There is true to what Squinky.Labs says, unfortunately
thats equally tactile but less convenient because you have to constantly remap, unless you only use a single fixed rack.
Without some form of electronic scribble strip, figuring what a knob does is indeed annoying. I’m thinking of something akin to the Mackie C4 but with high res encoders. For me it’s actually more complex, as I really want to control instruments and effects both inside and outside of Rack.
My thoughts are to send the data to intermediate software (in my case a self-written LuaJIT system called MIDImangler), and then route the interpreted data to the instruments. I realized this isn’t the same as one knob/one parameter, but that is quite impractical for large music systems, and even with real modular synth hardware, you’re limited by having only two hands and a maximum arm span. Deadmaus needs an assistant if he wanted to move more than two fairly nearby controls on his setup.
You can often turn two knobs or move multiple sliders with one hand. Footswitches can be useful too, as can modules that act as macro controllers and feed out to multiple controls.
All that is possible with real modules, or with a virtual modular and suitable controllers.
But the latter sadly still doesn’t feel like the former. There is the scribble strip problem but the real difference, for me at least, is that in the virtual world I always assign controllers after I have designed a patch and want to play it, whereas with the real modular I designed the patch in the first place using knobs, sliders, switches, capacitive touchplates etc. Playing and patching is seamless, and very much a jamming experience most of the time, not two separate activities (unless I want to make it so). It makes the patches that I design in the two environments very different. With VCV it is generally more of an intellectual activity (and I do make much more ‘interesting’ patches) and often they are primarily generative in nature; with hardware it is more visceral, and I am very focused on what I am hearing, though there remains the intellectual challenge of problem solving as you never have, or have enough of, what you need, and have to find another way of doing it. I don’t think ‘Oh, it would be fun to play this patch’ because playing it is naturally part of what the patch ‘is’.
4 Likes
@caowasteland , I know you are engineering your own device (which I think is awesome!) but if anyone wants something like this and hasn’t seen the endless-encoder-based Faderfox EC4, check it out:
Loopop has a good video on it (as for most other things).
With a few firmware upgrades (apparently in progress) it has some really neat interoperability possibilities for Rack. I’ve also heard a rumor that a larger one is in development, although obviously COVID is forcing a lot of companies to change or postpone plans.
@Nik, I thought that was really well said. I’ve never put it as clearly to myself, but some of my long-term hopes with module development involve ways to, in your terms, shrink the playing/patching dichotomy in Rack; keyboard or controller driven cable control being the first thing, but ultimately patterns like “touch a module and its controls auto-spill onto a control surface.” For me, it’s about avoiding (or at least postponing) the configuration step. “Let me think about how I want to set this up to play it” is the biggest creativity- and flow-breaker for me.
I think the right approach could get most of the control anyone would want with almost no configuration. Doing it really well would require VCV controller metadata per module, but some of that can be gotten through the API, and a little more of it can be squeezed out for open-source modules through static code analysis.
Anyway, a topic for another thread and another time (and after V2 is out). But there are some incredible possibilities, I think.
2 Likes
Other than its price 
that thing is super-cool (and a really nice technical solution to the electrical limitations)!
Unfortunately, it looks like it only runs with the DROID (though that’s pretty cool too) and I’m not sure whether it’s an i2c connection or proprietary. It may need to be run by something that has the capacitance limits in mind but something like this which did I/O over MIDI or CV would be very special. Like the 16n or Sweet Sixteen but with motors.
That said, it’s extremely inspiring and I hope some of the interface ideas on display in that video stay around!
um, may I politely request more information about what’s going on in these pictures as soon as possible please 
So I had similar dreams about building a 16’m’ device for controlling VCV. I realized the cheapest route maybe buying something off the shelf like the presonus faderport etc, but then you don’t get the hi-res. So I kept an eye for old things with motorized faders to pick up cheap and and hack.
Now for the more technical bits, I ended up getting this broadcast digital mixer ‘Lawo Sapphire’ (2 of them). The controller uses CANBUS protocol (like in your car) over ethernet cables to communicate with the mixer. I bought a little canbus device and reverse engineered most of the CAN communication. Removed the mixer and instead have a python program that can translate the can messages to midi/osc. Presto, 24 motoroized faders, 36 endless encoders and a crap ton of buttons and oled screens. I don’t have the screens fully figured out yet, and some of them are dull from age, but it’s still amazing, when you do something like in that video.
The faders send values from 0-649 which I was rescaling to 0-127 when using midi-cat and hence osc’elot was born.
4 Likes
That is AWESOME.
1 Like
Incredible !!!
2 Likes
The main issue is that most endless encoder implementations have mechanical encoders, and those have poor angular resolution: 24-ish jumps per rotation. Using an as5601, you can resolve up to 4096 changes per rotation. Also, most controllers are going to send MIDI 1.0 which restricts you to 128 levels if you stick with what’s well specified. I recall a discussion here in which Bruce and I participated not too far back about how poorly 14bit CCs are specified. Bruce provided interesting historical notes about the standardization process. Bruce is hereby dubbed VCV hardware history/color commentator. 
1 Like
That thing looks more than a little expensive.
Yes very, it’s used in TV stations and such. I picked up an older model, the mixer is a 19" rack unit, with only the master board and no other cards. That and the fact that it was old saved me a bunch. I hoped I would be able to use at least the faders and worked out that each fader would cost me less than $35. (Still expensive). Luckily I was able to decode it enough to use it without extracting the faders.
If someone wants to try something crazy like this, this was something I was considering before getting what I have now:
You should do enough research and assume no one will be able to help you before you spend that dough though.
I’m staying far from motor faders as electro-mechanical stuff is more complexity to go wrong. With an Hall effect encoder, there’s no physical contact to muck up, and the meaning of a position can be anything the software decides it is. Just do all your +/- deltas mod 4096.
1 Like
Yes I’m very interested in whatever you come up with, I would love to have something like a diy monome arc
1 Like
Agreed!
I remember it well! I still think, as I maintained in that discussion, that 14-bit MIDI is in practice adequately specified for professional use, but I am getting more interested in other protocols these days and will be very interested to see what you do with the as5601 in any case.
You may have seen the Yaeltex version 2 custom controllers? They’re still in beta but extremely well made. The platform can support a LOT more than the current suite of controller types and the firmware is extremely well written.
I mention it because their flagship custom controller is a push-button detented endless encoder with a NeoPixel 16 ring around it, configured with the bottom three “fused” and 13 steps. It’s not as cool as the Arc’s full ring but it’s more practical and you get independent color control. TONS of feedback options and a very responsive dev team.
While I’m thinking about just a smooth grease damped shaft, using a neopixel ring is an intriguing prospect.
Incidentally, the as5601 detects gap change, so push-buttony things are possible.