PSYN-tology 3 [Reso Sweep]

This is the kind of sound got me into synths in the first place. The Reso Sweep patch reminds me of my Poly-800 (still with me, unfortunately all but retired). Let's look under the hood ...

Have you noticed how many patches only utilize one oscillator? Having four, it seems that someone could come up with some great patches (a la Crystal) with four simultaneous but unique sub-patches controlled by the four EGs (through the fifth EG A). Unfortunately only one (really two) filter section. But I digress. This fat sound starts with a sawtooth wave and is generated in part by the Unison section and a healthy (38%) Detune amount. If I understand the architecture correctly, Unison uses three oscillators per played note up tp the polyphony limit. In that case, I'd consider the choice of 8 note polyphony to be an odd one. [2 and 2/3 note chords?] Might want to increase polyphony here to avoid voice-stealing. It also appears that the patch author flirted with adding Portamento to this patch, but enabling it with the default Time gets lost in the slow envelope attack.

The filter choice is a 12dB slope Lowpass filter with a moderate amount of Resonance added. This increases the level of the frequencies at or around that set by the Cutoff Frequency. You can increase this close to self-oscillation. That's not a bad thing; you can hover near some squeaky sounds that'll pop with the right velocity touch. Notice that the settings for the second 24dB per octave filter appear to have been tweaked, but abandoned as perhaps unnecessary. Try enabling it and alternating between filters 1 & 2 enabled.

Examine the two EGs that are active. The setups are almost identical, with some subtle differences. The Amplitude envelope jacks the Sustain up but backs off on the Release time a bit with a slow Release Shape, when compared to EG 1. Both Attacks are slow, but the EG A comes in near 700ms, and EG that controls the filter is double that time. The Velocity knob here is the key; by setting it to 100%, you have the ability to 'play' the filter with your keyboard technique. [You do have a velocity-sensitive KB controller ... you don't? ... go back to PSYN-tology 0 [Disabled]].

EG 1 routes it's envelope to the Cutoff Frequency of the first filter section at 100% Depth in the Modulation 'matrix'. This makes for a very expressive patch with a minimal amount of effort. Try it with some automated 'stab' delays for a killer 'endless-ending' while your drummer shows that his extensive training in high-school marching bands have finally paid off.

OK, perfect patch, so now let's fug it up. The easiest mod is to go for a fast attack, while retaining much of the original character. Decrease the Attack time of both EG a & EG 1 to somewhere around 5 ms. While anything faster may or may not mess with the filter attack, there is an audible psycho-acoustic 'thump' that will occur with your amplitude envelope set to <5 ms. Use that as a rule of thumb. Now you've got a sound capable of a more rhythmic feel, but a bit too much release time for many circumstances. This time back the Release Time of just the EG 1 down to about 40ms. and listen to the note 'ring through'. Nice resonance, and it has the timing of my bass player when he takes his medication. Now drop the EG A Release down to zero, and comp it up like a spaced-out clavinent. Now this sounds good with some judicious tempo delays.

Reload the original patch by using Replace Synth and PSYN in Project5 and choose the Reso Sweep patch again. This time just cycle through the first filter options and audibly note the differences: Lowpass 12, Highpass 12, Bandpass 12 and Band Reject 12. Spend some time with this last one: drop the Resonance to zero and manually sweep the Cutoff Frequency with your mouse while playing a note. Hear how that chopping a section of the frequency spectrum out can actually give you a phasing/flanging type of sound? Add a dash of automation and you'll agree that it's a great time to be alive.

One last thing: After the last experiment above, load the patch above Reso Sweep in the new [F] patch bank [Sweet Sweep] to see how a Band Reject filter, some heavy Resonance, a second filter and a little variation introduced by altering the Amplitude Envelope and using EG 1 to modulate two filter parameters can completely change the sound of what is essentially the same patch.

Well, thanks for the good work. Had a little time tonight, so cranked out PSYN and started working my way through the lessons. Great stuff. One question (okay several, but they are all about filters!!) I looked this up in the help file, but didn't see it. How do you route to the different filters?.... can you route osc1 to filter 1 and osc 2 to filter 2? or can you route the groups (osc 1-2, osc 3-4) to the different filters? Also, is filter 2 a low pass 24 db roll off filter? When the filters are in serial mode do you have to route all of the output of one into the other or can you say route 50% of filter 1 to the outs and the other 50% into filter 2.

TIA

ORIGINAL: b rock

This is the kind of sound got me into synths in the first place. The Reso Sweep patch reminds me of my Poly-800 (still with me, unfortunately all but retired). Let's look under the hood ...

