[sdiy] More on Flanging...

[Gene Stopp]

Hi DIY,

I finally set up the tape decks for the audio portion of my search for the
perfect Flanger effect (as opposed to the motor speed variation portion).
Being that it is extremely hard on the dermis of my thumb to slow down the
capstan itself manually (along with the danger of spreading skin oil all
over the tape path), I pulled the back cover off of the deck so that I could
slow down the motor fan instead. The results are fantastic.

True tape flanging sounds crisper, deeper, and more intense than anything
I've tried yet. It is what I've been looking for. It's better than the
effects found inside any of my DSP boxes (Midiverb 4, Eventide Instant
Flanger, SPX-90, Effectron II, etc.). I think that this goes a long way to
validate my theory that the two audio paths need a combination of full
fidelity, matched fidelity, and total lack of aliasing artifacts or whatever
to get the real deal. Sad but true - it's like the difference between a
Hammond organ clone sound and a real Hammond. For the time being I'm stuck
with big heavy clunky things with motors in them :)

Some observations - flanging is not possible without minor pitch shifting. I
had this concept that ideal flanging is comparing a signal to a delayed but
not pitch shifted copy of itself, so that the comb filter would be generated
by harmonic reinforcement and cancellation. This is true for a static flange
sound, but a sweeping flange sound must include an element of pitch shift
which may distort the comb somewhat since the harmonics are pitch shifted
relative to themselves! But obviously any time you store something and
retrieve it later, and there is a difference in the storate rate versus the
retrieval rate, there shall be pitch shift. In the DSP realm you can stuff
or drop samples to keep pitch constant, but not with analog tape.

Also, there is a latency involved with real-time tape flanging, 145
milliseconds in this case. OK for sustained sounds, but bad for tempo-based
stuff like drums. Doing it out of real-time is fine for producing a
recording, but I want it live. It's just a personal thing.

Next, I have realized that there is no possibility of adding resonance or Q
with this setup. The only thing that would happen if you try is a bunch of
echos. However resonance is usually added to emphasize the effect, and
tape-based flanging is so profound that it really doesn't need emphasizing.

As a result of my AC inverter experiments I think I am moving in the
direction of building a voltage-controlled small inverter and tapping in to
the capstan motor only on the tape deck. Powering the whole machine off of
the inverter would require too big of an inverter, much more than I need.
I'll probably try to build an inverter using a standard-sized filament
transformer driven by a pair of HEXFETs. See
www.chipcenter.com/circuitcellar/askus/mar00/Q1_00_43b.htm


Best Regards,

- Gene