Crowdsourced mixer design

[TomWhitwell]

I do feel there’s something interesting in a multi-in multi-out device controllable from Norns, without any interface. I have (but never built) one of the 6-in, 8-out audio boards for Teensy. The idea of a routable, mixable, effect-able thing that could be a matrix mixer or a traditional mixer with lots of sends, or six effects boxes or a multichannel panner seems really interesting, with a lovely library like softcut to make control radically easy and nice-sounding. I have no idea how much processor power you’d need to manage 48 nodes, let alone to start adding 48 compressors or 8 channels of reverb or whatever. I suppose it’s exactly the kind of hardware that is in my Motu Ultralite, but I’d never use it in that way. That tension feels quite Monome-ish to me.

Edit: details of that 6in 8out codec board: https://oshpark.com/shared_projects/2Yj6rFaW

[Lervais]

If I owned this I would love it if the span knob could go beyond an incremental nearness-esque spread to a state where the middle channel is mono, and all others are panned hard left and right respectively, depending on which side they are on. In order to mix mono and stereo sources in a traditional way when needed.

Maybe a different way of achieving this could be a knob attenuating the spread curve. Similar to quality & span on 3Sisters , if that makes sense? Something like this:

[zoundsabar]

Jumping into the fray, here’s a design trying to be minimal and flexible in a modular environment. Eliminating bells and whistles and concentrating on mixing and routing.

Like the latest @karst design, it has direct, post fader outs per channel, which act as sends, and allow for stem recording of the tracks.

Each channel would have:

• Active direct post fader (pre-panner) send out on each channel. (No send level knob needed. Mixing multiple sent signals would be done outside this minimal mixer. A pre-fader send would be done by multing the input.)
• L/R panner.
• A or B bus switch that can be toggled by swiping across the row of buttons. (If using the A bus, toggling to B acts as a mute, but doubles as a submixer, similar to some Mackie mixers “Alt” bus.)
• Fader (could be a knob)
• Signal present / OL LED
  (no EQ or solo)

Bus masters A and B would have:

• Fader (could be a knob)
• Mute switch
• A cascades into B unless there are plugs in the A outs

Master out would have:

• Stereo meters
• Headphone out

All ins and outs would be 1/8" jacks for eurorack (could be 1/4", in a larger standalone version).

The idea is to make the mixer as simple and as flexible as possible. Anything that can be done outside of the routing and combining functions is not included. So no EQ or drive. If there were room it would be nice to add solos.

[jlmitch5]

If you did transformers on the output a la rip, your extra knob could be a dc offset added to the signal which will get rip to fuzzy territories. If you did that you’d probably want the ground lift switch too.

[bobbcorr]

Sadly, it was my understanding that the Poltergeist is no longer being made? Or is it simply in a production hiatus?

Also +100 for your wonderfully informative posts. It’s a real joy reading experts discussing their craft in Understandable Language.

[addamm]

My post-poll vote might be to just work off @karst’s design at this point I’d love to see a feedback-friendly device and wonder if Bastl’s Dark Matter module might spark an idea.

If we have a knob to change between modular and line level it might be necessary to include at least a single overload indicator led per channel. That said, I wonder if instead of an overall drive knob (which I am down with) if the individual channels could just drive in a pleasing (ha) way when pushed. For instance, if that line/modular level knob could be used to drive that individual channel (this might be very tricky to implement and fiddly to manage). I’d love to see it go down (up?) to instrument level too

In that blank knob space I wonder if it might also be possible to spread things across L and R in a comb fashion similar to the serge reseq. So odd inputs would pan left and evens Right to some degree. Admittedly this might be a terrible idea. Maybe this could be separate LR outs altogether. Something something phase inversion switch…

[zoundsabar]

Thanks! You are too kind.

Koma doesn’t have any Poltergeists in stock, but I don’t think they have officially been discontinued. They have closed their Berlin space. I wonder what other changes they are going through these days.

[Jonathan_Riley]

I scoured the internet and sent an email asking about the poltergeist going back into production, still haven’t heard back. Heard from some folks who work at a synth shop a couple of months ago that there aren’t any plans to restock…

[Gahlord]

Knob to attenuate an assignable cv maybe?

[renegog]

@addamm

if the individual channels could just drive in a pleasing (ha) way when pushed. For instance, if that line/modular level knob could be used to drive that individual channel (this might be very tricky to implement and fiddly to manage)

the bastl dude does this really well, handling levels from euro down to whatever knicknacks (not enough gain for instruments), and also has a nice driven character to it when pushed. not sure how they did this but it should be open source for people that know how to read those things?

[karst]

spent a little time looking into potential targets for size. I like @TomWhitwell’s suggestion to make stuff compatible with hammond enclosures. the sandwich stack would be totally valid (and probably the easiest way to make something low-profile), but those who wanted a real case could just buy a ready-made aluminum enclosure. putting all the jacks on the top means you can drop it into a case without having to modify anything to get access to the full i/o (power excepted, but that’s a later conversation I think).

there’s two sizes that seem to work. the 1550G is 222x146x51 (~8.7”x 5.8”x 2”), so roughly the footprint of an 0-Coast (but would sit higher off a desk). 1550J is 275x175x63 (~10.8”x 6.9”x 2.5”), so a bit bigger than a grid (and, again, much higher profile). the smaller case is a bit smaller than my previous sketch, the bigger one decently larger.

any thoughts on desired size?

as an experiment, I tried to squeeze just the features from the previous sketch into that 222x146 footprint while leaving small border around the edge (5-7mm) to be able to fit the pcb. unsure if that border is sufficient, but the short answer is that I think it works, but just barely. any additional physical affordances beyond maybe a tiny switch (think RIP ground lift switch) would be impossible at this size.

size comp to last sketch

definitely less luxurious, but possibly livable

I’m really squeezing stuff in. took cues from the minimum distances I can find between jacks and knobs on the Mannequins modules, so at this size it would be crowded, but I think could still be decently ergonomic in use.

