I was wondering if anyone can provide information, or point me to a good article, about certain numeric parameters that appear when editing a voice on the motif. When no unit is indicated, I am sometimes unsure of how to interpret the numbers.

When no specific unit is indicated (like “Hz” Hertz, or “dB” deciBel) you can assume it not applicable. You can be sure, in most cases, the larger the value means more, higher, while lower values tend to mean less, lower. There are exceptions, however.

Like here is one question about the cut off parameter in element edit: is there a reason why isn't it indicated in Hz ? For example, is the cut off frequency relative to the pitch of the note being played?

Yes, to both questions. There is nothing that says a certain frequency is going to be produced by a given key. It can vary from program to program.

In general, when working cutoff frequency in programming it is done “by ear”... the MIDI Keyboard which includes some 128 Notes, C-2 through G8, would stretch beyond our audible musical range.

The Filter Types, which are 1 to 4 pole Low Pass, are 6, 12, 18 or 24dB per octave in their roll off — used to attenuate frequencies above the Cutoff Frequency. If you were to place the Cutoff Frequency at A440 a LPF24 Filter would render A880 24dB lower in volume.

The way you would use the filter is you would be raising and lowering the cutoff frequency until the balance of harmonics feels/sounds right. It is not an academic endeavor as much as it is a ‘when it feels right it is right’ thing.

If you want a reference point... when Cutoff Frequency is set to 127 on the scale of 0 to 255, the center frequency is A440.

You can experiment by assigning a BEF6 (Band Eliminate Filter 6dB per octave) this Filter can notch out a narrow frequency band. Reproduce the pitch A440 - it does not matter which note is used to reproduce A440, when A440 is sounding a Cutoff setting of 127 will reduce the sound output specifically. One click toward 126 or 128 you will hear are louder than 127.

You will see Hertz when working with the EQs, you will see dB when those are actually used to measure increase or decrease.
In general, when you see “+0” as the default, you are applying an offset to the currently stored value. You can add to the current setting or subtract from the current setting.

What about the time parameters with respect to envelopes? Is the length of the overall envelope dependent on the pitch or the particular oscillator?

Again, “musician time” is measured in measures and beats, not seconds and milliseconds. Clock timing values are mostly meaningless to musicians — they simply are not used. To be sure, you can make it an academic endeavor but mostly it is done by ear, for several reasons. Envelopes can be flexible and can adjust according to how much velocity is used to trigger the note, and can be scaled across the keyboard. Both of these factors make the use of seconds and milliseconds of less use. Again it is a feel thing.

Velocity can be used to make an envelope faster or slower, shorter or longer... in general, the more energy used in creating a sound the longer the sound will last. Key “Scaling” can be used so that the speed of envelope changes as you go up or down the keyboard... in general, the lower the pitch the slower and longer the envelope, and the higher you go the quicker and shorter the envelope. Think of the piano strings (the lower strings ring out longer than the high strings if hit with equal velocity… (scaling).

Fixing envelopes to time (secs and milliseconds) would not be particularly musical... and may not improve the task of setting envelopes, at all, as the time values are not intuitively known by players. It, again, is best a “feel” thing.

This is not to say you couldn’t use such time values but they are certainly not necessary to recreate the envelopes of musical instruments.

Also the time values (increments) are not linear... not at all. Rather, they are weighted in such a way that there is more density in areas known to be useful (shorter time values are denser). For example, the amount of time between a setting of 40 and 41 is much much much smaller than the time represented by values 120 and 121. Why? … because you need more resolution in the shorter time variations than the longer time variations. You don’t need a Release Time than last 90 seconds and another that lasts 90.5 seconds.... but you do need much higher resolution in the shorter time values.

How do you know what you’ll need? Again, settings are made “by ear”.

Hope that helps. Thanks for the question.

Thanks Bad Mister, that helps a lot, exactly what I was looking for.

I always thought that filter types meant low pass, hi pass, band pass, and so on. But when you talk about 1 to 4 pole and the dB per octave, this is something else, right? I think this pole stuff may be a key piece of the puzzle.

I see what you mean about the time values not being linear. A decay time of 64 is definitely not half as long as a decay time of 127.

Point well taken about using my ear.