Have you noticed how many patches only utilize one oscillator? Having four, it seems that someone could come up with some great patches (a la Crystal) with four simultaneous but unique sub-patches controlled by the four EGs (through the fifth EG A). Yes, and I think another important tutorial should focus on how to decide what another waveform/osc. will contribute to your patch creation. I mean, if you start from scratch to try and build a patch, then it's difficult to imagine what impact another oscillator will have without knowing how these work. Follow me? whereas we are using subtractive synthesis for the most part, patch creation is an additive process for me. I'd like to see more on the issue of waveforms and their design factors. Unfortunately only one (really two) filter section. But I digress. This fat sound starts with a sawtooth wave and is generated in part by the Unison section and a healthy (38%) Detune amount. If I understand the architecture correctly, Unison uses three oscillators per played note up tp the polyphony limit. How does it determine what 3 notes? Is it a triad? What is voice stealing...and why is it so bad? I imagine it would give you a fatter sound with the added benefit of not hogging so much CPU, as Unison can be pretty intensive...no? In that case, I'd consider the choice of 8 note polyphony to be an odd one. [2 and 2/3 note chords?] Might want to increase polyphony here to avoid voice-stealing. It also appears that the patch author flirted with adding Portamento to this patch, but enabling it with the default Time gets lost in the slow envelope attack.

The filter choice is a 12dB slope Lowpass filter with a moderate amount of Resonance added. This increases the level of the frequencies at or around that set by the Cutoff Frequency. You can increase this close to self-oscillation. Um...insert technical linky thing here!? Self Oscillation I imagine is something to do with ring modulation? That's not a bad thing; you can hover near some squeaky sounds that'll pop with the right velocity touch. Notice that the settings for the second 24dB per octave filter appear to have been tweaked, but abandoned as perhaps unnecessary. Try enabling it and alternating between filters 1 & 2 enabled. Curiously, what exactly does the VELOCITY knob in Filter 1 do? Seems to affect the depth or attack of the "WOW".

Examine the two EGs that are active. If EGA modulates the volume or amplitude...what does EG1 modulate? The setups are almost identical, with some subtle differences. The Amplitude envelope jacks the Sustain up but backs off on the Release time a bit with a slow Release Shape, when compared to EG 1. Both Attacks are slow, but the EG A comes in near 700ms, and EG that controls the filter is double that time. The Velocity knob here is the key; by setting it to 100%, you have the ability to 'play' the filter with your keyboard technique. [You do have a velocity-sensitive KB controller ... you don't? ... go back to PSYN-tology 0 [Disabled]]. n00b aL3Rt!! Layman's terms: the harder you hit the note...the faster the WOW or filter occurs. Try lightly touching a note to hear the difference between that of a hard hit note! Dont forget us n00bs, B-Rock!

EG 1 routes it's envelope to the Cutoff Frequency of the first filter section at 100% Depth in the Modulation 'matrix'. yikes...potential disaster ahead! This makes for a very expressive patch with a minimal amount of effort. Try it with some automated 'stab' delays ...well...do I need to ask? WHAT is A stab DELAY? for a killer 'endless-ending' while your drummer shows that his extensive training in high-school marching bands have finally paid off.

OK, perfect patch, so now let's fug it up. The easiest mod is to go for a fast attack, while retaining much of the original character. Decrease the Attack time of both EG a & EG 1 to somewhere around 5 ms. While anything faster may or may not mess with the filter attack, there is an audible psycho-acoustic 'thump' that will occur with your amplitude envelope set to <5 ms. Use that as a rule of thumb. Now you've got a sound capable of a more rhythmic feel, but a bit too much release time for many circumstances. This time back the Release Time of just the EG 1 down to about 40ms. and listen to the note 'ring through'. Nice resonance, and it has the timing of my bass player when he takes his medication. Now drop the EG A Release down to zero, Hopefully you meant as low as it will go, cuz mine won't go under 7.0 ms and comp it up like a spaced-out clavinent. Now this sounds good with some judicious tempo delays.

Reload the original patch by using Replace Synth and PSYN in Project5 and choose the Reso Sweep patch again. This time just cycle through the first filter options and audibly note the differences: Lowpass 12, Highpass 12, Bandpass 12 and Band Reject 12. Spend some time with this last one: drop the Resonance to zero and manually sweep the Cutoff Frequency with your mouse while playing a note. Hear how that chopping a section of the frequency spectrum out can actually give you a phasing/flanging type of sound? Add a dash of automation and you'll agree that it's a great time to be alive. I agree man! Great Job!

