# Can Frames function as a vector mixer?

Looking at Frames, it has four inputs and a mix output… that gives me a thought:

Would it be possible to use Frames as a vector mixer, where the mixed signal remains more or less at a constant perceived amplitude compared to each individual input signal, and where panning can be done across two axes?

I guess that with the channel response curves set to exponential and not linear, and with some careful amplitude settings for each input on every keyframe, it should at least be possible to manually approximate a typical vector mixer swirl pattern that we know from synths like the Prophet VS and Wavestation. Correct? But I fear that it could be a bit complicated to set up, and maybe it would be a better idea to save some more cash for a Planar…

Do you mean that you have 4 separate signal sources, that notionally lie at the corners of a square. Each signal is fed into a Frames input. As the Frame knob.CV is advanced, the four signals are mixed down to one signal in a manner that simulates what someone following an arbitrary path across that square plane? So the path could be a circle, and advancing the Frames knob would progress the listener around that circle? With three inputs, the path would be a triangle, and with two inputs, just a line, which is just stereo panning. Of course, other more complex paths would be possible. Is that the idea? If so, that could be done in the Frames software. Is it a useful technique? Can you point to any handy examples of what you are thinking of on YouTube or SoundCloud?

Yes, that’s what I mean. It’s a really great technique for making good drones. The magic usually happens with modulations somewhere in the center.

An example can be heard in this Planar demonstration: (audio example starts at 1m40s)

https://www.youtube.com/watch?v=y25BwF0ikUs

I’ve had some more time to think about this. I believe that in a vector mixing context, Frames works best as a CV source for modulating separate crossfading modules (or Planar), and is not that suited to do the actual audio mixing, unless you can accept intermittent audio level drops and boosts due to a lack of crossfading logic between the four Frames channels.

Yes, but Frames has a digital brain, and thus cross-fading logic could be added. There is plenty of space for additional code in Frames, unlike Braids or Tides where the firmware storage is essentially full. The main limitation of Frames is that it only has one CV input.

OK, let’s suppose that such a rewrite of the firmware would be possible. A vector mixer with keyframed and animated X and Y positions would only require two parameters, not four. What stuff could the two last parameters be used for… maybe control overall gain, or some sort of cross modulation between channels? VCA matrix style.

Yes, but there’s only one CV input, so X could be controlled by the frame knob/ CV input, but Y would only be controlled by one of the 4 knobs. What I had in mind was a set of predetermined paths to be followed across the plane, with the frame input advancing the position along that path. One path could be the perimeter of the square, another could be a circle, but you could have star patterns, Lissajous figures etc. Plus the four knobs could set other parameters, although not dure what. Of course, both schemes could be supported - there’s heaps of room for additional code.

Attached is a patch illustrating the Frames functionality I’m talking about: The four outputs from Tides are multed into 2x3 crossfaders, that are being CVed by Frames. The first set of crossfaders (controlled by Frames channel 1 and 2) form a vector mixer where the final mix is sent to the audio input of a VCA. The second set of crossfaders (controlled by Frames channel 3 and 4) are similarly mixed and sent to the CV input of that same final VCA.

If this functionality were in the Frames firmware, one could simply send the four outputs from Tides to each channel input in Frames, and then get the result on the Frames Mix output. No 2x3 crossfaders + final VCA would be necessary.

(Each group of three cross faders could be replaced with a Planar)

The method described in my post above, leaves the four individual channel outputs unused So what can be done about that? That’s when I was hoping one could have each individual output be the individual channel modulated by the blend of the others. However, this is when the patch complexity grows dramatically! A bit too ambitious, maybe? Or is this a typical example of things that don’t require a lot of resources in software, but are completely unrealistic in hardware?

To give a simplified example, I’ve attached a patch illustrating the output of one such individual Frames channel. Tides output 4 is routed directly to the VCA input, instead of into the crossfader. Tides output 3 goes directly into the last crossfader in the vector mixer, where it is crossfaded together with the crossfaded mix of Tides outputs 1 and 2. This blend of the three “non-4” outputs is then sent to CV in the VCA.

EDIT: I now realize that my example is more complicated than this; The final cross-modulation would depend on what values that are present for Frames params 3 and 4, that I didn’t include in the patch example.

> Or is this a typical example of things that don’t require a lot of resources in software, but are completely unrealistic in hardware?

Not even unrealistic…

Frames’ individual output N is always equal to input N x a positive gain.

The core of the module is a quad VCA.

> Frames’ individual output N is always equal to input N x a positive gain.

Thanks for sharing that info! I suspected that there was a restriction like that.