Keyfollow and FM and Tracking question
-
lunarfootprints lunarfootprints https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=314029
- KVRer
- 3 posts since 8 Oct, 2013 from New York City
I'm trying to do FM without the FM module. So far OSC1's TuneMod is connected to an LFO, and the rate of that LFO is being 100% modulated by KeyFollow. This keeps all my sidebands in line when I'm doing amplitude modulation, but an article I'm reading says I have to have the modulator signal double for every increase in the carrier frequency (OSC1). Any way to do this? I thought of having OSC1's tune KeyFol by .5, but that obviously doesn't help. Help!!
-
- KVRAF
- 1617 posts since 11 Dec, 2008 from Minneapolis
Unfortunately I think this approach isn't completely workable, although still an interesting exercise. One issue is the range of rates available, I believe LFOs scale from 0 * base rate to 8 * base rate, so with .1sec at maximum: .0125 seconds or ~80 hz (control rate issues may arise, not sure exactly how that works out), while oscillators scale from 8hz to 12,500 hz. I believe there's also an issue of linear versus exponential FM - FMOs the former, tune-mod the latter.
My guess is, for the article you're working off of FMOs are appropriate - FM of the 'PM' / phase modulation type, which John Chowning developed and the DX-7 made legendary. Your experiment actually does make some sense in ACE or Bazille, though. Bazille really has a whole family of FM techniques available.
My guess is, for the article you're working off of FMOs are appropriate - FM of the 'PM' / phase modulation type, which John Chowning developed and the DX-7 made legendary. Your experiment actually does make some sense in ACE or Bazille, though. Bazille really has a whole family of FM techniques available.
-
lunarfootprints lunarfootprints https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=314029
- KVRer
- Topic Starter
- 3 posts since 8 Oct, 2013 from New York City
Super helpful response. It did seem like even the highest end of pushing up the rate barely gets into the speed I want to be in. Only thing that I don't get is why zebra's FMO's don't produce sidebands below the carrier frequency. They should be there, right?
I did find a nice way to create really fast loops with the MSEG, by making the distance small and the loop speed high, but sometimes at high speeds (higher, I think, than the LFOs can reach) the modulated aspect just gets blurry (perhaps that's why they don't have audio-frequency modulators in Zebra?). Only the wave editor seems to get around this with the adjustable resolution. Am I right to assume that the rest of Zebra's parameters have pretty low resolutions? I was morphing through "bandworks" and I could hear the discrete steps, where I wanted the change to sound continuous. But I'm not sure if it was the steps taken by Bandworks or those taken by the LFO, or both.
Is this something other softsynths get around? Maybe why zebra seems to be the less-popular underdog? As opposed to, say, Sylenth or Massive? I may repost this question in a new topic.
I did find a nice way to create really fast loops with the MSEG, by making the distance small and the loop speed high, but sometimes at high speeds (higher, I think, than the LFOs can reach) the modulated aspect just gets blurry (perhaps that's why they don't have audio-frequency modulators in Zebra?). Only the wave editor seems to get around this with the adjustable resolution. Am I right to assume that the rest of Zebra's parameters have pretty low resolutions? I was morphing through "bandworks" and I could hear the discrete steps, where I wanted the change to sound continuous. But I'm not sure if it was the steps taken by Bandworks or those taken by the LFO, or both.
Is this something other softsynths get around? Maybe why zebra seems to be the less-popular underdog? As opposed to, say, Sylenth or Massive? I may repost this question in a new topic.-
- KVRAF
- 4197 posts since 23 May, 2004 from Bad Vilbel, Germany
-
- KVRer
- 14 posts since 18 Nov, 2009 from uk
Yes they should do, but I'm not 100% sure what you are trying to do... posting your patch would help!lunarfootprints wrote:Super helpful response. It did seem like even the highest end of pushing up the rate barely gets into the speed I want to be in. Only thing that I don't get is why zebra's FMO's don't produce sidebands below the carrier frequency. They should be there, right?
A lot of synths, both hardware and software, have slower control rates for modulators. For example, I'm pretty sure my old Nord Micro G1 updates at around ~400Hz... This is because it is less cpu intensive to "downsample" control signals, freeing up resources for more polyphony or other DSP.lunarfootprints wrote:Am I right to assume that the rest of Zebra's parameters have pretty low resolutions? I was morphing through "bandworks" and I could hear the discrete steps, where I wanted the change to sound continuous. But I'm not sure if it was the steps taken by Bandworks or those taken by the LFO, or both.
In FM terms, this downsampling process means that you cant actually modulate a signal beyond the control frequency without introducing serious aliasing. I cant comment for sure on zebra vs other synths, but i suspect others have the same limitation. In fact, Bazille and Ace are probably some of the few soft synths that actually have a lot more audio rate modulation capabilities than others (so maybe try one of those instead of zebra).
-
- KVRist
- 431 posts since 27 Sep, 2005
As I understand, the main problem in this approach (besides the limit of LFO frequency) is the fact that oscillators frequencies are scaled exponentially, and LFO's frequencies are scaled parabolically. So there is no way to do perfect keyfollow of LFO frequency.