The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Synths for sale  Search Gearslutz Go Advanced
DSI Rev2... to sell or not to sell? THAT is the annoying question.
Old 4 weeks ago
  #151
TJT
Lives for gear
Quote:
Originally Posted by creativespiral View Post

Important to note: The per-voice offsets are fairly stable, per oscillator, when you hold down a note... there my be some very minor drift, or VCO frequency jitter as described above, but overall they are stable offsets. If you try to create voice variance with an LFO routing or Osc Slop, you don't get the stability of the offsets... instead you end up with frequencies constantly shifting around with a more artificial drifting motion. For patches with quick decay (stabs, leads), this may not be something your really notice, but if you're playing chords, and holding down notes, there is a substantial difference between having modeled voice offsets, and trying to fake it with S+H, LFOs, or Slop controls.
Cheers, Jason
Can you give a comparison between your patches and a dco (or vco) synth with some lfo routing or oscillator slop that shows this "substantial" difference?
Old 4 weeks ago
  #152
Gear Guru
 
zerocrossing's Avatar
Quote:
Originally Posted by creativespiral View Post
Hey guys... I'm a few days late to this thread, but I'm the author of the http://www.VoiceComponentModeling.com website (and the Blade Runner video that was posted)... I'm gonna have a bunch more content uploaded soon, and I'm just putting finishing touches on a full bank of 128 presets using voice modeling techniques... the bank will be available soon... in my opinion, the patches sound pretty darn good... the classic synth patches sound a lot more like synths of yesteryear, and patches based on ensembles of acoustic instruments sound much more organic as well (strings, brass sections, wind instruments, plucks, organs, etc..)

Background: Late last year, I set out on a journey to try and understand what gives classic analog poly synths their "character".... ie: why are they warmer / fatter / more lush sounding than modern synth counterparts. I was able to take a bunch of recordings/measurements from a variety of classic poly synths, including the OBX, OBXA, SEM 4-Voice, Prophet 5, Prophet 10, Korg PolySix, Roland Jupiter 4, Jupiter 8, Yamaha CS-80, and MemoryMoog. For some, I captured samples from multiple different copies of the synth. In addition, I got readings for a variety of other VCO monosynths, DCO synths, and digital synths... suffice to say, I have done a lot of research on this to compare different aspects of each synth. There are various differences from synth-to-synth. Of course, the topic that always gets discussed in forums is Filter Types, which definitely has a significant effect on sound. Other aspects discussed are unique oscillator shapes, envelopes and amplifier sections.

The two "new areas" of research I was focused on are:
1. Voice-by-Voice Variance to a variety of aspects (osc freq offsets, filter offsets, envelope offsets, etc)
2. VCO Harmonic Jitter / Frequency Jitter (lookup my video on VCO Harmonic Jitter for more on this topic)

The VCO Harmonic Jitter is present in all VCO synths, classic to modern.. its a relatively high frequency, but low amount of frequency modulation that occurs on each oscillator (usually less than 2-3 cents swing, at a fairly high frequency) The effect of VCO freq jitter is hardly noticeable at the fundamental, but in the upper harmonic set, it has an effect of washing out / blurring some of the higher frequency content, which I believe results in a less harsh / more warm sound. It's very subtle though, and in the grand scheme of things, the effect less important when there's a lot of voices / modulation going on, with most poly synths. If you're playing mono lines, or more raw patches though, it is worth exploring if you want a vintage mono-synth type of tone.

In the big picture, Voice Variance may be the most important characteristic of recreating warm/organic sound of classic synths. I compared the classic poly synths (listed above), and they all have significant voice variance. First up is Oscillator fine tuning offsets. As you play each key on a classic poly synth, every voice will have unique offsets to tuning, and those offsets will be fairly stable if you hold down the note.

I tested each voice of each synthesizer, and each oscillator separately, over multiple octaves. The most common behavior is that each voice will have "intonation issues" throughout the keyboard range -- often in lower registers, the voice will be offset sharp, and as you go up the keyboard for a given voice, it will become more flat, relative to the target pitch for that given key strike. Note: There are several variations on voice-by-voice tuning depending on the synth, and voice-allocation method.

The result is that each voice in the instrument will have unique detuning offsets between its two (or three oscillators), creating a phasing/detuning motion for that note. And then that effect is multiplied when you play chords. Each voice in the chord will have unique tuning offsets (and each oscillator within that voice)... the net effect is that there is a significant amount of "natural phasing / natural detuning" happening.

This, in my opinion, is THE BIGGEST DIFFERENCE when comparing classic synth of yesteryear to their more modern counterparts... especially if trying to get classic sounds out of DCO or digital synths. (Note: even modern VCO boards could benefit from voice modeling, as they are generally more tamed than vintage gear) Modern electronics are inherently more stable, and DCOs/Digital synths are maybe just "too perfect" if going for classic synth vibe, but with these type of modulation techniques, the Rev2 may actually be able to sound more like a classic VCO poly synth than even modern VCO synths... since you can dial in exactly how much voice variance exists and to a variety of characteristics.

Important to note: The per-voice offsets are fairly stable, per oscillator, when you hold down a note... there my be some very minor drift, or VCO frequency jitter as described above, but overall they are stable offsets. If you try to create voice variance with an LFO routing or Osc Slop, you don't get the stability of the offsets... instead you end up with frequencies constantly shifting around with a more artificial drifting motion. For patches with quick decay (stabs, leads), this may not be something your really notice, but if you're playing chords, and holding down notes, there is a substantial difference between having modeled voice offsets, and trying to fake it with S+H, LFOs, or Slop controls.

