thankyou thankyou thankyou !
If the attenuator is a VCA, you can hook up an envelope detector to the filter output, invert the output and use it to control the VCA gain. A simple control loop that (with some tuning) holds the filter constantly at the brink of self-oscillation, with some interesting transient action, kind of like a hybrid filter/compressor. I used to do this with Serge before I sold my preamp/detector panels. This approach works in general for exploring other types of feedback, with delay lines, resonant EQ, and so on. it’s basically a way of extending the range of musically useful results while avoiding too much distorted chaos…
thanks for the kind words sir!
alright, fun things learned about re-calibrating 3sis for v/8:
- temperature-compensated really means what it sounds like. touching the pcb threw pitch off pretty hard, since my hands were cooler than the module. wild. took about 10 minutes to re-stabilize after that first touch.
- the fluctuations made it pretty hard to calibrate. the first time I nailed the 80 -> 320 check, 10 minutes later I was getting 80 -> 313.
- eventually, after each adjustment, I let the module sit out of the case for a few minutes to re-stabilize and just kept trying.
- after what felt like a million tiny adjustments, things evened out to 80 -> 320(ish)!
thinking over the helpful info Ezra provided about self-oscillating filters, I wonder if the Doepfer A-110–4 might fit this particular bill. it’s a VCO with a sine/cosine core, which seems like it’d be the best match for the sinewave generated from 3sis. of course, we lose the whole ‘filters as oscillators’ thing, but I think the FM pair between this and 3sis might get close to the same fun with more stability.
(this all feels insane, but also oddly necessary! thank you all for being on this journey.)
The octave trick that @Galapagoose mentioned for calibrating Mangrove might also help? It worked wonders for me. You sequence octaves rapidly while you’re calibrating and you can hear (or see w tuner) the pitch accuracy over the span. I attempted 3sis calibration once long ago and it was indeed tricky.
re: temperature calibration - this is a naive question for someone like @Galapagoose or others with better understanding of analog circuits than myself, but is there a reasonable way to retrofit temperature compensation?
my (incomplete) understanding is that it typically involves a special “tempco” resistor, right? I suppose I am curious if there is a possibility that some minimal rerouting/part replacing might be able to achieve temperature compensation.
only a noob reply, but my guess would be that this would be incredibly difficult on a post-factory Three Sisters, given the SMD construction
I suspect you are right…but on the bright side, SMD components are much easier than thru-hole to remove
I don’t think the takeaway of this thread is don’t get a 3 Sisters!!
for sure! I’d hope the takeaway is “get two”
the Doepfer suggestion is meant to explore options besides getting three, for the particular use case of creating complex voices through filter modulation (which 3Sis excels at) with/of clean sines rather than traditional frequency modulation.
super glad to learn that there is a legitimate (+ functional) difference between triangle-core sines and the sines of oscillating filters! that is honestly a huge relief.
Any recommendations for a decent envelope follower? I’ve been thinking about picking up an Ears in part because it’s got one…
Maths can do this (Cold Mac’s follow is a clone of the ARP 2600 envelope follower as well).
For Maths, patch your audio signal into the signal input on Maths, and adjust the rise/fall/curve knobs to taste, until the LED output shows you that you’re getting a smooth envelope rather than weirdness.
I think you’d find Nicola Ratti’s work fascinating. His album ‘Pressure Loss’ is 8 lfos and 8 filters. Don’t mean to divert this thread too much, but it related to pinging filters and he does a lot of this if you’re looking for more examples.
The main one I’ve had experience with is the Serge preamp/env. I still have very limited experience with Eurorack in general… and of course nothing at the moment.
Not sure Maths will be enough if it’s like the Serge DSG. I think you need to half-wave rectify or square the signal before putting it through (does Maths also have this?) then you’ll be able to do something with the rise/fall times to get a good envelope. But every time I tried patching something like this, I still got better results with the Serge preamp/env module directly.
ok - so I think the main takeway from the convo here is that I’m gonna have to do a bit of exploration before setting on anything (funny how this is usually the takeway). I’ve got my eyes on the doepfer a-11-4, but I think I’m gonna spend some more quality time with max and figure out some tones I like before making any big decisions. I’ll keep u folks up to date.
i’m really enjoying these convos, so ¯\(ツ)/¯
i’ll dig out my 0.5 version of less concepts (app) and share it here. i originally included two pinged sines before deciding it should speak midi, but keeping the sines in there would’ve made it impossible to distro through github (size). brb
adapts http://maxforlive.com/library/device/4672/low-pass-gate inside of less concepts, though y’all might find great use for the m4l device as well!
it does, if you fiddle with it: I guess the patch would become signal -> maths ch. 2/3 (attenuator) -> OR out (half-wave rectifier) -> ch. 1/4 (slew) -> corresponding channel output. I want to maintain that if the rise time is shorter than the fall time, you’ll end up with a positive-going envelope even without rectification, but I confess I haven’t actually tried it
How would you deploy the CM For this?
Well, to follow @ht73’s patch description to a T definitely will require some case-by-case thinking (or rack-by-rack, as it were), so let me instead translate their words into a patch schematic and leave the rest to you.
So. Let’s call our filter, delay line, what-have-you module and let’s work “backwards”. Mult the output of module twice (Stackcable? Mult module? you decide). Listen to one copy and patch another through a VCA (or LPG). Now, we need a negative-going envelope added to a positive offset so that if the envelope is at zero, the VCA will be open to taste. If your VCA has a bias control, we’re already halfway there. If not, you’ll need to do some mixing. So, let’s create an envelope from our final copy of the output. One way to do this is via Cold Mac, mentioned below, another (potentially), would be to use Maths as I mentioned above. Somewhere along the way, though, we need to invert the envelope and add the offset. Then we patch the envelope we ended up with to the CV control of the VCA. Likely we’ll want to add some attenuation to taste. In the end, we want to mix the output of the VCA with the input of module.
As far as Cold Mac, the main observation is surface-level: patching a signal into slope turns follow into an envelope follower; in this configuration I’m basically ignoring the rest of the module.
If we wanted to be annoyingly clever (and hey, finding annoyingly clever uses for Mannequins modules is something of a pastime for me) and weren’t inputting audio to anywhere else in Cold Mac, we actually could save ourselves one mult: the mac output of Cold Mac mixes all the (audio-rate) inputs and attenuates them according to the value of survey. In fewer words, mac is a 6-to-1 AC-coupled mixing VCA with survey as its CV control and the knob as a bias. So, we could just patch our inverted and attenuated envelope into survey (no need to worry about offset) and our mac would be what we would mix with the input to module.