Couldn’t find a good spot for both sizes of output jacks, but I’m sure something could be figured out. How would people feel about only 1/4” (or, alternatively only 3.5mm) for the main outs?

I love the suggestions around using transformers for the main outs like RIP and either having a dedicated DC offset knob to get that fuzz (or maybe you could just patch a signal with offset into one of the channels), but adding transformers to this would increase the cost, weight, and height of this fairly significantly. maybe there’s a way to make that kind of addition optional to the user?

[jlmitch5]

Really love seeing your process on the design.

Re: transformers. Agreed that they would be a significant cost add and are big and heavy. Maybe there’s some easy way to put pins on the back of the pcb main outs so that they normal to the jacks, but could interface with some like @sonoCircuit’s EXI[S]T transformer boards for the 1u adaptation of rip if users wanted to go that route.

Thinking more about it…cv control over “span” mentioned above seems like such a cool idea (and I’d say keeping with the idea that what makes this design special is that it’s based on the nearness concept), especially for rhythmic/metered music, where you could patch up an in time control signal to snap the width out (envelope) or make the width undulate (lfo)

I did start to think (what If those gain knobs were push button to control mute) but then thinking through the ux, it’d be super annoying if you went to mute/unmute and you get these volume jumps or clipping coz the control moves a little bit, so scratch that idea.

EDIT one thing to think about with the span concept is the pan law. I think you’d want it so that at fully mono things are a little quieter in order to not have a volume jump, but that might make whatever’s in the center decrease in volume too much. Maybe each channel pair fanning out from the center has a different pan law, to keep things feeling contiguous? Interesting puzzle there

[_mark]

@karst

hammond enclosures.

Amazing that you can even get walnut side panels -

[karst]

I’m sold on cv over span. I said there’s not much room for additional features, but I can make some room near the span knob for a jack. there’s room to sneak most of the master controls a bit north. if we wind up needing an additional control over the stereo field (@Lervais’ ‘curve’ param) then it might get messy, but one thing at a time.

I’m not sure if the best solution for gain is the knob or some kind of switch (like the instrument/line switch on the land mixer), but my instinct is that continuous control over gain is going to be better for internal feedback patching within the mixer.

just a heads up: if people are into this idea and want to keep pushing it forward, we’re rapidly approaching the point where it’ll take people who actually understand circuit design to make any meaningful progress. I’m having a lot of fun mocking stuff up and will help in anyway I am able, but I think we’re still looking for the person or people who are interested in tackling the engineering side of this thing.

[jlmitch5]

I’d be happy to try building prototypes/testing etc. and continuing to think through some of the design ideas. This seems like a really cool idea and I’d love to help it come to fruition. But the circuit/pcb design is above my electronics skill set.

[sonoCircuit]

Like how your designs are taking shape. As for ergonomics, I’d place the gain pots above the faders. This gives better access to the faders and accidental gain changes are less likely. But that’s just my personal taste.

@karst

How would people feel about only 1/4” (or, alternatively only 3.5mm) for the main outs?

If you wish to have a low profile faceplate/PCB sandwich design like in your sketch you’ll have to ditch the 1/4" Jacks in favour of 3.5mm. Their hight is substantially more than 3.5mm jacks and you wouldn’t be able to mount them on the same PCB and have the mounting nuts be at the same hight as the other jacks. One could have a PCB cutout and the 1/4" jacks mounted separately but that would increase the profile.

Here’s a picture of MN Rosie. They did this for their 1/4" jack.

[karst]

@sonoCircuit

I’d place the gain pots above the faders.

I liked the visual separation between the rows of knobs on the larger design, but I agree with you on the compact design.

I’m now thinking the gain knobs sit closest to the i/o jacks (this is maybe also helpful schematically??).

my reference for the inclusion of the 1/4” jacks is the RIP design (with the 90° small pcb connector).

they extend ~24mm behind the front panel, which wouldn’t make the sandwich version unspeakably large (clearly including transformers would be a far bigger burden). I figure removing of the upper right-hand corner of the pcb would be doable to include them. That said, I’m very open to varying opinions on output jacks, maybe this is something that could be determined by each builder.

side note, if supporting the full gain range from instrument to line to eurorack proves impossible, I wonder if it would be somewhat easy for each builder to choose if they want a instr/line or line/euro range.

[sonoCircuit]

@karst

I’m now thinking the gain knobs sit closest to the i/o jacks (this is maybe also helpful schematically??).

Yes, that would be helpful in respect to PCB design.

@karst

my reference for the inclusion of the 1/4” jacks is the RIP design

Ah ok go it. I somehow assumed the goal was to have a shallow PCB sandwich like the 16n.

@karst

supporting the full gain range from instrument to line to eurorack

AFAIK a high gain circuit would be needed for this and this introduces a lot of components if you want a clean signal. I’d opt for a switch to choose between two circuits, depending on what is plugged into the input. Or like you said: let the builder choose what suits them best.

[karst]

@sonoCircuit

Ah ok go it. I somehow assumed the goal was to have a shallow PCB sandwich like the 16n.

I think the goal is still up in the air. My working assumption is that having 1/4” jacks for the master outs is worth the height increase, but this is not a deeply held belief.

[jlmitch5]

Seems like the you could protoype with 3.5mm might be helpful and then migrate just that part later on in the process (but maybe that sort of iterative design makes more sense in software than hardware)