I’ve got a sequencer sending V/OCT and and a clock sending GATE voltages to Bogaudio’s ASSIGN in order to get 4-channel polyphony. The gates from that pass through Count Modula’s POLY GATE MOD, in order to have some overlap in the gates. The polyphonic V/OCT from ASSIGN goes to VCV CV>MIDI V/OCT, and the the lengthened polyphonic (4 channels) gates from POLY GATE MOD go into GATE of CV>MIDI. The midi device is “IAC Driver Bus 1”
I have GarageBand listening to IAC Driver Bus 1 and have selected “Female Chamber Choir” for an instrument. All is well so far, with all four of the polyphonic channels working. Positively angelic. What comes back via “Blackhole 2ch” to VCV AUDIO is non-polyphonic (but stereo) audio.
Now, I want to do is introduce a Bogaudio SLEW module between the output of the POLY GATE MOD and the VOL of CV>MIDI. It sort of works, but only while channel 1 is high, so I only hear audio for approximately 1/4 of the time, ie. when the channel 1 gate is high, and it raises/lowers the volume on all channels whose gate is also high at the same time as channel 1.
I tried connecting the poly gates directly into the VOL jack without the SLEW module, in case it couldn’t deal with polyphonic signals (which it does, actually). Similar results, but more choppy.
If I manually adjust VOL with one output of an 8VERT it raises and lowers the volume on all four channels.
By way of experiment I merged four outputs of an 8VERT and plugged that poly signal into VOL, it also raises and lowers the volume of all four channels, the same as when using a single (non-poly) voltage. And with the poly signal, only adjusting channel #1 has any effect on VOL.
Am I misunderstanding how midi VOL should work, or is it some misfeature of GarageBand?
I suppose that would be best. I’ll see what I can do to minimize it, and get back to you.
If I do that, then according to my Generic Technology Trouble-Shooting Guide rule #4, I will not be able to reproduce the issue and the problem goes away.
John Rose’s Generic Technology Trouble-Shooting Guide
Plug it (or some peripheral device) in.
Turn it (or the peripheral device) on. Repeat a couple of times if necessary.
Read the instructions. Are you sure it’s supposed to do what you think it’s supposed to do?
Call a technician. When they show up, you will not be able to reproduce the fault.
Open up the case, look inside and say “Hmmm.” Close it up again and it might work (even if there are leftover screws).
Blame the Telecom Company / Electric Utility / ISP / Sunspots/ Sun Transit / Rain Fade / etc. Basically any large faceless institution/entity that will probably not care about your puny problems. It’s out of your hands, but do try to get a ticket number to follow up on.
I compiled this guide over 27 years of working with broadcast systems from the studio and portable equipment to AM/FM/Shortwave transmitters and satellite up/downlinks. Mostly analog hardware. You typically escalate to the next step about 1/5 as often as escalating from the previous step.
Not very, now that I’ve adapted it to use far fewer few plugins (VCV Fundamental, Bogaudio, and NYSTHI) and deleted all the unused/disabled extra modules left over from my default “New” patch. I only left enough for “proof of concept” and trouble-shooting the issue of interest.
I set it up for only two channels of polyphony. Play with the top two 8VERT knobs to see how the VOL levels behave independently.
I had GarageBand use the “External Headphones” for output, rather than hassle with the extra complication of using the Blackhole virtual device to pass the audio back to VCV Rack AUDIO. That comes later.
Plug-ins used: VCV Fundamental, NYSTHI, Bogaudio,
Other details:
macOS Sonoma 14.7
GarageBand 10.4.11
VCV Rack Free 2.5.2
My cable colours usage:
blue - clock, gate, trigger
red - audio (Oh … none in this patch, because GarageBand handles all the audio.)
Huh. Here’s a potential work-around: the GB Orchestral > Choir sounds all have their own Attack knobs (no numbers, just adjust to taste). It does lengthen the envelope attack, but there’s nothing for the tail ends. Using their Reverb setting sorta’ does it, but it isn’t quite right, and I’m still hoping to find a general solution for any of the GB instruments.
The CV to MIDI module can only send on one single MIDI channel. You can play polyphonic notes over the inputs V/OCT, GATE, and VEL, but all the other inputs are in fact mono.
The 3 inputs V/OCT, GATE, and VEL are combined to MIDI note-on messages when GATE gets high and MIDI note-off messages when GATE gets low.
The input VOL refers to MIDI CC #7, and a poly input here is simply ignored.
When you want to achieve that different polyphonic MIDI notes have different velocities, then the CV inserted into the VEL-input is taken at the moment the corresponding GATE value changes to high.
When you want to use volume (MIDI CC #7) as some kind of envelope for a single note, then you have to use MPE where every single MIDI note has its own dedicated MIDI channel. For this to achieve you need more than one CV to MIDI module, and each module must be set to a different MIDI channel.
Well, there’s a big piece of the puzzle, the kind of thing that should be mentioned in the manual. I could’ve stopped at Rule #3.
That’s what I did last night, using four CV>MIDI modules. I SPLIT the V/OCT, GATE, and VOL signals such that each CV>MIDI module had it’s own.
It sorta’ worked, and gave me per-channel control of the volumes, but the results are less than perfect. I get a lot of stuttering, especially when the VOLs are ramping down, for some reason. The effect is similar to using too small of a Block size in the AUDIO module. I tried changing the number of threads, but no joy there.
At any rate, using the “Attack Time” knob in GarageBand nearly does just what I was hoping for, and adding reverb helps to smooth out the tail ends.
Using BlackHole to bring the GB audio back through Rack for further processing has been working out well, once I got everybody using the same sample rate. SWMBO is becoming pleased with the results too. She’s writing a play about three out-of-work sirens (the mermaid kind) stuck on their island with fewer sailors being tempted by their singing. My patch will provide the singing. It’s got a hint of crashing waves too.
I deem this to be what will have to pass for a satisfactory conclusion to this thread.
This is because the CV value with a precision of 23 bits (over 8 million values) is truncated/quantized to a MIDI CC value with a precision of 7 bits (only 128 values).
The scope on the left shows the CV value and the scope on the right shows what the ramp looks like when it is truncated to MIDI CC.
I’m talking about complete dropouts for individual notes, maybe up to 250 milliseconds (?) or so of silence, not just “noise”. I don’t get the stuttering with a monophonic volume cv into a single CV>MIDI module.
This screenshot shows one of four “midi volume” tracks highlighted in yellow. You can see the other three in fainter yellow. There should not be any vertical lines. I can edit them in GB to get rid of the stairsteps (that aliasing you mentioned?) and give them all smooth slopes with fewer points, but it still sounds bad. Now I’m thinking it is a limitation of GarageBand.
Huh. Here’s a potential work-around: the GB Orchestral > Choir sounds all have their own Attack knobs (no numbers, just adjust to taste). It does lengthen the envelope attack, but there’s nothing for the tail ends. Using their Reverb setting sorta’ does it, but it isn’t quite right, and I’m still hoping to find a general solution for any of the GB instruments.
