of tempcos and that CV will instead be processed servo like. There are already many examples of this approach.
I enjoy the You Tube channel https://www.youtube.com/wingsofpegasus which has recently been doing various exposes of the excesses of pitch correction on singers, featuring realtime pitch detection animations. I have found the different vibrator strategies very interesting: it seems that many singers have a downward vibrato near the top of their range but an upward one at the bottom. And on those extreme notes, there is often less or no delay before the onset of vibratro. They are not so much using vibrato for affect there, but for pitch fudging: the human ear figures out what pitch it is supposed to be not based on some averaging of the pitch but based on what pitch makes musical sense in context. We know that various real instruments don't have centred vibrato too (guitars bend mainly up, unless there is a whammy bar; single reed instruments bend mainly down, etc.)
Sung vibrato can often be 100 to 200 cents wide, or more for operatic singing.
So I wonder whether a simpler(?) alternative (in some applications) to ideal linearity (like digital) or ideal compensation (like dial-a-tempco) involving complex circuits or highly accurate/expensive components might be to automatically add the slightest vibrato (instead of or as well as compensating.)
For example, if an expo converter of some design is known to go up to 12 cents flat at the high end when hot, add a rough detector circuit (thermistor and diodes?) which on high notes adds in just enough positive vibrato (e.g. 0 to 13 cents) that the pitch oscillates between the flat note and to roughly the linear pitch.
I guess this comes back to what the purpose of linearity is:
- so that two unsynched oscillators track perfectly without phasing, that is a hard problem (which might be better handled by some phase-locked loop circuit rather than perfect linearity?); or
- so that notes stay in tune regardless of temperature, then good expo converter or servo is probably good; or
- so that an oscillator is musically useful over a wide range of notes and temperatures, then perhaps this low-precision automatic vibrato idea might be simpler and more pleasing, at least for patches where slight vibrato is not disruptive. Perhaps noise could be used too.
Statistics: Posted by ricko — Sat Oct 12, 2024 11:37 pm