Also, if you're trying to model an ensemble of violins in a patch, and holding chords, the analogous example would be that for every player in the violin ensemble, there is another person constantly turning the tuner knobs on each violin as they are playing.... its a highly un-realistic scenario.. and can be improved by voice modeling techniques like described on the website.

In addition to oscillator tuning offsets on a per-voice basis, there are dozens of other characteristics that can be modeled voice-by-voice, to create more realistic classic synth and acoustic ensemble patches. Varying envelope attack, decay, release speeds slightly produces a very organic sound, and makes patches with filter sweeps sound amazing (ie: Tom Sawyer type patches). Varying pitch settle effects, glitches, frequency jitter, noise, osc shape and other characteristics per voice can result in very organic sounding patches... definitely takes the "edge" off more recent synths and gives them some of the character of classics.

The Rev2 works extremely well for Voice Modeling... its a testament to the design and flexibility of DSI/Sequentials Gated Sequencer and Mod Matrix. I originally thought that many other synths would be able to accomplish this as well, but actually, there are very few that can do Voice Modeling at this point.

For now, there's no way I'm even considering letting go of a Rev2... until other manufacturers build in voice modeling capabilities. In the thread I posted on the Sequential forum, there's been some discussion with Pym and it sounds like DSI/Seq is taking the idea seriously and will have some sort of dedicated implementation for voice modeling in future boards. I'm hoping it becomes a widespread practice, as it makes recreating classic synth sounds much more realistic, as well as acoustic ensemble type instruments.

Also, in response to another comment I saw... once you get the hang of it, you can set up voice modeling in the Rev2 in a matter of a minute or less... It's not a big deal to program it into patches. Also, I have templates saved for a bunch of variants of voice modeling that I can now use as starting point for new patches.

Anyways, on the http://www.VoiceComponentModeling.com website, there is a bunch more info, if you're interested. Also, for anyone interested in some example patches, shoot me an email and I'd be happy to share some patches with you. Also, follow the thread on Sequential's forum for additional info: https://forum.sequential.com/index.p...ic,3449.0.html

Cheers, Jason
Great video. I watched it again today. You’re totally right about the jitter. I was calibrating my Dominion 1 and there it is, clear as day. Maybe that’s one of the reasons I love it so much. Such a great tone. Maybe it’s missing from some of the modern VCO polys like the Prophet 6... I’ve not looked at it as closely. I love it anyway though. Keep up the good work. I’m missing my Tetras a lot, so I wouldn’t surprise anyone if I pick up another pair or a REV2 around Christmas time, or if my grandmother dies. (Inside Gearslutz joke... that I’m not even sure I understand.
Old 4 weeks ago
  #153
Lives for gear
 

All very well but I would say if you want VCO sound, buy a VCO synth. We have the technology! If you have a vanilla toned DCO DSI with a ton of modulation that can be employed to try and make it sound more like a VCO then great, but as a general idea for synthesis it's what the VA synth makers have been doing for decades. It does work, but ultimately it's advanced turd polishing (ATP Synthesis, website coming soon).
Old 4 weeks ago
  #154
Here for the gear
 

Quote:
Originally Posted by rockmanrock View Post
All very well but I would say if you want VCO sound, buy a VCO synth. We have the technology! If you have a vanilla toned DCO DSI with a ton of modulation that can be employed to try and make it sound more like a VCO then great, but as a general idea for synthesis it's what the VA synth makers have been doing for decades. It does work, but ultimately it's advanced turd polishing (ATP Synthesis, website coming soon).
What VA synths do you know of that offer per-voice modeling of oscillators, filters, amps and envelopes? I don't think its as common as you might think. Monophonic unison detune is prevalent, and there are some other synths that offer "Slop like" controls, but those are not the same as poly voice modeling... its a totally different animal.

I did some research a few weeks ago and figured out that SonicProjects OP-X and u-He Diva VSTs have some limited capability for voice offsets for oscillators. But both pale in comparison, in terms of capabilities and flexibility that the Rev2 has. (I have recently purchased OP-X Pro II and its awesome, btw)

With the Rev2 you can model basically anything though... you can set virtual voice counts and target attributes on a per voice basis: osc frequency, shape mod, noise, envelope attack, decay, release, resonance, audio mod, vca levels, and even complex macro type behaviors like oscillator settle envelopes, glitch envelopes, harmonic freq jitter, and other characteristics common to classic poly synths from the 70s/80s... everything can have offsets unique to the voices... just like the voice boards on synths of yesteryear.

Given similar filters and similar general architecture, I'd say the Rev2 may be more capable of modeling classic poly synth sounds than even some modern VCO poly synths.
Old 4 weeks ago
  #155
Here for the gear
 

Quote:
Originally Posted by TJT View Post
Can you give a comparison between your patches and a dco (or vco) synth with some lfo routing or oscillator slop that shows this "substantial" difference?
I recorded a quick demo, below. At the end there is an A/B test with Slop first, then Voice Component Modeling. The longer you hold chords, the more noticeable the differences are... with voice modeling, there is still substantial phasing/detuning happening, but each oscillator is stable in its offsets, so there's not additional "artificial motion" happening. It's a natural / more controlled detuning, like you find when you actually measure oscillators from classic synths... I tested OBX, OBXA, two MemoryMoog, three Jupiters, two CS80s, multiple Prophets and others.