Thanks again,
Dave

I was using the LPF24 as an example of the musically useful Low Pass Filter. Low Pass Filter 24dB/oct is one of 18 Filter Types: LPF24D, LPF24A, LPF18, LPF18s, LPF12, LPF6, HPF24D, HPF12, BPF12D, BPFw, BPF6, BEF12, BEF6, Dual LPF, Dual HPF, Dual BPF, Dual BEF, LPF12+BPF6, thru

The Low Pass Filter (LPF) is the most used filter types in acoustic instrument emulation... Reason: the harder that the musician hammers, strikes, plucks, blows or bows a musical instrument, the richer it becomes in harmonic content. This bit of earthly acoustical physics means that softer sounds have more low frequency content and as you increase the energy used to start the musical vibration, the more high frequency content is heard - for the most part. LPFs come in different intensities—which determines how severe the frequency roll-off is. As velocity increases the LPF is typically programmed to respond by allowing/revealing more high frequencies. This sounds natural to our ears. And when too much of this harmonic content change is missing, our ear/brain is not fooled... and we start questioning what we are listening to...

High Pass Filters (HPF), allow high frequencies to pass and initially block the lows... assigning cutoff movement to a HPF gives us a bit of the unnatural ... the result can be startling and profound — science fiction, synthy. Because it is the opposite to what we are used to..

On your Motif XF call up a Voice called “Long HiPa”
The following is from a tutorial for MONTAGE dealing with on High Pass Filters..........:
If you are listening in stereo, (and we sure hope you are) the first thing you should notice is that the filter movement is such that the high frequencies are heard first and as time goes on we hear more low frequencies being allowed through. We are hearing the result of the Amplitude Envelope Generator (loudness) and Filter Envelope Generator (timbre change). And there seem to be several distinct areas moving within the stereo field. You can feel it come from the center and split down both sides of the stereo field.

The "HiPa" spelling is revealed as a hint that this is using "High Pass" Filtering. A High Pass Filter is one that initially allows high frequencies to pass and blocks low frequencies. Why this sounds "synthy" and "atypical" to us is because in the emulative world of sound designing, it is the Low Pass Filter that does the lion's share of work when attempting to mimic instrument behavior. This is because, in general, the harder you hammer, strike, pluck, blow or bow a musical instrument, the more high frequencies it produces. Therefore, an LPF made sensitive to Key-on velocity is most often used to create this effect - and the harder you play the brighter the sound. Velocity is used to raise the Cutoff Frequency of an LPF, which means the timbre gets brighter. Here however, we are using the Envelope Generator to create movement of the Cutoff Frequency and the movement is in the opposite direction. An "envelope" describes movement over time.

The HPF24D translates to High Pass Filter 24dB per octave, Digital. For every octave you go below the Cutoff Frequency the signal will be down 24dB. This means that if you play an A440 and it is 0dB on the meter, the "A" at A220 is -24dB relative to the A440. Frequencies below the Cutoff Frequency are reduced by 24dB for each Octave you go down. This is consider a steep curve (also called a 4-pole Filter).

...taken from the MONTAGE tutorial: MONTAGifying Motif XF Performance: “Creepin’ Worm”

The tutorial goes on to discuss the Band Pass and Band Eliminate Types as well... and because the MONTAGE is based on the XF, and like having eight Motif XFs, you should be able to follow everything that is discussed about how the filter parameters work and are defined. Let us know.

Bad Mister wrote:

"If you want a reference point... when Cutoff Frequency is set to 127 on the scale of 0 to 255, the center frequency is A440.

You can experiment by assigning a BEF6 (Band Eliminate Filter 6dB per octave) this Filter can notch out a narrow frequency band. Reproduce the pitch A440 - it does not matter which note is used to reproduce A440, when A440 is sounding a Cutoff setting of 127 will reduce the sound output specifically. One click toward 126 or 128 you will hear are louder than 127."

Bad Mister,
I've been experimenting with this, and it doesn't work the way I expected. I don't hear a difference in volume between 127 and 128. I must be misunderstanding something.

I understand"Center frequency" to mean the frequency around which the BEF is set: the frequency corresponding with very bottom of the notch. So, if the cut off frequency is set to 127 and if I play A 440 on the keyboard, the filter should be cutting out the fundamental.
But I don't think that's what I'm hearing. It sounds like it's cutting out a much higher frequency, like an overtone.

Also, am I correct that with respect to band filters (whether they be elimination or pass), there is some functional overlap between the resonance level and pole settings? They both seem to control how steep a drop there is from the cut off point.

Adding this a few hours later:
I had an idea to try the filters on a sine wave, since I believe they have no overtones, only the fundamental. I'm figuring that all the filter could do to a sin wave is reduce the volume. That way I can tell what frequency the filter is affecting. So here is what I found:
Using Bef12
Cut off at zero
A 440 goes silent

But…!
If I mess around with the Cut off key follow parameter, I can make other notes go silent.
A setting of 46 makes C 1 go silent. And in the same vein respectively:
46: C 1
58: C 2
5: C 4
18: C 5
23 : C 6
25: C 7
No number can make C 3 go silent.

Hmmm. I don't see the pattern.

Sorry.
Like a lot of people these days, I have a lot of time on my hands.

My best to everyone,
Dave