The SonicSmith ConVerter E1 is just becoming available: it’s allegedly the fastest pitch-tracking device yet, and it produces v/oct pitch output, plus CV envelope out.
As a fretless bass player and occasional slide guitarist, I’m kind of excited about being able to get non-quantised pitch control. MIDI’s good, but its limitations start showing when you add vibrato or deliberately play a little sharp or flat.
The problem is getting this into VCV-rack with the least possible added latency. The simplest path I can think of is a USB voltmeter with a module that continuously samples its output, and this leads to the question: is there already a module for this that I haven’t found yet?
Alternatively, does anybody have a reasonable guess as to how hard it would be to write one, as their first VCV-rack module?
I did find two pitch-to-voltage modules, but latency is the killer when a guitar is involved, especially when you get down to bass frequencies.
Sounds like you are asking “how do I get a CV into VCV”, right? the typical answers will be “any DC coupled sound card, but the stuff from expert sleepers is best”. I’m sure others will have something more accurate to say on the matter.
Interesting piece of gear! Will definitely keep an eye on it.
Yep, per @Squinky, Expert Sleepers is best. Get the ES-9 if possible. You can just plug the E1 straight in to it. Relatively many interfaces have DC-coupled outputs, but DC-coupled inputs, which it has, are very rare (and make it, quite literally, a continuous USB voltmeter). input (ADC) latency won’t be an issue as it will be minuscule (like the equivalent of moving a few feet back from a speaker) and, more importantly for live tracking, perfectly stable. Also, if you have or build a bigger hybrid system, the ES-9 will have a million other uses.
Cool! Definitely cheaper/smaller than the ES-9 and multifunctional. Those interface using Os’s DC-via-AC trick, right (plus the corresponding Rack or VST decoder)? Or do they serve as a USB interface as well?
I imagine a good commercial VST pitch tracker (Waves, iZotope, Melodyne, Antares, etc) will give better results than a hardware unit. VCV Host will have MIDI output soon, so you’ll be able to send your instrument’s audio to a Host module and generate a MIDI note + pitch bend signal in VCV Rack.
regarding Waves. since their v12 upgrade i dont consider them good anymore. mine dont even show up in Host FX and in other environments the GUI’s are reversed and components out of place. the only thing i have where they work properly still is Reaper.
Thanks! It looks like this is the answer I was looking for.
It even looks like it’ll handle both v/oct and envelope CV for all 6 modules, if I take the full plunge and get a converter for each string for full-poly adaptation of guitar to modular.
The need for calibration had occurred to me, but that article goes into much more depth than I’d reached so far.
That looks incredibly useful; thanks!
Admittedly, this discussion also has me thinking that the smart way to go could be to do the initial tone generation and pitch-sensitive modulation in hardware, then pass the audio into the computer and drive the rest via MIDI and an envelope-follower. A bit of latency is unlikely to be a major issue for those stages.
Nice! If you do this, please send audio clips of your ultimate guitar synth I did see in their video that they’re offering a hex-pack…
If I were you and were thinking of going the hex route, I’d start with a workflow that patched the CV in and did audio generation in the box (and put the $$$ you’d save by not buying Eurorack oscillators in sets of six into the most powerful computer you can get). You’ll have so much more flexibility that way, and you can record your performances and resynthesize from them as you like.
Once you have the CV in Rack (which, with the ES-9, will happen exactly as quickly as getting audio into Rack, since it doesn’t distinguish) the audio generation is going to be effectively instantaneous (happening within one block). If you’re running 96kHz at 256 samples (not at all unreasonable for six-voice polyphony), that’s a constant 2.7ms, which is like stepping three feet back from your speakers. (Add ADC and DAC delay and you might be doubling that, so six feet back). The key point is that it’s constant; most musicians find no trouble adapting their playing to much longer delays than this as long as the delay length is stable. (The issue with MIDI latency is that the timing can actually be inconsistent in many contexts, but with straight CV you get to avoid that).
Also bear in mind that if you were driving hardware analog oscillators outside and doing in-line effects like filtering/VCA/etc. in the box, you’d be incurring exactly the same delay anyway (unless you were monitoring the raw oscillators before the interface, which might not be a very inspiring sound!) And if you were driving hardware digital oscillators, there would be similar latency as with an interface into Rack–possibly more, depending on the oscillators. There’s no getting around a few ms of delay with anything digital.
Prove it out on one string first if money is at all an issue, as a guitar player I know the road to satisfaction with pitch converters is long and winding, and best taken one step at a time. My favorite guitar-computer interface so far is Jamstik’s studio guitar.
And that’s speaking as someone who has worked on my own in music instrument manufacturing. It’s a genuinely hard problem due to the slowness of the wavelength of the low notes. Without fret scanning you’re left with trying to determine the pitch based on the voltage output of the pickup, and the best thing I’ve seen in the software front are Jam Origin’s Midi Guitar 2 which uses predictive AI, and that actually works pretty well. I’m not sure what secret sauce is in Jamstik’s studio guitar, but I prefer it to Midi Guitar 2, and those are the two best off the shelf guitar midi interfaces that I’ve played with. I would give up on using waveform recognition for bass guitar, if you calculate the time constants of the lowest wavelengths and assume you only need to recognize one or half a wavelength (I think usually you need more to be stable), I think you’d find the latency to be inherently unacceptable. Although, give it a try, it’s possible with bass you don’t need to get the timing so exact, but I find that hard to believe. I think even bassists want a snappy fast response to their plucks.
It seems like the best thing you could do is take a guitar and tune everything way up, so that the frequencies are really high, use pitch tracking to identify the pitches because it will get them quickly that way, and then scale back down in software to bass range. I might have some blind spots, so I’d welcome any disagreements or arguments.
I did see in their video that they’re offering a hex-pack…