With Slop, you get a different result every time you play something, since its based on a macro type control of multiple free running LFOs that move each oscillator pitch. Sometimes you'll hit a chord and it will sound great, but then you might hit another chord and it will sound like crap because the two oscillators for those voices have either diverged to extreme tuning points, or one or more of the oscillators is "in the middle of its LFO cycle" and accelerating with artificial motion... ie: you will sometimes get a nice detuned sound and other times, lots of weird dissonance.

Old 4 weeks ago
  #156
Lives for gear
 
Xero's Avatar
depends what you want, for me, this is easy

is it called rev2?

does it have ssm 2040 filters?

no?

sell it
Old 4 weeks ago
  #157
Lives for gear
 
markodarko's Avatar
 

Quote:
Originally Posted by creativespiral View Post
If you try to create voice variance with an LFO routing or Osc Slop, you don't get the stability of the offsets... instead you end up with frequencies constantly shifting around with a more artificial drifting motion.
Thanks for the post, Jason.

With regards to the above, I completely agree. I’ve been using Note Number / Keyboard assigned to the pitch of an oscillator on the Rev2 and Peak since I’ve had them as the built-in “slop” is not useful for analog mimicking at all. It’s useful as an effect, but not for the purpose that we’re talking about.

On the Rev2 I believe I had Note Number set to -1 and the pitch adjusted to +26 (if I recall). That allowed a subtle variation of pitch across the keyboard and was most notable when using 2 oscillators where that trick is only applied to one of them.

After returning my Rev2 I do miss it, I have to say. I think the 30 days wasn’t long enough for me to get to know it. Whilst other synths it's as though I’ve always known them, the Rev2 is a slow-burn, I feel. With that in mind I intend on buying another one in a few months time and this time keep it.
Old 4 weeks ago
  #158
Lives for gear
 
markodarko's Avatar
 

Quote:
Originally Posted by creativespiral View Post
What VA synths do you know of that offer per-voice modeling of oscillators, filters, amps and envelopes?
I believe Alchemy can do this as it has a step sequencer that can be assigned to any destination but I’ve not looked into it. It’s only from memory. Also, it’s only on Logic now so not really applicable if people don’t have Logic.
Old 4 weeks ago
  #159
ozy
Lives for gear
 
ozy's Avatar
 

Quote:
Originally Posted by markodarko View Post
it always needs a lot of work to coax sounds out of it. It seems to have a sweet spot >this< big,
Excuse me if this will sound a bit harsh, but...

... you can't fix a 1$ swing with a 500$ bat.

I use three different "REV 1", two contain 150 patches emulating the Prophet T8, one is dedicated to emulating the Prophet 5 factory sounds.

Well... that makes for a quite wide "sweet spot" imho

Did it take me time? Yes it did.

Last edited by ozy; 4 weeks ago at 11:37 AM.. Reason: graphic clarity
Old 4 weeks ago
  #160
Gear Maniac
 

Now I don't play each and every instrument I have every day... It took me about a year to really get my Tempest's osc's and envelopes. Fantastic instrument but it was hard for me to make it sing.
Old 4 weeks ago
  #161
Lives for gear
Quote:
Originally Posted by vlz View Post
That's the REV2 -- only per voice. But it shouldn't be that hard for the OS to offer a global option as the Prologue does.
If you have Logic Pro it's very easy to add an infinite number of mono lfos to it.
I guess other daws can do the same...
Old 4 weeks ago
  #162
Gear Guru
 
zerocrossing's Avatar
Quote:
Originally Posted by rockmanrock View Post
All very well but I would say if you want VCO sound, buy a VCO synth. We have the technology! If you have a vanilla toned DCO DSI with a ton of modulation that can be employed to try and make it sound more like a VCO then great, but as a general idea for synthesis it's what the VA synth makers have been doing for decades. It does work, but ultimately it's advanced turd polishing (ATP Synthesis, website coming soon).
I’m all for a good VCO synth, but a lot of us were “stretched out” by the feature rich VA synths of the 90s, and simple VCO synths aren’t up for the types of sounds we’re used to making. The 08 architecture synths are a lot closer to what I like in a synth, but without the pesky aliasing.

Of course, there are synths that have a ton of features and VCOs, but the two that I can think of will cost between 4 and 8 times the amount of a REV2. Not really an option for many of us. So, using a little technique like voice parameter offsets is a great way of having your cake and eating it too. Some synths like the Peak and DeepMind already have something like this built in.
Old 4 weeks ago
  #163
Lives for gear
Sell that P.O.S

Rev 2 gets you ZERO cool points round here.

Besides....I don't know, sometimes I get feelings...and...I've just got this feeling....no...I KNOW...that YOU NEED to buy a NEW SYNTH ASAP
Old 4 weeks ago
  #164
Gear Addict
 
jurfin's Avatar
I find it strange that there’s such a strong opposition from some to the VCM technique. It has made a noticeable difference and improvement on my Rev 2 patches.

I was always reasonably happy with LFOs and slop before, but this is better. Try it out, along with the jitter technique. It sounds great
Old 4 weeks ago
  #165
Lives for gear
 

Quote:
Originally Posted by creativespiral View Post
What VA synths do you know of that offer per-voice modeling of oscillators, filters, amps and envelopes? I don't think its as common as you might think. Monophonic unison detune is prevalent, and there are some other synths that offer "Slop like" controls, but those are not the same as poly voice modeling... its a totally different animal.

