Six voices not enough for pads?

[LPF83]

I find myself often frustrated when doing sound design on six voice synths when it comes to pads.  This is due to the envelope retriggering that occurs with lots of note changes.

To understand what I'm referring to, starting with the init patch, set both the amp and filter envelopes attack and release knobs long, maybe around two oclock to create a very basic pad sound.  Now hold two notes in the lower register with your left hand, three or four notes in the upper register with your right hand, and tap the right handed notes syncopated with the left to hear the envelopes become erratic, some of the notes suddenly have a fast attack.

I understand this is not how pads are usually played, but it rears its head often during sound design (and sometimes during playing) on synths with a low voice polyphony count.  Its so much better with the 10 voices of the Prophet 10 or the 16 voices of the Rev2... in fact to really get the same effect on the P10 I have to put in 5 voice mode.

I confess to not even fully understanding why this happens, since the net effect is an attack that seems higher than the maximum value defined by the envelope, but I know its not an exclusive problem to Sequential synths (even my Hydrasynth does it).

Is there some sort of workaround for this for designing pads on lower poly-count synths (P6, OB6, P5)?

[Sacred Synthesis]

I have the same problem even with the eight voices of the Prophet '08.  Patches with slow attacks and long delays are killers.  I find that lifting in between note strikes - whether chords or melodies - solves the problem.  It's with a smooth legato style that the problem is at its worst.

[LPF83]

I have the same problem even with the eight voices of the Prophet '08.  Patches with slow attacks and long delays are killers.  I find that lifting in between note strikes - whether chords or melodies - solves the problem.  It's with a smooth legato style that the problem is at its worst.

Indeed, even the 8 voices of the Hydrasynth isn't really enough to overcome the issue, although there are some retriggering options there which I have not fully explored.  I've noticed there are very few pads in the factory patches of the Prophet6 which are what we think of as a true analog pad sound, a lot of them use modulation to create movement that helps mask the issue.  There are a couple in there which seem not so immune to the issue, though I haven't dissected them to see exactly how they might be accomplishing it... I assume probably just optimizing/shortening the release to minimize the erratic note syndrome.

Its funny because I've had the P6 and OB-6 for quite a while now, and only recently decided to sit down and try to create good pads for the P6 before realizing how challenging it is.  I guess adding a P6 module to the KB and polychaining is a workable option, but it's an expensive one.

[Sacred Synthesis]

Yes, the only meaningful solution is increasing the number of voices.  It's fine once in a while to invent a work around.  But in the name of creating the precise type of pad you want to, and then playing it just as you need to, work arounds are no good.  Even playing in a detached way when using a long attack and release type pad (as I suggested above) is self-defeating, in that you're breaking with the intended smooth and dreamy musical effect.  So, this issue provides a strong rational for 10-16 voice instruments.  Some people consider this to be voice count overkill, but absolutely not.  You can always use fewer voices than you have available, but you can't use more than are available.

[LPF83]

I'd be content enough just to fully understand the phenomenon.  I can understand a more recent note needing to steal the voice of an earlier pressed one, and even needing to retrigger the envelope, but I haven't wrapped my head around why the note wouldn't be retriggered with the slow attack rather than what seems to be the maximum env amount.

[creativespiral]

The best workaround is probably to just keep envelope release timings a bit lower/quicker and use an ample amount of onboard, or external delay/reverb to take care of the release phase.  That's generally my strategy for big pads on a synth like P6, P5 or Polybrute.   

Regarding understanding why the envelopes steal like they do: there is actually another type of Retrigger mode that could be implemented, with firmware development -- "Always Retrigger from Start".   It was discussed a while back in the Pro 3 forum... and I added a graphic to show the differences:  https://forum.sequential.com/index.php/topic,4246.msg43715.html#msg43715



Optimally, you want to have both options available - standard retrigger, along with retrigger from start mode.  The "Always Retrigger from Start" option still has a voice stealing sound, but its nice to have as an alternate option.

Ultimately if there was a slew control / smoothing variable on retrigger from start, that would be the ultimate in flexibility.  This would be a great area of development for all Seqs synths... especially if its a shared library of code / IP that can be reused in future synths with similar envelope architecture.   That plus ADR contour shaping would really be great advances for envelopes / sound design.   

[LPF83]

The best workaround is probably to just keep envelope release timings a bit lower/quicker and use an ample amount of onboard, or external delay/reverb to take care of the release phase.  That's generally my strategy for big pads on a synth like P6, P5 or Polybrute.   

Regarding understanding why the envelopes steal like they do: there is actually another type of Retrigger mode that could be implemented, with firmware development -- "Always Retrigger from Start".   It was discussed a while back in the Pro 3 forum... and I added a graphic to show the differences:  https://forum.sequential.com/index.php/topic,4246.msg43715.html#msg43715



Optimally, you want to have both options available - standard retrigger, along with retrigger from start mode.  The "Always Retrigger from Start" option still has a voice stealing sound, but its nice to have as an alternate option.

