P12: High resonance filter sweeps have (annoying) noticable steps

[hype1]

Hi DSI,

I am a big fan of your Prophet 12. It is the main used synth in my studio for a while now.

One thing annoys me though on almost in every patch I make: the stepped filtering (in semitones). I do a lot of patches for ambient and abstract works and I tend to use a lot of resonance as well. When I have a patch going on that is super dreamy, monotonous and abstract, on most synths I am turning the cut-off knob very slowly to bring in some change of that nice creamy resonance peak. On the P12 however it annoys me that those subtile change I wish for are very much stepped. When I have a monotonous patch, the filter plus resonance manual slow sweeps will introduce semitone stepping (melodic stuff, feels very 'digital') that I really don't want.

Is there a way to turn this stepping off or perhaps could you please consider implementing an option for turning this behavior off in an update?

Thanks from a fan.

[jdt9517]

I do not have that problem on my P-12.  Make sure your LFO is set correctly.  I believe there is a setting on the LFO's which allows for stepping in time to the clock.  Make sure that your LFO is set for a regular sweep.

Edit:  Reading your post a little more carefully, it looks like you are trying to do the sweep by manually moving the cutoff/resonance knobs.  You will get stepping using that approach.  You need to either use a pedal or set it to an LFO to not get the stepping. 

[hype1]

Haven't tried it out yet, but heard connecting a pedal doesn't make a difference.

Is this your experience?

[jdt9517]

I tested it last night after responding your post.  The pedal will smooth out the swells.  I don't totally understand the tech, but turning the knob creates incremental increases or decreases in filter.  The pedal, OTOH, makes it behave like an analog adjustment - silky smooth!   

[hype1]

Not sure why you would get other results other than not testing thoroughly but I just added an analog pedal and the stepped behavior is definitely still there, most noticably with high resonance settings.

[DavidDever]

Might be worth confirming which s/w version you are (both) running, as well....

[jdt9517]

Good point David.   Mine is 1.4.0.  Here is a recording I just did with a basic string sound and modulating the HPF with a pedal.  Other than not great pedal technique in one area, the sweep is smooth and no digital stepping. 

https://soundcloud.com/jdt9517/2017-01-27-hpf-sweep-pedal-p-12

[hype1]

System version 1.4

Thank you for your effort but I am talking about Low Pass Filtering plus I don't hear much resonance going on in your patch.

For clarity just use a basic patch (Global Settings > Basic Patch > Write) and set resonance full and play with the cut-off frequency of the Low Pass Filter. Slowly turn knob, or use an expression pedal (I just soldered a potentiometer to tip-ring-sleeve on a balanced jack, exactly the same circuit used inside expression pedals) and you will hear very clear stepping when turning or moving slowly.

[DavidDever]

It probably is worth pointing out that some frequencies would be more pronounced than others, by dint of their position on the harmonic spectra of the note that is being held; if you notice the same thing happening when using the envelopes to modulate the cutoff, then it's pretty safe to assume that it's normal operation.

[hype1]

It is normal and designed operation. But I am looking for a way to change this normal operation or request DSI to come with a solution like an option for continuous filtering instead of in semitone steps.

[tumble2k]

Just in case what DavidDever said wasn't clear, if you sweep the filter with high resonance you will pick up discrete overtones of the waveform. This will sound like stepping, but it isn't actually stepping. It's just that a periodic waveform has specific overtones at the root frequency of the tone and multiples of that frequency. For example, a sawtooth wave at 440 Hz is made up of sine waves at 440 Hz, 880 Hz, 1320 Hz, 1760 Hz, etc. As you sweep the filter over this waveform you will get discrete tones at each of these frequencies but no tones in between these frequencies. It will sound a little like stepping.

This is not a function of the design of the filter or its control. It's a manifestation of a Fourier transform applied to a periodic waveform.

[hype1]

As you sweep the filter over this waveform you will get discrete tones at each of these frequencies but no tones in between these frequencies. It will sound a little like stepping.

Not true, try it.

Also, from the manual:

Frequency: 0...164—Sets the low-pass  lter’s cutoff frequency over a range of more than 13 octaves, stepping in semitones. The cutoff transistions smoothly across the values when the frequency is swept.