One last thing: After the last experiment above, load the patch above Reso Sweep in the new [F] patch bank [Sweet Sweep] to see how a Band Reject filter, some heavy Resonance, a second filter and a little variation introduced by altering the Amplitude Envelope and using EG 1 to modulate two filter parameters can completely change the sound of what is essentially the same patch.

Man, this was a clear and concise tutorial ...and most importantly... A FUN walk-thru of this patch! Thanks B-Rock! Totally diggin' this. You have a lot of experience...I see you can start from scratch and build your own stuff...i can't wait to be at your level!

Yes, and I think another important tutorial should focus on how to decide what another waveform/osc. will contribute to your patch creation. I mean, if you start from scratch to try and build a patch, then it's difficult to imagine what impact another oscillator will have without knowing how these work. Follow me? whereas we are using subtractive synthesis for the most part, patch creation is an additive process for me. I'd like to see more on the issue of waveforms and their design factors.

Good idea. I started with the assumption that digging into the osillator waveforms might be redundant and repetitious, but now I see this as a glaring omission. This series was never meant to take on a linear form, and I hope that by the end of it I'll have hit upon all the major points. The waveform tutorial will be an easy one to whip out. Maybe I can find the perfect patch to demonstrate with, but I think this'll have to be from a 'scratch patch'.

How does it determine what 3 notes? Is it a triad? What is voice stealing...and why is it so bad? I imagine it would give you a fatter sound with the added benefit of not hogging so much CPU, as Unison can be pretty intensive...no?

When your reach the limit of polyphony in any synth, it uses a pre-determined or user-selectable algorithm to determine where the synth will get it's next voice to play. Most times the synth 'steals' the oldest (earliest in the time domain) Note On that's still pressed or sustaining (ringing through). In this example, I purposely set the polyphony to a low number (3) & used 'voice-stealing' to clean up the overabundance of long-release 'sustaining' notes that cluttered up the patch. I haven't done any tests regarding CPU usage with or without Unison, but it has to get those two extra notes from somewhere. Edit: A re-read tells me that you're aking what actual additional notes Unison adds. It's not a triad. There are three exact copies made for each Note On MIDI message. This is not to be confused with the 3 notes that I chose in Polyphony. I selected that number so that anyone playing a four-note chord or a lead riff/ bassline combination would experience the 'voice-stealing' firsthand. I can see where these coincidentally identical numbers might lead to confusion.

Um...insert technical linky thing here!? Self Oscillation I imagine is something to do with ring modulation?

Let me try to unearth a good one. Self-oscillation is a property of filters that happens this way: Think of the Resonance control as a feedback control on a delay or a regeneration control on a flanger. You continuously feed back a portion of the output signal back to the input, centered around the Cutoff Frequency. Eventually, that'll cause an oscillation. [Yes, your filter has now become an oscillator. Confused yet?]. It's the same thing with a microphone being too close to a speaker, or the 'runaway' echo effect, or your flanger whistling some strange 'harmonic'-like sounds. It's all oscillation. This not necessarily bad, but, uncontrolled feedback can cause some unintentional results (including deafness!). If you bring one of the above close to self-oscillation, but not quite there, you'll be rewarded with some wonderfully strange artifacts. Yet another brush to paint with. Ring modulation is quite a different process. Coming soon to a forum near you ...

Curiously, what exactly does the VELOCITY knob in Filter 1 do? Seems to affect the depth or attack of the "WOW".

Both filters have a Velocity control. Normally the control is mid-way at zero. Anywhere but centered gives you more and more influence over the filter with your playing stlye. Clockwise toward +100% affects the filter this way: the harder you hit the keys; the more the filter 'opens up'. It actually moves the setting of the Cutoff Frequency upwards with faster velocity. Counter-clockwise toward -100% inverts the process: the softer you hit, the more the filter opens up, or harder = CF knob going down (bassier sound). This is in addition to any LFO or EG that you hae controlling the Cutoff Frequency.

If EGA modulates the volume or amplitude...what does EG1 modulate?

All five EG's modulate both of the Dest 1 and 2 destinations in it's Modulation (matrix) section, if selected and by the amount set in the Depth there. The shape of the modulation control source sent is determined by the DADSSR controls above it. In addition, EG A also sends it's shape to an unseen VCA (voltage-controlled amplifier) or equivalent. This controls the overall volume/amplitude/loudness over time for the entire synth with it's DADSSR controls, and simultaneously sends this shape to whatever destinations are selected there.

n00b aL3Rt!! Layman's terms: the harder you hit the note...the faster the WOW or filter occurs. Try lightly touching a note to hear the difference between that of a hard hit note! Dont forget us n00bs, B-Rock!