Ultimately if there was a slew control / smoothing variable on retrigger from start, that would be the ultimate in flexibility.  This would be a great area of development for all Seqs synths... especially if its a shared library of code / IP that can be reused in future synths with similar envelope architecture.   That plus ADR contour shaping would really be great advances for envelopes / sound design.   

Yeah, I think Always Retrigger from Start and/or slew to take the edge off would be welcome fixes.  I wonder if there are any synths currently that do this?  And if not, is there some technical hurdle that makes implementing it challenging?

[thedigitalman]

I am very often frustrated with the 6 voice limit of my Prophet 6. As a compromise, I often use one of my moogs for the low end and save the 6 voices for fat chords.

[Sacred Synthesis]

I don't have a six-voice instrument.  I find eight voices to be restrictive enough.  But if I did have a six-voice synthesizer, I'd use it for either right or left-hand parts - basically arpeggios, triads, bass, or melodies.  It would never be the instrument I was fully facing, the main instrument, complimented by the others.  It would serve quite well in those capacities, but not as the base instrument in a multi-instrument set up. 

[LPF83]

I don't have a six-voice instrument.  I find eight voices to be restrictive enough.  But if I did have a six-voice synthesizer, I'd use it for either right or left-hand parts - basically arpeggios, triads, bass, or melodies.  It would never be the instrument I was fully facing, the main instrument, complimented by the others.  It would serve quite well in those capacities, but not as the base instrument in a multi-instrument set up.

I only recently started really noticing this on the Prophet6 and OB-6 because as 49 key keyboards, I very rarely play them with both hands.  More like left hand on the Prophet10, right hand on P6/OB6... or maybe I'd already recorded a lot of two handed work on the P10 so I'm using the P6/OB6 for a right-hand lead sound on top of it (which they're both astounding good at).   What changed was when I got the Komplete Kontrol S61 to replace my previous (Novation) 61 key primary controller.   Because it's so good at controlling other synths, I've started regularly playing the P6/OB6 through it just to get the extra octave and convenience. 

Where I did notice it before was when putting the P10 into polyunison-2 mode, which of course reduces the voices from 10 to 5.  The Rev2 and Prophet 12 handle pads nicely, of course, although I do stack a lot which in effect reduces them each to 8 and 6 voice synths respectively.

[Julienne Fractals]

Polychain. 2 x Prophet 6's = Sweet.

[LoboLives]

Just get a Dexibell with unlimited polyphony and you'll be happy. LOL

In all seriousness there's two ways around this.

Either get a P6 desktop and poly chain them or get rid of the P6 all together.

Unfortunately that's it. The P6 and OB6 are basically "accept the instrument for what it is...warts and all....or don't use them." type of approach which I appreciate.

To be honest, the polyphony argument is never ending....I've seen people complain about 16 voices, 32 voices, and even 125 voices and even saw a complaint on the new Kurzweil which is over 200 voices. I honestly can't take the argument or demands seriously anymore.

[LPF83]

Just get a Dexibell with unlimited polyphony and you'll be happy. LOL

In all seriousness there's two ways around this.

Either get a P6 desktop and poly chain them or get rid of the P6 all together.

Unfortunately that's it. The P6 and OB6 are basically "accept the instrument for what it is...warts and all....or don't use them." type of approach which I appreciate.

To be honest, the polyphony argument is never ending....I've seen people complain about 16 voices, 32 voices, and even 125 voices and even saw a complaint on the new Kurzweil which is over 200 voices. I honestly can't take the argument or demands seriously anymore.

Oh don't get me wrong, my P6 and OB6 aren't going anywhere (I don't see ever selling them, they are excellent at what they do).  For pads I have a third option available which is "use other synths for that", so it's not a deal breaker for me by any means.  And creativespiral is right, the release can be shortened and reverb/echo added to take up the slack, so there's another option.  Then of course there's multi-tracking (most of those amazing vintage Prophet 5 tracks were multitracked), and my favorite workaround of all which is just finding ways of playing the pad to avoid the peak-jump issue (I think it all becomes more interesting actually when the music itself becomes shaped by the limitations).

But, I thought it's worth discussing to see what if any workarounds folks had come up with...  even if someone is researching whether the P6/OB6 is the right synth for them -- if two-handed pads with long releases are their primary need, they will at least want to be informed of the limitation or the potential need to polychain (although I have not ruled out that possibility yet -- I like the P6 enough that I would be more likely to expand than replace -- plenty of modules on the used market too).

Sometimes synths with high voice counts do in fact suffer surprising polyphony deficits.  Especially in many VA synths like the Virus where the onboard FX reduce the available voice count... in those cases a "voice" is almost more like a slice of CPU than an actual "note" as we talk about them on an analog synth.  The mention of Kurzweil reminded me of a K2000 I owned briefly back in the early 90's....  24 voices with a note stealing algorithm that had a great reputation at the time, yet I still felt voice limited and ultimately sent it back.  So, I mean to each his own, but I can understand the other side.

[creativespiral]

Yeah, I think Always Retrigger from Start and/or slew to take the edge off would be welcome fixes.  I wonder if there are any synths currently that do this?  And if not, is there some technical hurdle that makes implementing it challenging?

