Automation vs Internal Modulation

[keyxmusic]

Hi,

When playing with Bazille CM, I noticed that when I used automation to draw a sharp shape, it didn't sound as sharp as, for example, internal saw ramp down LFO. That made me wonder if this was a Bazille CM problem, a DAW problem (Cubase 9.5) or just in general how VSTs work - internal modulations are always more accurate than automation?

Also, on that note, is there technically a better method between direct automation vs MIDI learn and automate the MIDI CC?

Thanks!

[daily patcher]

DAW automation can certainly play a part. This series will enlighten you:

[Urs]

DAW automation is like a remote control for turning knobs by mouse. It sounds scatchy if we don't smooth it out. So we smooth it out.

The reason is, the plug-in gets the mouse/automation information only when the value actually changes, and only from time to time. The plug-in might get 10 update sper second, or maybe 100. But not 44100 or 96000. In Bazille the LFO creates 88200 updates per second, if run at 44.1kHz. That's a proper resolution. 10 is not. 100 is not. Not even 1000 is. That's why LFO is tight and automation is somewhat smoothed out and fuzzy.

[keyxmusic]

Thank you all!! This is some really good info that I've never seen mentioned else where.
What is the MIDI CC automation update rate compared to plugin direct automation? I know it's not as accurate value wise, since it's only 7 bits, but how about timing wise?

[Urs]

Thank you all!! This is some really good info that I've never seen mentioned else where.
What is the MIDI CC automation update rate compared to plugin direct automation? I know it's not as accurate value wise, since it's only 7 bits, but how about timing wise?

If you use the internal MIDI modulation sources (AT, MW, PB, CtrlA/B) you get a very accurate update rate. It's time stamped and the smoothing is selectable in the preferences section.

[keyxmusic]

I see. But what if I just MIDI learn something and automate it via MIDI CC? Is that gonna give results just as slow as DAW automation?

The plug-in might get 10 update sper second, or maybe 100.

Is that number DAW-dependent? What about DAWs that have their own modulation system such as Bitwig or Max for Live? Surely they have to be a bit more precise with their approach?

[Urs]

I see. But what if I just MIDI learn something and automate it via MIDI CC? Is that gonna give results just as slow as DAW automation?

Yes, pretty much.

The plug-in might get 10 update sper second, or maybe 100.

Is that number DAW-dependent? What about DAWs that have their own modulation system such as Bitwig or Max for Live? Surely they have to be a bit more precise with their approach?

I can't speak for these. They might process plug-ins with smaller blocks of samples to increase precision. However, we still need to smooth parameter changes to avoid scratchy mousing. That's not going to change.

The MIDI modulations/articulations I mentioned are intended to be used as interface for external modulations. I can only stress that using these is the cleanest and best solution for timing critical stuff.

(the best way would be modulation signals on separate audio channels, but that seems overkill and not supported at all in current DAWs)

[keyxmusic]

Thanks again Urs, this has been very eye opening for me

[muki]

> In Bazille the LFO creates 88200 updates per second...

question:
is *this* "the thing" that dave smith means, when he emphasizes the very high scanning rates (kilohertz/megahertz?) in his synthesizers ?

[Urs]

> In Bazille the LFO creates 88200 updates per second...

question:
is *this* "the thing" that dave smith means, when he emphasizes the very high scanning rates (kilohertz/megahertz?) in his synthesizers ?

Possibly... the old vintage Prophets (the polyphonic ones) had extremely bad update rates for digital control, you could hear stepping on many knobs (they didn't have digital LFOs though...). But even the Prophet-08 can have audible stepping on its envelopes. So it makes sense that Dave points out that his newest stuff has a very high update rate on digital control/modulation.

[muki]

if i recall it corrctly he emphasizes it esp in re prophet 6/ob6
but also the new rev2 and X i think...

question2:
if i take a softsynth
apply a lfo to an oscillator pitch resulting in exagerated vibrato so to say
and i hear kind of zipper noises, clicks or crackles
it might be due to too low scanning rates, i.e.: not in the 2x overdsampling range... ?

[Urs]

question2:
if i take a softsynth
apply a lfo to an oscillator pitch resulting in exagerated vibrato so to say
and i hear kind of zipper noises, clicks or crackles
it might be due to too low scanning rates, i.e.: not in the 2x overdsampling range... ?

LFO -> oscillator pitch should be free of zipper, clicks and noises. You're modulating a derivative (car analogy: driving speed rather than distance driven), so there should not be any significant discontinuity, regardless of modulation update rate. That is, the car may look drunk, but it doesn't teleport.

(If you encounter such in any of our synths, I'd like to hear it...)

[muki]

> If you encounter such in any of our synths

none of yours!

i tried it with hive, repro1 and diva (dual vco)
all of them zipperfree and fine

it's another newly released animal...

[Urs]

Ok, cool

I can think of a scenario where a modulation might sound like that, when the modulating waveform/envelope has a steep spike and a reasonably tuned oscillator is occasionally doing like half a cycle or so in a few samples.

[muki]

> Ok, cool

ui! zu früh gefreut...
email ergeht an support