I did some research a few weeks ago and figured out that SonicProjects OP-X and u-He Diva VSTs have some limited capability for voice offsets for oscillators. But both pale in comparison, in terms of capabilities and flexibility that the Rev2 has. (I have recently purchased OP-X Pro II and its awesome, btw)

With the Rev2 you can model basically anything though... you can set virtual voice counts and target attributes on a per voice basis: osc frequency, shape mod, noise, envelope attack, decay, release, resonance, audio mod, vca levels, and even complex macro type behaviors like oscillator settle envelopes, glitch envelopes, harmonic freq jitter, and other characteristics common to classic poly synths from the 70s/80s... everything can have offsets unique to the voices... just like the voice boards on synths of yesteryear.

Given similar filters and similar general architecture, I'd say the Rev2 may be more capable of modeling classic poly synth sounds than even some modern VCO poly synths.
Yes, OP-X did spring to mind. In some cases this microtuning stuff will also be being done as an under-the-hood part of the VA engine. The VA developers version of what you're doing.

My point is, with real VCO polys available and more about to saturate the market like the monos have done, it's going to be far easier and better to get hold of a real VCO synth. I don't think an expensive Curtis voice-on-a-chip DCO synth is worth it, that's why I've stuck with my Matrix 1000. They sound alright, but they always sound like what they are. For owners who enjoy cranial brickwall therapy then yes, all this programming will get them somewhere, but for anyone else looking for VCO sound, it's a red herring.
Old 4 weeks ago
  #166
Gear Guru
 
zerocrossing's Avatar
Quote:
Originally Posted by markodarko View Post
With that in mind I intend on buying another one in a few months time and this time keep it.
Unless, of course, we find out what Dave’s replacing the 12 and 2 with and that looks better.
Old 4 weeks ago
  #167
Gear Guru
 
zerocrossing's Avatar
Quote:
Originally Posted by rockmanrock View Post
Yes, OP-X did spring to mind. In some cases this microtuning stuff will also be being done as an under-the-hood part of the VA engine. The VA developers version of what you're doing.

My point is, with real VCO polys available and more about to saturate the market like the monos have done, it's going to be far easier and better to get hold of a real VCO synth. I don't think an expensive Curtis voice-on-a-chip synth is worth it, that's why I've stuck with my Matrix 1000. They sound alright, but they always sound like what they are. For owners who enjoy cranial brickwall therapy then yes, all this programming will get them somewhere, but for anyone else looking for VCO sound, it's a red herring.
Did you watch his demo? I’d hardly call it “cranial brick wall therapy.” The entire video is about a half hour long, but he’s going into a lot of explanations and details that once you have that knowledge, you’ll be able to do it yourself in probably 5-10 minutes. Make a couple of template presets and you’ll really never have to do it again after that.

And again, let’s see the VCO poly module that’s got 16 voices, 4 LFOs, 4 EGs, Mod sequencer, poly aftertouch (useful for Blade Runner type sounds) bit mod matrix and effects for $1700. I don’t see it happening any time soon.
Old 4 weeks ago
  #168
Lives for gear
 

Quote:
Originally Posted by jurfin View Post
I find it strange that there’s such a strong opposition from some to the VCM technique. It has made a noticeable difference and improvement on my Rev 2 patches.

I was always reasonably happy with LFOs and slop before, but this is better. Try it out, along with the jitter technique. It sounds great
Got no opposition to it as its an interesting way to expand what one can get out of the Rev2, just I agree with the point that if people want the VCO sound, go get a VCO based synth as it will just do it out of the box.
Plus people have made great music making the most of the DCO characteristics, particularly with Junos.
Old 4 weeks ago
  #169
Lives for gear
 

Quote:
Originally Posted by zerocrossing View Post
Did you watch his demo? I’d hardly call it “cranial brick wall therapy.” The entire video is about a half hour long, but he’s going into a lot of explanations and details that once you have that knowledge, you’ll be able to do it yourself in probably 5-10 minutes. Make a couple of template presets and you’ll really never have to do it again after that.

And again, let’s see the VCO poly module that’s got 16 voices, 4 LFOs, 4 EGs, Mod sequencer, poly aftertouch (useful for Blade Runner type sounds) bit mod matrix and effects for $1700. I don’t see it happening any time soon.
If you go back to markodarko's thread opener, it sounds like he has a general problem with the character of the synth. No amount of crafty programming is going to change that. It's just papering over the DCO/"I really want a VCO"/"this sounds off" cracks. I have done a ton of this in various forms over the years and really it's a waste of time if you can get the thing that really does the job. Trying to get FM to sound analogue. Trying to get a sampler to sound analogue. Trying to get a DCO to sound more like VCO. Then you get the real thing and you realise how futile a lot of that effort was. If you already own a DSI and like it then great, these techniques are useful, but I really wouldn't recommend getting an expensive DCO at this point. Especially when he's had it once and didn't like it!

The module you've described sounds like Dave's next synth! 3340 oscs through SSM filters with modern modulation and effects. With really great sounding VCO synths you don't even need so much modulation. On a DCO/VA it's all needed to try and raise the dead.