Yes, the Moog Sub 37 and Moog One both have this Always Retrigger From Start option available.   It's called Envelope Reset on their products. In addition, the Moog One has some other really nice envelope options, like a Hold stage between Attack and Decay, the ability to switch timing of stages to ADR Sync Mode, which syncs ADR timings to BPM divisions for syncopated envelope motion, and ADR Contour/Curve Shaping, which allow for snappier bass/lead sounds, or smoother pad/string swells.   

Envelopes are really the heart of sound design.  If possible to add some, or all of these envelope options to current and/or future Sequential instruments, that would be at the top of my list of general feature requests / IP coding development, for both mono and poly instruments.   (that, and adding a mod transforms layer in the matrix, which is a real game changer for advanced sound design -- with all sorts of simple operators for math transforms (square, root, basic functions), value limiting, quantizing, slew, and multiplexing with other mod sources)

[the8bitdeity]

Optimally, you want to have both options available - standard retrigger, along with retrigger from start mode.  The "Always Retrigger from Start" option still has a voice stealing sound, but its nice to have as an alternate option.

Ultimately if there was a slew control / smoothing variable on retrigger from start, that would be the ultimate in flexibility.  This would be a great area of development for all Seqs synths... especially if its a shared library of code / IP that can be reused in future synths with similar envelope architecture.   That plus ADR contour shaping would really be great advances for envelopes / sound design.   

Yeah, I think Always Retrigger from Start and/or slew to take the edge off would be welcome fixes.  I wonder if there are any synths currently that do this?  And if not, is there some technical hurdle that makes implementing it challenging?

In my experience, Elektron offers some of the best options for envelopes, you get a number of curve options with and without reset. Hydra also has a bevy of options that make good use of the interface.

[MPM]

If 6 isn't enough you can always multitrack for recording, and use shell/power chords live.

I've always got the other hand on the OB~6 or Rev2, so 6 is no issue for me

[MPM]

The best workaround is probably to just keep envelope release timings a bit lower/quicker and use an ample amount of onboard, or external delay/reverb to take care of the release phase. .   

AMEN

[LPF83]

If 6 isn't enough you can always multitrack for recording, and use shell/power chords live.

I've always got the other hand on the OB~6 or Rev2, so 6 is no issue for me

When playing one hand at a time its really not an issue.  Multitracking to fill in lack of voices isn't really a possible solution for me..  because the issue is the sudden jump to the envelope peak (which is kind of a difficult anomaly to cover without simply lessening the envelope amount, and that's not always a feasible solution).

But I was doing a bit of experimenting on the P10 tonight with a sound which is particularly bad for this (even with 10 voices available), and my god what a difference it makes to put it in Prophet5 voice allocation mode instead of round robin, the issue basically goes away.   So maybe just implementing P5 voice mode on other Sequential synths would be the answer?  Yes it can change how a sounds when played, somewhat, but same could be said for most other solutions.

[creativespiral]

When playing one hand at a time its really not an issue.  Multitracking to fill in lack of voices isn't really a possible solution for me..  because the issue is the sudden jump to the envelope peak (which is kind of a difficult anomaly to cover without simply lessening the envelope amount, and that's not always a feasible solution).

But I was doing a bit of experimenting on the P10 tonight with a sound which is particularly bad for this (even with 10 voices available), and my god what a difference it makes to put it in Prophet5 voice allocation mode instead of round robin, the issue basically goes away.   So maybe just implementing P5 voice mode on other Sequential synths would be the answer?  Yes it can change how a sounds when played, somewhat, but same could be said for most other solutions.

This is a very good callout.   I found the same thing on Polybrute... almost always use the "Reassign" mode (aka: P5 / per key allocation)    It does make a huge difference for pads, and would love to see this feature added on more current synths (like P6,T5) and/or upcoming synths.   Ultimately it would be nice to have the "Reset/Backtrack" allocation type as well... which always tries to use the lowest voice number.

[LPF83]

When playing one hand at a time its really not an issue.  Multitracking to fill in lack of voices isn't really a possible solution for me..  because the issue is the sudden jump to the envelope peak (which is kind of a difficult anomaly to cover without simply lessening the envelope amount, and that's not always a feasible solution).

But I was doing a bit of experimenting on the P10 tonight with a sound which is particularly bad for this (even with 10 voices available), and my god what a difference it makes to put it in Prophet5 voice allocation mode instead of round robin, the issue basically goes away.   So maybe just implementing P5 voice mode on other Sequential synths would be the answer?  Yes it can change how a sounds when played, somewhat, but same could be said for most other solutions.

This is a very good callout.   I found the same thing on Polybrute... almost always use the "Reassign" mode (aka: P5 / per key allocation)    It does make a huge difference for pads, and would love to see this feature added on more current synths (like P6,T5) and/or upcoming synths.   Ultimately it would be nice to have the "Reset/Backtrack" allocation type as well... which always tries to use the lowest voice number.

Interesting, I'd be interested to hear how that sounds.  I assume its use the lowest voice number that's in the release stage, and if none available in release take one from sustain, as opposed to prioritizing stealing of the lowest note played?