On my PC I typically run with 2 or 3 threads. In this case I happened to be set at 3 threads, and I got ~55% avg 75% max as long as I wasn’t recording. When recording the max would often spike to 150+ and I would get live pops, but the recording is clean.
I just checked, and on my PC the sweet spot is 4 threads, giving ~50% avg, 60% max.
On my M1 MacBook Air everything hums along beautifully with 1 thread. Rack x64 yields ~50% avg, 52% max, and Rack ARM yields ~38% avg, 42% max
That REFTONE is a red herring. I just needed a controller that would give me integral values between 2 and 5 to control the number of sequencer steps, and that was the first module I found that worked. I’m sure there are better, more intuitive choices. I really need to update my KNOB 5 to have integer and semitone quantizer options. That was my original plan, but I got lazy in that release.
All of the Solar 42 drone voices are designed to be tuned by ear. There are no drone V/Oct inputs - all FM is linear. Voices 4 and 5 are V/Oct, but my emulator does not provide any quantizer (but it is easy to add via the patch bay). All quantization in the Solar 42 is provided by the capacitive key pad, and I did not try to emulate any portion of that. Besides quantizer, the keypad provides arpeggiator and 16 step sequencer, and more.
Definitely read the Solar 42 manual (link in first post) for the approximate range of each oscillator. I did my best to tune the range of each audio oscillator to match the Solar 42 hardware. But there wasn’t any documentation for the range of the LFOs or the voice 3,8 modulators. I just used what I thought sounded good.
Yeah, that is why the Solar 42 is probably a niche product. The workflow is intentionally very old school. I opted to preserve that ethos in my emulator. It might be a good cheap test to see if you might be interested in the Solar 42 hardware.
Wait, you meant in the patch? I think it’s probably the Surge Wavetable VCO, but Supermassive can make pads sound like a choir with the right settings.
Version 1 of the emulator had a serious bug – the FM CV for drone voices 1, 2, 6, 7 was completely broken. It didn’t effect the demo patch because it didn’t use the FM CV.
Below is a fixed version 1.1 that also uses polyphony to reduce the number of Bernoulli Switches from 6 down to 1, without changing the character of the demo patch.
I also noticed that I have not quite implemented voices 1, 2, 6, 7 correctly, as I am using linear FM, but the Solar 42 is neither linear nor exponential. I hope to release a version 2 soon that corrects those voices, and also adds external inputs, and some other inputs/outputs to the patch bay.
Sure, as long as you didn’t attempt to use the patch bay CV input for voices 1, 2, 6, or 7 - it was totally non-functional. Version 1.1 fixes that bug.
That sounds great! Something to be said for just stacking up tons of VCOs. I tend to use polyphony these days instead, just makes things easier to patch. It’s really easy to make ridiculously thick sounds.
That does sound beautiful. However, unless I am missing something, your emulator is missing an important feature of the Solar 50. It looks like each voice in your emulator has a knob to simultaneously modify the pitch of all 5 VCOs. But the hardware has a feature that causes the knob to cause the 5 VCOs to cross modulate each other once you go past 12 o’clock. Presumably this uses the voice mix as FM input, and the knob serves as an FM attenuator from 12 to 6.
Voices 1,2,6,7 in the Solar 42 also have this feature, which I implemented in my emulator. I also tuned the range of each VCO to match the range specified in the manual.
I have completed my version 2 of the emulator that adds some missing patch bay inputs, as well as adds additional inputs and outputs that don’t exist in the Solar 42 hardware.
The manual states the drones are neither V/oct nor linear, but I have nothing to match against, so I don’t know how to make changes to better emulate the hardware. I opted to keep using through zero linear FM for my emulator drones.
The big important change to version 2 is the new polyphonic V/Oct send/return inserts for all 8 voices (26 oscillators!). By default the sends are patched directly to the returns, but you can break that connection and insert polyphonic quantizers to very easily get standard tuning. I have also changed the design a bit so that all VCOs are tuned to the standard C4, and all pitch manipulation is done via CV so it can be quantized.
Tuning by ear is an important philosophy behind the design of the Solar 42. But as @purf demonstrated, not everyone wants to take the time to tune all those oscillators by ear. @k-chaffin - I’m thinking you in particular might be interested in these changes. “Out of the box” the emulator sticks to the Solar 42 philosophy, but it doesn’t take much effort to break out and use more modern work flows.
Here are a pair of very different example patches using the emulator, of which the second demonstrates how to use a quantizer with the emulator.
The sonic possibilities are extensive with the Solar 42. I challenge anyone to see what they can do, patching anything they want into the patch bay, but sending everything through the emulator, and only using effects after the emulator left/right filter outputs. And if you take up the challenge, please post what you have done here, or at least add a link.
I had a little time again and listened to your Solar 42. Really fantastic. Then I tried your patch and unfortunately without success. It forces my computer to its knees. Of course it’s not your fault. I only have an i5. But you did a good job anyway.
