emusonacid! legacy

i like to use resonant sounds as contrast to dry sounds in a tune but i can never quite figure out at which value to have the resonance. also, if i want to play it melodically from the keyboard i would need to know how to map it to key tracking; what start value to use and how strong the modulation should be. i've tried doing it by ear but haven't really gotten anywhere.

so, if i want to figure out which values of resonance correspond to which values of pitch, how would i go about doing that?

edit: i suppose i should mention that i'm talking about the vocal filters

Are what you are looking for a list of the hertz value for tones using western equal tuning?

I don't know if it's useful in this situation, but you might try looking at the patch-cord amount table in one of the Proteus2000 manuals (since they're all freely available as PDFs). They didn't bother to include this table in the EOS manual. In the P2K manual, it's right after the diagrams of velocity curves. It lists amounts for (pretty near) semi-tone steps. But I don't know how well that translates to filters rather than pitches.

I don't really have any examples to pull reference from, the nearest being some in the old Morpheus... but many of the vocal filters in the Morpheus have three values (morph, key-track, transform 2), while the ones in EOS have only two (freq, res) which roughly translate to (morph, key-track), since "key-track" doesn't always really mean "key-tracking", and gets used for other sorts of parameters in some filters. I can't pretend to completely understand E-mu's Z-plane filters. Although it would have meant a TON of parameters to keep track of, I sometimes wish E-mu had given full control of the six parametric filters that make up a Z-plane filter, or more graphical representations of how the Z-plane filters behave across the different sides of the cube (or square, for 2 parameter filters). It would make it a lot easier to visualize the result you want and how it relates to the filter responses.

It really is pretty much trial and error... patch key->res (since filter res is only read at note-on) on the assumption that "res" on the EOS filters maps to "key-track" in the Morpheus filters. The amounts used in the Morpheus patches vary from +15 to +127 (but seem to average around +63/64, or doubling steps... 0, 15, 32, 64, 127...), with no initial offset. But that might not translate over to the EOS filters.

thanks for the extensive info. i'll try to track those PDFs and see what i can find in there. however, reading your response, i had an idea that might help figure it out. i figured i'll use the lowest and the highest settings and see if i can find which notes they're at and then if that works out i can just divide it by 256 (i think that's the range).

i'll post more info when i've tried this out.


thanks again, guys.


on second thought, i don't really understand that last paragraph regarding key tracking. are you saying that key track in EOS has different values depending on what parameter it's mapped to? i probably misunderstood that but come to think of it i also don't understand how strong key track is. for instance, if i were to use the HPF and map key track to it (at highest value), what range would it cover if i set the HPF at the lowest value initially? could someone please shed some light on this?

No, when I say "key-track" in the last paragraph, I'm talking about what the Morpheus architecture called the second filter parameter destination. As Morpheus used "morph" for the first (single real-time) parameter, and "key-track" for the second parameter (only read at key-on), the names used on the EOS filters were "frequency cut-off" for the first parameter (variable in real-time) and "resonance" for the second (only read at key-on) parameter. And "key-track" as a filter parameter destination is different from actual key tracking note numbers as a modulation source.

I think you're idea about comparing the highest and lowest values is a good direction to go in. I think it might get a bit tricky depending on whether you're calculating from a piano-style 88 note keyboard or from the full 127 note MIDI note number range.

Key-tracking can essentially provide a number value between 0 and 127. This could be mapped onto a filter parameter either as an absolute value or added/subtracted from a starting value. If you started the filter cutoff parameter at 0 on a low pass filter, for example, then key track would run from fully closed down (0) to wide open cut-off (127), increasing one "value" more open with each semitone up the keyboard. I think the parameters generally only have a 128 width value, either as -127 to 0, or 0 to 127, or -63 to +64 (depending on the type of mapping in the cord setup... some parameters can only be mapped one way... + mapped to 0 through 127, ~ mapped -63 to +64, or < mapped -127 to 0) The high pass filter example would work similarly (+ mapped from fully open at 0 to closed down at 127, so that higher notes raise the cut-off frequency... or < mapped from closed down at 0 to fully open at 127)

If you have the EOS 4.0+ manual, page 262 has a useful list of which modulation sources can be mapped in different directions (+, ~, <) and which modulation destinations are read only at note-on (')

Resonance is typically an artifact that occures when the poles of a filter become too close to each other. Basicually they start to squeal. In absyth, I can emulate resonance by going into the spectrum of a user defined wave form and push the highest end of the spectrum up. That pitch is going to be super high, much above what you would be playing. The spectrum in absynth has harmonics at intervals of octaves. So I have upto 64 different notes I can play by tweaking the harmonics.

I guess one thing I could do is take a sound, like a square wave and apply Q (resonance) to it in absynth and then try to recreate that sound using two oscillators, one with a dry square and one using just sine waves and thus isolate what frequency it is. See, I could load up a sine, and by selecting the proper harmonic, maintain the fundemental (though it wont really be there) and mimic the sound. Then if this works, I could sample something basic like a square and put it into the sampler, add Q and then record it into cubase and then try to mimic the sound again using absynth again and thus be able to tell what range the Q is effecting in the body of the sound. It would only take 15 minutes or so I would figure if things work correctly.

But there is something problematic with this. Resonance is not directly effected by the pitch you play but by the centre frequency of the filter. For example. If I have a completely open filter in absynth and ramp up the Q and then start to modulate the centre frequency, the sound will start to squeal almost instantly and will continue to squeal all the way down to a point where the filter completely mutes the sound that is being filtered. This would occur regardless of what note that is being played.

The other thing that could be come problematic as well. The sound itself will change by applying Q because the shape of the filter is being changed. It's becoming narrow and as such a dry version running lpf open with a Q of zero will sound "normal" but with an lpf running open with the Q up all the way may squeal or just sound like a slightly high passed version of it as the filter is "belling" at the left and the right of the of the centre freq. Of course depending on the filters architecture. This is why when you up the Q of a sound, depending on the centre freq, you get an increase in bottom and an increase in upper harmonic content with a deep in the middle of it all to a degree. And of course at a specific range the squealing will stop as the upper part of the bell will be below the nasty freqs of the 1 or 2k range.

Of course, I am sure you already know this stuff though. Oh... and as to what frequencies are going to present in the resonance aspect of a sound are going to be near transient. 12000 or higher. Maybe bottoming out around 1 or 2k since it often times hurts my ears, but this is going to be extremely high.

What could occur here though is if you find this particular freq you could run your original sound with a sine wave or a set of sine waves at those particular harmonics that would represent a desired harmonic/melodic interval in association with the fundemental frequency of the original dry sound. You could in a sort of way fake it, essentially.

It's up to you. Give a shout and I think I can sort you out if I am underingstanding you post correctly. Hope this helps.

snh