Point taken, DA. But, as you've seen above, this isn't always true.

yikes...potential disaster ahead!

Modulation Matrix always refers to the Modulation section in any EG or LFP in PSYN.

well...do I need to ask? WHAT is A stab DELAY?

My term. Whenever I turn on a delay for only one phrase or note, and let the delays trail off, I call it a stab delay (or reverb). Like you're jamming on a riff, and only one note 'stabs' the delay for an accent effect. The reason I suggested that in this part of the tutorial is that the sweeping filter and a continuous pulsating delay would quickly descend into chaos. Hmm ... unless that's the effect that you're after!

Hopefully you meant as low as it will go, cuz mine won't go under 7.0 ms

Ooops! That one slipped past the proofreader. Tell you the truth, I rarely even notice the actual numbers in the settings; instead, I prefer to program by ear. I've been forced into paying attention in writing these tutorials, but I'm going in kicking & screaming (and evidently missing some!). Thanks for firing back.

LOL...B-rock..sorry to cause an avalanche...it can really get away on ya, eh?

No probs...thanks for explaining all that....I think I got my head around that...now..ON TO #4!

Sorry, only got to 3 yesterday. These are good enough that I wanted to be at the PSYN to actually make the tweaks. Sometimes I just read the articles...

Take Care

b rock

These are very good tutorials, and I, like others (?), am still on the first few - It only goes to show me that I need to slow down and focus on one thing at time. Unfortunately, my desire to learn all about this synth is greater than my ability to comprehend what's being shown to me.

I find that when I complete the lesson, I have to later go back to it and repeat the lesson -and each time I do, some things that were not even noticed on the previous runthrough -suddenly jump out at me. Then I run to the PSYN help file and read the about the area(s) you just discussed and strangely it sorta make some sense where before - I had no clue what they were talking about!

I am one of those consumers that have always bought and used the sounds as provided - with zero background in synths. These tutorials are opening the previously unapproachable secrets of how they make the sounds work.

While I am not foolish enough to think that i "get it" yet - I am experiencing moments where it almost comes into focus-sorta on the peripherial vision thing -

Anyway sorry for rambling, just wanted to let you know I appreciate these and with some serious plugging away on my end - your efforts will not be in vain..(I am just a bit slower than the average bear on this stuff.

If I understand the architecture correctly, Unison uses three oscillators per played note up to the polyphony limit ... How does it determine what 3 notes? ...I imagine it would give you a fatter sound with the added benefit of not hogging so much CPU, as Unison can be pretty intensive...no?

You really got me thinking on this one, DA. The statement that AFAIK the Unison function created three identical copies and detuned them got me pondering a better road to 'fattening'. While the Unison/Detune controls certainly have their place, it's a kind of a quick fix.

Now what if we took an alternate route by cloning two or more instances of PSYN instead? Lowering the Polyphony down to a few notes might keep the CPU hit manageable. Set up a relatively simple patch the way you want, and use Clone Synth to make some identical copies. Now adjust those copies' envelope, modulation and filter settings ever so slightly and trigger the copies all on the same MIDI channel with your KB.

What the advantage here? Now you have control over many more options than simple detuning, not the least of which include timbre variation and spatial positioning. This method results in a much fatter sound to my ears; much more like an ensemble (or not, if you like). It works around the single stereo out limitation to allow panning each clone's output. The timbre changes allow for subtlety (although you can do some of this in a single instance), you have 8-12 oscillators at your disposal, and you can run separate effects processes on the outputs.

No earth-shattering insight here; it's been done before. I just thought it needed to be restated for clarity and to kick-start another possibility. I haven't thoroughly investigated the CPU hit comparison, but preliminary results were promising. When you try panning a few PSYNs around with some subtle differences and effect variations, you might never reach for the Unison button again.

Yes, you can do that... but I trick I like even more!! is to add a different synth! Talk about fattening up a patch.

Good ideas..... great thread... looking forward to other lessons... should do from lesson 4 on this weekend!

Take Care

Hey man...cmon! No fair!

You can't add a tutorial to an EXISTING tutorial!

I am not amused! I'll never get through all these and remember to come back to this one again!

No problem... I definitely got my money's worth . I have learned a lot, but slowed down on the reading as my vacation is over. Am slowly working my way through however.

Take Care