One note per timbre (like old analog)?

acemonvw · Post by **acemonvw** » Mon Jun 13, 2011 1:15 am

So... I have a strange question and I tried searching and realized I don't even really know how to search this, so here goes:

My old analogs, Oberheim matrix 1000 and roland juno 106 have 6 voice chips (well, I think the matrix has 2 oscillators per voice). When you press a note, voice chip 1 is activated and sounds out based on the parameters you've set. When you lay down another note, voice chip 2 is activated... and so on. I learned this the hard way from buying an oberheim matrix 1000 with bad voice chips and had to figure out which ones were good/bad (but paying only $50 for the unit seemed worth it).

Anyway, I got to thinking, can one model the way these synthesizers work on the radias? If we think about each timbre as a voice chip, could you theoretically play one note and it causes timbre one to play, then when you press down another note it causes timbre 2 to play, then so on to timbre 3 and 4... This would give you only 4 notes of polyphony as when you first play a note only timbre 1 would sound.

It seems to me that playing multiple notes of one timbre starts sounding really thin, maybe it's that the notes are cancelling each other out and I haven't quite figured out how to master that... but, it seems like it would be interesting to see if what I suggested above could be done.

Thanks guys!

Morshu · Post by **Morshu** » Sun Oct 09, 2011 2:28 pm

asign a slight out of tune ness to each timbre and use a very tiny amount of s&h lfo on pitch in each timbre, vary it until it sounds natural. add a little detune if you want, vary that too. you'll end up with something that sounds more natural and analog

X-Trade · Post by **X-Trade** » Sun Oct 09, 2011 5:33 pm

Radias and R3 have the 'analog detune' parameter for this. it applies a random pitch to each note as it is played. so the S&H LFO isn't necessary there, not for pitch anyway but you could use it for filter cutoff for example.

There isn't a way specifically to get each timbre to play a different note of the same chord, like I think you're asking, except for perhaps sequencing it yourself or using multiple MIDI keyboards aligned underneath one another.

If you're only play two notes - duophonically - then you could assign two timbres to mono, one set to highest note priority, the other set to lowest. although then you will get both timbres playing when you hit one note. could be interesting. You could then further layer a full poly voice on top of that.

acemonvw · Post by **acemonvw** » Sun Oct 09, 2011 5:39 pm

Interesting, thanks guys.

Yeah, I was aware of the analog detune, but I find that when multiple notes are played it starts sounding distorted and individual notes don't sound very good together (when using saw waves at least). My Oberheim just has a really cool sound and I know that they work off of one voice chip per note. I'll give the duophonic sound a try, seems interesting, and more or less what I was looking for.

When I route the radias through the spectralis' filterbank, the thing sounds amazing when playing monophonic... the second I add more voices it sounds like crap. I can't tell if its the fault of the filterbank or the radias, but like I said earlier, I feel like polyphony ends up getting distorted... Anyway, that's off topic.

Thanks again for the responses!

MarPabl · Post by **MarPabl** » Sun Oct 09, 2011 6:22 pm

Guys, do you think the Mod Sequencer could be used to alternatively mute each one of the 4 timbres ?

Morshu · Post by **Morshu** » Sun Oct 09, 2011 6:46 pm

you could but it wouldn't exactly do what you want, it'd be close.

X-Trade · Post by **X-Trade** » Sun Oct 09, 2011 6:46 pm

If it gets distorted when you play more than one note, you may find it is simply a volume thing.

Try a compressor or limiter in the master FX slot.

CharlesFerraro · Post by **CharlesFerraro** » Sun Oct 09, 2011 9:05 pm

X-Trade wrote: There isn't a way specifically to get each timbre to play a different note of the same chord, like I think you're asking, except for perhaps sequencing it yourself or using multiple MIDI keyboards aligned underneath one another.

I think the closest you're going to get is assigning one mono timbre keysplit across the four octaves.

