Synth Forum

Notifications
Clear all

Filter Modelling

23 Posts
2 Users
0 Reactions
558 Views
Antony
Posts: 745
Prominent Member
Topic starter
 

There was a thread a while back, in which I recall @Andrew claimed the Filter EG did not modulate Cutoff, only the amount of filtering.

I can't find the thread or be bothered searching harder.

EDIT:- Found it

https://www.yamahasynth.com/forum/how-to-make-cutoff-automatization

I have dug out and set up my MODX7 today and wanted to investigate.

Keeping things basic with single LPF Filters:-

The Cutoff Parameter does alter the Cutoff Frequency. Above this Cutoff Frequency, higher Frequencies are attenuated by 24dB per Octave (or 18 or 6 depending on the chosen filter).

This is easily demonstrated by choosing a Saw Wave Element, and Reducing LPF Cutoff until the Saw Wave sounds like a Sine Wave (I.e. the Fundamental Note). The LPF is cutting off the Higher Harmonics of the Saw Wave.

The Cutoff Range (0-255) is typically a lot greater than the Harmonic Range (Harmonic Spectrum) of a given "Oscillator" (i.e. A single Element in AWM2 or Summed Carriers in FM-X).

Therefore, in order to hear Cutoff Modulations, the Cutoff Frequency must first be set in the vicinity of the Played Note's Timbre, such that any movement in Cutoff is heard as a change in Timbre.

It is easier to verify this with Cutoff/Key = 100%, i.e. the Cutoff Frequency tracks the actual Key (Note) being played. This way, changes can be heard for any key played.

In the case of Filter EG, With Hold Level = 0 & Attack Level = 127, a Key-On Event will increase the Cutoff Frequency from its current Value (e.g. 105) by an Amount dependent on the Positive FEG Depth Value. If FEG Depth = 0, there will be no Change to Cutoff. If the FEG Depth = +64 (Max), the Cutoff Frequency will undergo its maximum deflection (Since Attack Level = 127).

It is hard to be sure of the Range of this maximum increase in Cutoff Value. By my Tests, whatever the value is, it seems to be less than 255.

If true, this means the FEG cannot be used to increase the Cutoff Value from 0 all the way to 255. But this is OK, you probably wouldn't want to or need to.

However, it does mean that FEG movement, for small Depth values, may not audibly affect the Timbre of the Note, if the Initial Cutoff Value is outside the Note's Harmonic Spectrum (see above).

Also, for High FEG Depth Values, the Cutoff may move above a Frequency that has no audible affect on the Timbre. Equally, the Cutoff Value may max out at 255, and prevent the FEG from changing it any further.

This "going beyond useful maximum" is known as "Pinning". Any movement beyond the "pin" either has no audible affect, or is not permitted since the Max has been reached.

This will lead to short periods, where it seems the Cutoff is stationary and seems like the FEG is "Not doing Anything".

Therefore, it is important to set the Initial Cutoff and set the FEG Levels and Depth relative to what you want to achieve.

Watch out for other parameters that can change Cutoff Levels, FEG Levels or FEG Times... I.e. Velocity and "Per Key" Key Tracking.

 
Posted : 19/11/2022 5:23 am
Antony
Posts: 745
Prominent Member
Topic starter
 

Regards unpleasant "Squawking" Filter Resonance.

I too experienced this and figured out the Solution.

First consider what Resonance is doing. It is boosting (amplifying) a Frequency Band centred around the Cutoff Frequency Value, where ever it is set. The higher the Resonance, the Narrower the affected Frequency Band, and the greater the amplification.

In real world Analogue Filters, it's possible to narrow down Resonance to single Frequency Sine Waves (Harmonics)... e.g. "Pick Out the Harmonics". Analogue Filters can also Self Resonate... MODX Filters cannot.

The digitally modelled Filters on the MODX do not have quite the same finesse, but do serve to create facsimile Sounds you might hear an Analogue Subtractive Synth make.

As a test, choose a "Saw Wave" or "Square Wave" Element, and set the Filter Type to LPF24A.

While playing and holding a note, sweep the Cutoff Value down to a point where the sound disappears, then slowly move it back up to where the Sound comes back.

Set the Resonance to about 70, and while still playing/holding the Note, sweep the Cutoff Value back up through its range. You should hear the Resonance start picking out "Stronger" (louder) frequencies... these are the Harmonics of a Saw or Square Wave, that follow the Harmonic Series (Saw = Odd and Even, Square = Odd Harmonics only).

Turn the Resonance up a little more if you're not hearing it.

What you will find is that some Harmonics (when picked out) will be so loud, it will cause the sound to really distort (horrible Clipping).

If you look at the Bottom Right of the Filter Type screen, you will see the Gain Value. By default, it is set to value = 230 which is quite high, causing High Resonance to distort. You need to turn Gain down to stop the distortion or "squawking".

So, the digital modelled Filters are behaving like Analogue Filters, but they're not perfect models... they are approximations at best.

However they are "good enough" to mimic typical classic "Filter Sweep" sounds so loved on Analogue Synths.