It's like doing this instead of getting a new cat:

Old 4 weeks ago
  #170
Quote:
Originally Posted by creativespiral View Post
Hey guys... I'm a few days late to this thread, but I'm the author of the http://www.VoiceComponentModeling.com website (and the Blade Runner video that was posted)... I'm gonna have a bunch more content uploaded soon, and I'm just putting finishing touches on a full bank of 128 presets using voice modeling techniques... the bank will be available soon... in my opinion, the patches sound pretty darn good... the classic synth patches sound a lot more like synths of yesteryear, and patches based on ensembles of acoustic instruments sound much more organic as well (strings, brass sections, wind instruments, plucks, organs, etc..)

Background: Late last year, I set out on a journey to try and understand what gives classic analog poly synths their "character".... ie: why are they warmer / fatter / more lush sounding than modern synth counterparts. I was able to take a bunch of recordings/measurements from a variety of classic poly synths, including the OBX, OBXA, SEM 4-Voice, Prophet 5, Prophet 10, Korg PolySix, Roland Jupiter 4, Jupiter 8, Yamaha CS-80, and MemoryMoog. For some, I captured samples from multiple different copies of the synth. In addition, I got readings for a variety of other VCO monosynths, DCO synths, and digital synths... suffice to say, I have done a lot of research on this to compare different aspects of each synth. There are various differences from synth-to-synth. Of course, the topic that always gets discussed in forums is Filter Types, which definitely has a significant effect on sound. Other aspects discussed are unique oscillator shapes, envelopes and amplifier sections.

The two "new areas" of research I was focused on are:
1. Voice-by-Voice Variance to a variety of aspects (osc freq offsets, filter offsets, envelope offsets, etc)
2. VCO Harmonic Jitter / Frequency Jitter (lookup my video on VCO Harmonic Jitter for more on this topic)

The VCO Harmonic Jitter is present in all VCO synths, classic to modern.. its a relatively high frequency, but low amount of frequency modulation that occurs on each oscillator (usually less than 2-3 cents swing, at a fairly high frequency) The effect of VCO freq jitter is hardly noticeable at the fundamental, but in the upper harmonic set, it has an effect of washing out / blurring some of the higher frequency content, which I believe results in a less harsh / more warm sound. It's very subtle though, and in the grand scheme of things, the effect less important when there's a lot of voices / modulation going on, with most poly synths. If you're playing mono lines, or more raw patches though, it is worth exploring if you want a vintage mono-synth type of tone.

In the big picture, Voice Variance may be the most important characteristic of recreating warm/organic sound of classic synths. I compared the classic poly synths (listed above), and they all have significant voice variance. First up is Oscillator fine tuning offsets. As you play each key on a classic poly synth, every voice will have unique offsets to tuning, and those offsets will be fairly stable if you hold down the note.

I tested each voice of each synthesizer, and each oscillator separately, over multiple octaves. The most common behavior is that each voice will have "intonation issues" throughout the keyboard range -- often in lower registers, the voice will be offset sharp, and as you go up the keyboard for a given voice, it will become more flat, relative to the target pitch for that given key strike. Note: There are several variations on voice-by-voice tuning depending on the synth, and voice-allocation method.

The result is that each voice in the instrument will have unique detuning offsets between its two (or three oscillators), creating a phasing/detuning motion for that note. And then that effect is multiplied when you play chords. Each voice in the chord will have unique tuning offsets (and each oscillator within that voice)... the net effect is that there is a significant amount of "natural phasing / natural detuning" happening.

This, in my opinion, is THE BIGGEST DIFFERENCE when comparing classic synth of yesteryear to their more modern counterparts... especially if trying to get classic sounds out of DCO or digital synths. (Note: even modern VCO boards could benefit from voice modeling, as they are generally more tamed than vintage gear) Modern electronics are inherently more stable, and DCOs/Digital synths are maybe just "too perfect" if going for classic synth vibe, but with these type of modulation techniques, the Rev2 may actually be able to sound more like a classic VCO poly synth than even modern VCO synths... since you can dial in exactly how much voice variance exists and to a variety of characteristics.

Important to note: The per-voice offsets are fairly stable, per oscillator, when you hold down a note... there my be some very minor drift, or VCO frequency jitter as described above, but overall they are stable offsets. If you try to create voice variance with an LFO routing or Osc Slop, you don't get the stability of the offsets... instead you end up with frequencies constantly shifting around with a more artificial drifting motion. For patches with quick decay (stabs, leads), this may not be something your really notice, but if you're playing chords, and holding down notes, there is a substantial difference between having modeled voice offsets, and trying to fake it with S+H, LFOs, or Slop controls.

Also, if you're trying to model an ensemble of violins in a patch, and holding chords, the analogous example would be that for every player in the violin ensemble, there is another person constantly turning the tuner knobs on each violin as they are playing.... its a highly un-realistic scenario.. and can be improved by voice modeling techniques like described on the website.

In addition to oscillator tuning offsets on a per-voice basis, there are dozens of other characteristics that can be modeled voice-by-voice, to create more realistic classic synth and acoustic ensemble patches. Varying envelope attack, decay, release speeds slightly produces a very organic sound, and makes patches with filter sweeps sound amazing (ie: Tom Sawyer type patches). Varying pitch settle effects, glitches, frequency jitter, noise, osc shape and other characteristics per voice can result in very organic sounding patches... definitely takes the "edge" off more recent synths and gives them some of the character of classics.