So make your mono patch in the editor. Enable all timbres and copy your patch (alt, drag) onto the other timbres.
Make timbre one trigger from B2 and below
Timbre 2 from C3 to B3
Timbre 3 from C4 to B4
and Timbre 4 from C5 and above

You may want to change the note position of these splits to help you hit chords.
Note if you transpose the keyboard from the face the splits travel with the transposition.

acemonvw · Post by **acemonvw** » Thu Oct 13, 2011 2:46 pm

Thanks guys.

I think the splitting the notes is the best way to test how the radias might sound if it was a four voice-chip type analog synth. It's probably not a very useful method for using the synth, but I'm going to try it out and see if it sounds any different than just playing one timbre with 4 polyphony.

I tried the duophonic method described by X-trade, it worked, but I'm not sure I really generated a better sound, but it was a good idea!

MarPabl, I think that's worth a shot too.

Again, thanks for all the suggestions!

zalo · Post by **zalo** » Thu Oct 13, 2011 4:35 pm

You could set each timbre to a separate midi channel. Then use something like Max/MSP to step through the channels with each new note. That shouldn't be very hard to do.

Sans Nom · Post by **Sans Nom** » Wed Oct 19, 2011 11:56 am

Did you try the voice card trick of bobborries?

http://www.korgforums.com/forum/phpBB2/ ... 4&start=30

This is a very elegant way to reproduce the spread in tuning, panning, etc. of different voice cards. Forget the pitch destination of LFO 1, use analog detune instead (this way you get 2 random generators instead of 1). Also, try enveloppe parameters as destinations and add LFO 2 if possible.

acemonvw · Post by **acemonvw** » Wed Oct 19, 2011 3:36 pm

I had run through that back in May, but I don't remember what I thought of the end result. I'll give it another try, thanks for the reminder!

Re-Member · Post by **Re-Member** » Thu Oct 20, 2011 1:45 am

Sans Nom wrote:Did you try the voice card trick of bobborries?

http://www.korgforums.com/forum/phpBB2/ ... 4&start=30

This is a very elegant way to reproduce the spread in tuning, panning, etc. of different voice cards. Forget the pitch destination of LFO 1, use analog detune instead (this way you get 2 random generators instead of 1). Also, try enveloppe parameters as destinations and add LFO 2 if possible.

The Analog detune feature adjusts the tuning of both oscillators simultaneously and identically as if they were both synced together. To my ears, it doesn't sound very analog at all. The method posted by bobborries sounds more lively since the two oscillators end up being in a slightly different pitch from each other with each note, creating a subtle chorus effect. When people say analog is "warmer" sounding than digital, this is one of those factors.

Sans Nom · Post by **Sans Nom** » Thu Oct 20, 2011 11:04 am

Well the post of bobborries says "pitch" as a destination, which is the pitch of the whole timbre, which is equal to the analog detune feature from what I understand. In this case it's better to use analog detune, because its random generator is independant of LFO1.
Actually, the problem of using LFO1 as the unique source is that the destinations will be modulated in the same way, for example pitch will always be lowered while filter cutoff will be increased and panning sent to left. And this is something that can be noticed and doesn't feel natural and "analog". The best is to use LFO 2 also and use the mod sequencer or the effects for the classic modulations of the patch.

Re-Member · Post by **Re-Member** » Thu Oct 20, 2011 2:33 pm

Sans Nom wrote:Well the post of bobborries says "pitch" as a destination, which is the pitch of the whole timbre, which is equal to the analog detune feature from what I understand. In this case it's better to use analog detune, because its random generator is independant of LFO1.
Actually, the problem of using LFO1 as the unique source is that the destinations will be modulated in the same way, for example pitch will always be lowered while filter cutoff will be increased and panning sent to left. And this is something that can be noticed and doesn't feel natural and "analog". The best is to use LFO 2 also and use the mod sequencer or the effects for the classic modulations of the patch.

Whoops, didn't catch that. When I make patches on my other synths, I typically route an SH or RANDOM LFO to mod the pitch of a single oscillator, but on my R3, this actually is "OSC2Tune" to create the effect.