 
Posted : 19/11/2022 5:50 am
Antony
Posts: 745
Prominent Member
Topic starter
 

The other selection of Filter Models can be used for different effects.

The dual Filters can be used to create Vowel (or Voice Box, Vocoder) like sounds for example.

However, FEG and LFOs still only control whatever Filter Aspect is listed as "Cutoff". You can check the Reference and Parameter Manuals for more detail.

The FEGs and LFOs cannot be assigned to any other Filter Parameter other than Cutoff.

You may be able to Modulate other Filter Parameters using Assign Control and/or Motion Sequencer.

You can build "Envelope Type" curves in the Motion Sequencer (say to affect HPF Cutoff in HPF+LPF Type), but it's a bit clunky, you will need to persevere.

 
Posted : 19/11/2022 5:59 am
Antony
Posts: 745
Prominent Member
Topic starter
 

OK... found the original Thread

https://www.yamahasynth.com/forum/how-to-make-cutoff-automatization

No need to reread or rehash it.

The disputed point was that the Filter Envelope Generator does not modulate (automate) the Filter Cutoff Frequency (Cutoff Parameter Value 0-255).

Instead, it was suggested the FEG changes "the amount of Filtering" to loosely resemble a real Analogue Filter Envelope. I never really understood what was meant by this, hence my return to investigate.

Anyways, the FEG absolutely does Modulate Filter Cutoff, and nothing else (Resonance, Gain etc). Therefore this is identical behaviour to that in Analogue Subtractive Synths.

The problem the OP and others may have been experiencing is likely due to not setting initial cutoff and FEG parameters appropriately as outlined in my previous posts above.

 
Posted : 20/11/2022 12:09 am
Posts: 1715
Noble Member
 

Imagine the filter's graph being moved down, in the Y.

That's what's happening. It says it on the tin: DEPTH!

I knew I was going to regret not properly illustrating the inadvertent cutoff change occurs as a result of decreasing the depth of a sloped filter. And being more specific, in that context, about the individual note characteristics of the desired result.

Further, it's impossible to move the Cutoff LEFT, and maintain the depth, and especially impossible to increase the depth and reduce (move left) the cutoff - of a played note. This is the desired result, this is what makes individual note filter animations/modulations sound excellent on the synths that provide this.

And the other wonderful trick: increasing resonance as cutoff is reduced (as Cutoff value goes left, Resonance tracks with and increases in amount - up).

This is why I strongly encouraged trying out the free Vital Synth plugin, because it not only does this better than most all other synths ever, it also visibly shows this occurring in realtime, and you can modulate the extents and rates of this behaviour in ways that are enormous fun, such that you don't need to do a lot of knob and slider twiddling to get incredible dynamics of filter movements (depth, cutoff and resonance).

 
Posted : 21/11/2022 3:14 am
Posts: 1715
Noble Member
 

IOW: what you've done is found and begun exploiting the ranges within which values used cause sufficient inadvertent reduction and/or increase of the filter's range of influence that you can notice it. This is the best way to use these filters, but it's extremely limited, much more so than having direct, actual control over the Cutoff (and Resonance) during the Envelope Duration (time) of a Generated Envelope.

Again... this is why I've suggested the easiest and most powerful possible remedy, adding Cutoff and Resonance Envelope Generators, which have their own envelopes, and can therefore act in conjunction with the FEG, but at their own rates, thereby allowing all sorts of wonderful movements.

The existing FEG is kind of really only a Filter Depth Envelope Generator - with some range of inadvertent influence over the overall effect of a curved filter.. such that you can detect the reduction/increase of range filtered thanks to the slopes of the filter crossing at different points as the depth changes.

Note, if you use a Filter shape with a much gentler slope, you'll get more of the "feel" you''re looking for, at the expense of less initial filtering.

 
Posted : 21/11/2022 3:41 am
Posts: 1715
Noble Member
 

[quotePost id=119398]
And I think with your parameter envelope idea you could even set a rate for each parameter differently from the others.

[/quotePost]

Yes, that's the power secret of CEG and REG (cutoff and resonance envelopes), Yamaha could leapfrog into the present, and perhaps a little beyond it, when combined with Motion Sequencer and SuperKnob etc.

Here's what I should have shown originally, with regards inadvertent cutoff "movement":

 
Posted : 21/11/2022 5:14 am
Antony
Posts: 745
Prominent Member
Topic starter
 

I think you are confusing the Filter Diagrams (on the Filter Type Page) with Spectral Analysis Graphs... or something.

The Filter Diagrams, are just that... Diagrams, Visual Aids, Pictorial Schematics. They are not accurate Graphs.

The FEG Depth does not move the Filter Diagram up and down the Y Axis. The FEG Depth governs the Amount of the Envelope applied to the Filter Cutoff.

With any Envelope settings, FEG Depth = 0 has no effect on the Filter. As you increase FEG Depth, Filter Cutoff movement (Modulation) becomes increasingly dramatic/severe.

The FEG Depth is basically the same thing as an LFO Depth, or, Modulation Depth.