The Rev2 works extremely well for Voice Modeling... its a testament to the design and flexibility of DSI/Sequentials Gated Sequencer and Mod Matrix. I originally thought that many other synths would be able to accomplish this as well, but actually, there are very few that can do Voice Modeling at this point.

For now, there's no way I'm even considering letting go of a Rev2... until other manufacturers build in voice modeling capabilities. In the thread I posted on the Sequential forum, there's been some discussion with Pym and it sounds like DSI/Seq is taking the idea seriously and will have some sort of dedicated implementation for voice modeling in future boards. I'm hoping it becomes a widespread practice, as it makes recreating classic synth sounds much more realistic, as well as acoustic ensemble type instruments.

Also, in response to another comment I saw... once you get the hang of it, you can set up voice modeling in the Rev2 in a matter of a minute or less... It's not a big deal to program it into patches. Also, I have templates saved for a bunch of variants of voice modeling that I can now use as starting point for new patches.

Anyways, on the http://www.VoiceComponentModeling.com website, there is a bunch more info, if you're interested. Also, for anyone interested in some example patches, shoot me an email and I'd be happy to share some patches with you. Also, follow the thread on Sequential's forum for additional info: https://forum.sequential.com/index.p...ic,3449.0.html

Cheers, Jason
Great stuff!
Old 4 weeks ago
  #171
vlz
Gear Addict
 
vlz's Avatar
Quote:
Originally Posted by camarao View Post
If you have Logic Pro it's very easy to add an infinite number of mono lfos to it.
I guess other daws can do the same...
That won't be a help in all sorts of ways. To start with the lfos will be outside the synth signal chain and out of reach of the mod matrix. Furthermore, it is just like putting the synth through an effects processor.
Old 4 weeks ago
  #172
TJT
Lives for gear
Quote:
Originally Posted by creativespiral View Post
I recorded a quick demo, below. At the end there is an A/B test with Slop first, then Voice Component Modeling. The longer you hold chords, the more noticeable the differences are... with voice modeling, there is still substantial phasing/detuning happening, but each oscillator is stable in its offsets, so there's not additional "artificial motion" happening. It's a natural / more controlled detuning, like you find when you actually measure oscillators from classic synths... I tested OBX, OBXA, two MemoryMoog, three Jupiters, two CS80s, multiple Prophets and others.

With Slop, you get a different result every time you play something, since its based on a macro type control of multiple free running LFOs that move each oscillator pitch. Sometimes you'll hit a chord and it will sound great, but then you might hit another chord and it will sound like crap because the two oscillators for those voices have either diverged to extreme tuning points, or one or more of the oscillators is "in the middle of its LFO cycle" and accelerating with artificial motion... ie: you will sometimes get a nice detuned sound and other times, lots of weird dissonance.

Those sound like two completely separate patches. The first sounds totally out of tune. The second sounds in tune. Is that the lowest that slop will go on the rev.2? Because, that's not very subtle at all. It's more of a "detune" patch than a voice variance effect.

Have you tried doing a comparison using very slow random lfos at very low amounts,key synced to reset to a different value on every key trigger? Low enough amounts to not get that overly detuned effect,but more the offset phasing effect? I think that would be a more direct comparison.
Old 4 weeks ago
  #173
Lives for gear
 

Quote:
Originally Posted by rockmanrock View Post
With really great sounding VCO synths you don't even need so much modulation. On a DCO/VA it's all needed to try and raise the dead.
I don't feel that's the case at all. Only applies if you stick to a certain mindset & expect a particular sound.
Your comment here completely undermines the rest of your post.



I must say, it does surprise me how often when I buy second hand synths, anything with VCO's usually needs seriously tuning, its actually quite alarming how out a lot of them are when I receive them. Got to wonder what tone deaf ****ers are buying this stuff as its worse than some 14 year old who's just started out with the cheapest guitar they could get their hands on.

Perhaps the danger of VCM is it gives legitimacy to poorly maintained & out of tune instruments? A Rev2 without all that is easily capable of emulating the subtle variations inherent with acoustic instruments as that's all in a pretty small window whereas the Lifeforms SV1 I had for a while was all over the place, regardless of how good it sounded, outside what would be acceptable with any other instrument. Yet heralded as an example of how VCO's should sound.
Old 4 weeks ago
  #174
Lives for gear
Quote:
Originally Posted by vlz View Post
That won't be a help in all sorts of ways. To start with the lfos will be outside the synth signal chain and out of reach of the mod matrix. Furthermore, it is just like putting the synth through an effects processor.
I don’t get your point...
You can use a midi lfo to control any parameter inside the Rev 2.
Are you sure you tried this on Logic?
It’s a MIDI plug in.
Old 4 weeks ago
  #175
Here for the gear
 

Wow, there's some strange vitriol here toward the Rev2. The Rev2 is an extremely deep and capable synth, that covers far more territory than any modern VCO poly (save the one)... 16 voices, two full timbral synth engines, deep modulation matrix and sequencer, etc (btw: I love VCO synths too, and use mine often)

Rockmanrock... you mention the Matrix 1000... The Matrix 1000 is one of the most similar synths in history from an architecture standpoint to the Rev2. It uses CEM3396 chips that are almost identical to the CEM3397/PA397 that is used in the Rev2... its about the closest match there is. The Rev2 is seen by many as a highly upgraded Oberheim Matrix1000 or Matrix 6.