When for example I use the filter knob on my Moog Sub 37 I get the full range of resonance at slow movement of the filter knob. I think that resonance can happen at every frequency not just at multiples of the base frequency. Also I never notice steps when slowly turning the filter cut-off knob on my modular synth, but that's fully analog.

[tumble2k]

The Dave Smith synths set up their LPF frequency knobs with semitone resolution for editing. It's not so much meant as a performance tool but rather as a sound editing tool. If you want to smoothly sweep the LPF (or HPF) you need to use another modulation source, for example, the mod wheel, a pedal, or an envelope. If you do that you'll get the full resolution. This is different from the Sub 37. Again you need to remember that if you're using high resonance and your cutoff frequency is close to the fundamental frequency of your waveform you'll get discrete frequencies. As the cutoff gets higher they'll blend in more.

[jdt9517]

@hype1-  tumble2k's explanation is right on the money.  I did the HPF because I thought that was the filter you were complaining. The pedal woks famously on the LPF.  Yes, there will be semitones that stand out but Fourier explains that phenomenon quite well. Of course, the math is a little intense in that analysis. 😳

[tumble2k]

Just trying to help you fix your issues. Everything we've read so far indicates your synth can be modulate the HPF without stepping. We have used pedals/mod wheel to eliminate stepping and it works. You can also change the mod amount of the mod wheel to tighten the HPF range, which might help improve the mod wheel resolution.

If you don't want help we'll be happy to provide none.

[proteus-ix]

You guys don't read, do you?

This is how you talk to people who are trying to help you?

[Note: I've removed a foul comment from this post.

- Sacred Synthesis]

[hype1]

I admit being unthoughtful and ready to bite the dust, so bear with me:

Sorry, I was a bit tense last night and I have no reason at all to use my words like that. I apologize and want to thank you all for help.  (I would rather remove my negative comment for the sake of this thread but the forum does not let me...)

[Note:  I removed it.

- Sacred Synthesis]

"Yes, there will be semitones that stand out but Fourier explains that phenomenon quite well." It is this behavior that I don't find on other analog synths and therefore Fourier does not explain this behavior in my opinion and I already know that it is a feature implemented by DSI as stated in the manual which I quoted above.

The only thing I ask for is a work-around, or hopefully an option in the software to turn this off. So far we've found a pedal not changing things, and mod wheel doesn't change behavior either. Hopefully DSI staff will see this and are willing to think about a way to implement this as an option in the software.

[tumble2k]

I admit this wasn't on a Prophet 12, but I used the mod wheel to modulate the Prophet'08 LPF. I used a modulation amount of 127. I turned off the oscillators and put in a little noise. Then I set the resonance to max (4 pole mode) so the filter would self oscillate. With the second layer I played a single triangle oscillator. By using two layers I could listen to the filter beating against the oscillator and readily discern the number of filter resolution steps.

I found that the mod wheel was able to sweep the LPF frequency with fair resolution. I was able to get four steps per semitone when sweeping upward and more sweeping downward (why the discrepancy in direction I have no idea) around a frequency of 220 Hz.

hype1, please try to modulate the HPF with the mod wheel, turn resonance way into self oscillation and turn your oscillators off. If you still hear semitone steps tighten up your mod wheel modulation amount.

[jdt9517]

"Yes, there will be semitones that stand out but Fourier explains that phenomenon quite well." It is this behavior that I don't find on other analog synths and therefore Fourier does not explain this behavior in my opinion and I already know that it is a feature implemented by DSI as stated in the manual which I quoted above.

What we are talking about is not the digital stepping experienced when using the filter control knobs.  Rather, there will be certain harmonics which will stand out.  What is heard is more of an amplitude (volume) change of harmonics as the filter sweeps.  It's actually part of the effect you are looking for. 

The knobs incrementally change the filter in steps.  The pedal, mod wheel, and envelopes change the filter based upon voltage in the same way as any other analog filters are adjusted, e.g., Moog.  There shouldn't be digital stepping.   Please record a track with the problem, and link it here.  Then we can discuss more the issue more fully.  Thanks.

[Sacred Synthesis]

Fellas, there's no problem with having a somewhat heated discussion, but let's keep the foul language out of it, okay?  This is not GS.