The FEG moves the "Cutoff" left and right on the Filter Diagram.

It would look the same as you selecting the Cutoff Parameter Value, and then spinning the Encoder Wheel left and right. Except the FEG can do it a lot faster than your hands.

Probably the best way to visualise this and confirm it to yourself is do the following:-

  • INIT AWM2
    Choose Element "OB Saw"
    Go to Filter Type Page
    Choose LPF 24A
    Set Cutoff/Key = 0%
    Set Cutoff Max 255
    Go to FEG Page
    Set FEG Depth = 0 (so Envelope has no effect).
    Go Back to Filter Type Page
    While playing and holding C2 adjust the Cutoff Value down with Encoder Wheel.
    Keep going down until the Note decays to Silence.
    Nudge the Cutoff up until you hear the note become audible.
    Cutoff will be approx = 79-80.
    The Filter Cutoff Frequency is now Set at around 65Hz (i.e. C2).
    Set Filter Gain = 199
    Set Resonance = 79.
    Now increase Filter Cutoff slowly with Encoder.
    You will hear the High Resonance sweep through the Saw Wave's harmonics as you increase the Cutoff.
    Sweep up and down a bit, get a feel for it.

This sweeping of the Cutoff Frequency is Exactly what the FEG is automating.

 
Posted : 21/11/2022 6:24 am
Antony
Posts: 745
Prominent Member
Topic starter
 

A few excellent videos that really helped me get to grips with this.

https://youtu.be/qwrUUZWXOh4

https://youtu.be/pX4xVLbLhkU

https://youtu.be/otQ5lioCY_g

 
Posted : 21/11/2022 6:44 am
Posts: 1715
Noble Member
 

@Antony, whilst the FEG (a depth controlling envelope) gives somewhat of an approximation of what happens when a controller moves the cutoff, that's a byproduct of what it's actually doing to the extent (depth) of application of the filter.

It's going to take some time playing with a synth that actually gives you control over the cutoff, dynamically, per sounded note during the duration of that sounded note, before you can fully appreciate the difference.

I strongly suggest playing with the free Vital Synth, as this is the most visual and extreme and free example of this possible, and has some very extreme filters so you can fully appreciate this movement:

https://vital.audio

you'll gain a new appreciation for subtractive synths, too... as it's exactly this initial very fast move to the right and then eased out move left of a filter's cutoff that makes all the very best subtractive synth sounds... the very thing that's impossible with the MODX/Montage and why serious sound designers give it and all other AWM2 powered Yamaha synths a complete miss.

They don't have anything against Yamaha, they just don't like the lack of filter control.

Being able to rapidly move the cutoff frequency left and right, in isolation from Depth/Gain/Drive/Mix, is the most significant aspect of subtractive synth sound design since the late 1980's.

The sweeter the sounding the filter, the more poignant this particular kind of movement becomes, especially if it's possible to accelerate and then slow these movements.

 
Posted : 21/11/2022 7:25 am
Posts: 1715
Noble Member
 

A word of warning, once you can hear good filters and good, well controlled cutoff and resonance animation, you'll become a full cork sniffer for both filters and their control.

And may start finding yourself constantly trawling through second hand markets for older analogue gear. Which is far better than finding yourself addicted to modular gear. But still a nasty addiction, nonetheless.

 
Posted : 21/11/2022 7:40 am
Antony
Posts: 745
Prominent Member
Topic starter
 

@Andrew...

I'll accept that the MODX Filter Modelling is not on par with (not as sweet as) dedicated VA Synths.

I respectfully disagree that the MODX Filter behaviour and the FEGs impact on the Filter Cutoff is doing anything different compared to other Analogue or VA Synths.

I can hear it.

I posted test notes and video tutorials. You'll just have to work through them until you figure it out.

 
Posted : 21/11/2022 7:46 am
Posts: 1715
Noble Member
 

[quotePost id=119407]@Andrew...

I'll accept that the MODX Filter Modelling is not on par with (not as sweet as) dedicated VA Synths.

I respectfully disagree that the MODX Filter behaviour and the FEGs impact on the Filter Cutoff is doing anything different compared to other Analogue or VA Synths.

I can hear it.

I posted test notes and video tutorials. You'll just have to work through them until you figure it out. [/quotePost]

Miles ahead of you on this.

Am waiting for you to catch up.

 
Posted : 21/11/2022 7:51 am
Antony
Posts: 745
Prominent Member
Topic starter
 

You'll have a long wait. :p

 
Posted : 21/11/2022 7:55 am
Posts: 1715
Noble Member
 

[quotePost id=119409]You'll have a long wait. :p [/quotePost]

I know...

... it's on a list, along with waiting for Darryl to be able to hear how crusty the onboard reverbs are, and Bill to figure out this isn't Stackoverflow 😉

...and your review of the Hydrasynth!

 
Posted : 21/11/2022 8:58 am
Page 1 / 2
Share:

© 2024 Yamaha Corporation of America and Yamaha Corporation. All rights reserved.    Terms of Use | Privacy Policy | Contact Us