Also, just to address other complaints... the Rev2 is not expensive, based on its feature set. In fact, minus the Deepmind, I'd say its the best value on the market for an 8 or 16 voice poly synth.

In general on the DCO topic... you should all realize that the waveshaper / wave generation in DCOs is analog.. they should be called DCAOs... digitally controlled analog oscillators... you get all the uniqueness of analog osc shape, with its rich series of harmonic content, and no aliasing. The difference is that the freq clock is digitally generated, which results in each oscillator being very stably tuned. This is an issue that is specifically addressed with voice modeling.. you can virtually define a voice-allocation table, then set offsets to each voice, giving it as much or as little tuning variation per voice as you want. It's a perfect solution in my opinion, as you can get a hot and humid day VCO type of sound or a perfectly tuned VCO type of sound, with a little variation in the lookup table. Also you can target any number of other characteristics to mimic classic boards and their electronic peculiarities... or turn it off and have perfectly tuned oscillators.

Regarding it being too much effort. It isn't. Now that I know the process, I can setup a patch from scratch in about a minute, with rich voice variance to tuning, and other characteristics modeled on a voice-by-voice basis. Also, as zerocrossing mentioned, you can always setup, then save templates.. I have several templates saved for variations on 4-voice, 5-voice, 6-voice, and 8-voice models with different voice characteristics targeted.
Old 4 weeks ago
  #176
Here for the gear
 

Quote:
Originally Posted by haze015 View Post
Perhaps the danger of VCM is it gives legitimacy to poorly maintained & out of tune instruments? A Rev2 without all that is easily capable of emulating the subtle variations inherent with acoustic instruments as that's all in a pretty small window whereas the Lifeforms SV1 I had for a while was all over the place, regardless of how good it sounded, outside what would be acceptable with any other instrument. Yet heralded as an example of how VCO's should sound.
Yeah, I spent months investigating classic synths, and as mentioned, took samples/measurements from most of the legendary poly synths from the 70s/80s. The "poor tuning performance" is present in all of them to some degree or another. The natural phasing/detuning motion that occurs from this softens the sound and is largely responsible for what I would call the "warmth" or "lushness" or "organic" character of these synths. When you compare to most digital or DCO synths, which have perfect tuning per voice / per osc, there is a significant difference. But in general, people love the sound of classic VCO poly synths... and I do too! So, yes, there may be some legitimacy in poor tuning... but its not a black/white thing. The voice modeling techniques allow you dial in any amount of this per-voice detuning though, which is nice.

Also, interesting to note, the whole birth of DCO synths was based on the fact that synth developers in the early 80s were trying to build better, more capable polyphonic synths, and getting acceptable tuning with multiple voices and oscillators within each voice was very difficult with VCO poly architecture at the time. Nowadays, VCO electronics have better quality control and performance. They don't degrade as fast, and there has been a rebirth of VCO poly synths... although, even with modern VCO polys, there have been many cases documented of poor tuning performance... from what I've heard, even the cutting-edge Moog One has some issues.
Old 4 weeks ago
  #177
Lives for gear
 

Quote:
Originally Posted by haze015 View Post
I don't feel that's the case at all. Only applies if you stick to a certain mindset & expect a particular sound.
Your comment here completely undermines the rest of your post.
I'm referring to the use of modulation to emulate VCOs, analogue envelope shapes and as you've mentioned, general poor maintenance. With real VCOs we start with imperfection and the struggle is to get it in tune and keep it there. With DCO we start with near perfection and then try to artificially add life by adding these faults. If they're already there we don't need another 20 mod slots.

Quote:
Originally Posted by haze015 View Post
I must say, it does surprise me how often when I buy second hand synths, anything with VCO's usually needs seriously tuning, its actually quite alarming how out a lot of them are when I receive them. Got to wonder what tone deaf ****ers are buying this stuff as its worse than some 14 year old who's just started out with the cheapest guitar they could get their hands on.
I've had a hell of a lot of stuff with minor faults. People just live with the faults or don't even test things before they sell it.

Quote:
Originally Posted by haze015 View Post
Perhaps the danger of VCM is it gives legitimacy to poorly maintained & out of tune instruments? A Rev2 without all that is easily capable of emulating the subtle variations inherent with acoustic instruments as that's all in a pretty small window whereas the Lifeforms SV1 I had for a while was all over the place, regardless of how good it sounded, outside what would be acceptable with any other instrument. Yet heralded as an example of how VCO's should sound.
I think that is a sound that some people chase. Copying BOC with their tape warble and Aphex Twin with his off-scale SH101.
Old 4 weeks ago
  #178
Gear Maniac
 

Quote:
Originally Posted by creativespiral View Post
Hey guys... I'm a few days late to this thread, but I'm the author of the http://www.VoiceComponentModeling.com website (and the Blade Runner video that was posted)... I'm gonna have a bunch more content uploaded soon, and I'm just putting finishing touches on a full bank of 128 presets using voice modeling techniques... the bank will be available soon... in my opinion, the patches sound pretty darn good... the classic synth patches sound a lot more like synths of yesteryear, and patches based on ensembles of acoustic instruments sound much more organic as well (strings, brass sections, wind instruments, plucks, organs, etc..)

Background: Late last year, I set out on a journey to try and understand what gives classic analog poly synths their "character".... ie: why are they warmer / fatter / more lush sounding than modern synth counterparts. I was able to take a bunch of recordings/measurements from a variety of classic poly synths, including the OBX, OBXA, SEM 4-Voice, Prophet 5, Prophet 10, Korg PolySix, Roland Jupiter 4, Jupiter 8, Yamaha CS-80, and MemoryMoog. For some, I captured samples from multiple different copies of the synth. In addition, I got readings for a variety of other VCO monosynths, DCO synths, and digital synths... suffice to say, I have done a lot of research on this to compare different aspects of each synth. There are various differences from synth-to-synth. Of course, the topic that always gets discussed in forums is Filter Types, which definitely has a significant effect on sound. Other aspects discussed are unique oscillator shapes, envelopes and amplifier sections.

The two "new areas" of research I was focused on are:
1. Voice-by-Voice Variance to a variety of aspects (osc freq offsets, filter offsets, envelope offsets, etc)
2. VCO Harmonic Jitter / Frequency Jitter (lookup my video on VCO Harmonic Jitter for more on this topic)

The VCO Harmonic Jitter is present in all VCO synths, classic to modern.. its a relatively high frequency, but low amount of frequency modulation that occurs on each oscillator (usually less than 2-3 cents swing, at a fairly high frequency) The effect of VCO freq jitter is hardly noticeable at the fundamental, but in the upper harmonic set, it has an effect of washing out / blurring some of the higher frequency content, which I believe results in a less harsh / more warm sound. It's very subtle though, and in the grand scheme of things, the effect less important when there's a lot of voices / modulation going on, with most poly synths. If you're playing mono lines, or more raw patches though, it is worth exploring if you want a vintage mono-synth type of tone.

In the big picture, Voice Variance may be the most important characteristic of recreating warm/organic sound of classic synths. I compared the classic poly synths (listed above), and they all have significant voice variance. First up is Oscillator fine tuning offsets. As you play each key on a classic poly synth, every voice will have unique offsets to tuning, and those offsets will be fairly stable if you hold down the note.

I tested each voice of each synthesizer, and each oscillator separately, over multiple octaves. The most common behavior is that each voice will have "intonation issues" throughout the keyboard range -- often in lower registers, the voice will be offset sharp, and as you go up the keyboard for a given voice, it will become more flat, relative to the target pitch for that given key strike. Note: There are several variations on voice-by-voice tuning depending on the synth, and voice-allocation method.

The result is that each voice in the instrument will have unique detuning offsets between its two (or three oscillators), creating a phasing/detuning motion for that note. And then that effect is multiplied when you play chords. Each voice in the chord will have unique tuning offsets (and each oscillator within that voice)... the net effect is that there is a significant amount of "natural phasing / natural detuning" happening.

This, in my opinion, is THE BIGGEST DIFFERENCE when comparing classic synth of yesteryear to their more modern counterparts... especially if trying to get classic sounds out of DCO or digital synths. (Note: even modern VCO boards could benefit from voice modeling, as they are generally more tamed than vintage gear) Modern electronics are inherently more stable, and DCOs/Digital synths are maybe just "too perfect" if going for classic synth vibe, but with these type of modulation techniques, the Rev2 may actually be able to sound more like a classic VCO poly synth than even modern VCO synths... since you can dial in exactly how much voice variance exists and to a variety of characteristics.

Important to note: The per-voice offsets are fairly stable, per oscillator, when you hold down a note... there my be some very minor drift, or VCO frequency jitter as described above, but overall they are stable offsets. If you try to create voice variance with an LFO routing or Osc Slop, you don't get the stability of the offsets... instead you end up with frequencies constantly shifting around with a more artificial drifting motion. For patches with quick decay (stabs, leads), this may not be something your really notice, but if you're playing chords, and holding down notes, there is a substantial difference between having modeled voice offsets, and trying to fake it with S+H, LFOs, or Slop controls.

Also, if you're trying to model an ensemble of violins in a patch, and holding chords, the analogous example would be that for every player in the violin ensemble, there is another person constantly turning the tuner knobs on each violin as they are playing.... its a highly un-realistic scenario.. and can be improved by voice modeling techniques like described on the website.

In addition to oscillator tuning offsets on a per-voice basis, there are dozens of other characteristics that can be modeled voice-by-voice, to create more realistic classic synth and acoustic ensemble patches. Varying envelope attack, decay, release speeds slightly produces a very organic sound, and makes patches with filter sweeps sound amazing (ie: Tom Sawyer type patches). Varying pitch settle effects, glitches, frequency jitter, noise, osc shape and other characteristics per voice can result in very organic sounding patches... definitely takes the "edge" off more recent synths and gives them some of the character of classics.

The Rev2 works extremely well for Voice Modeling... its a testament to the design and flexibility of DSI/Sequentials Gated Sequencer and Mod Matrix. I originally thought that many other synths would be able to accomplish this as well, but actually, there are very few that can do Voice Modeling at this point.

For now, there's no way I'm even considering letting go of a Rev2... until other manufacturers build in voice modeling capabilities. In the thread I posted on the Sequential forum, there's been some discussion with Pym and it sounds like DSI/Seq is taking the idea seriously and will have some sort of dedicated implementation for voice modeling in future boards. I'm hoping it becomes a widespread practice, as it makes recreating classic synth sounds much more realistic, as well as acoustic ensemble type instruments.

Also, in response to another comment I saw... once you get the hang of it, you can set up voice modeling in the Rev2 in a matter of a minute or less... It's not a big deal to program it into patches. Also, I have templates saved for a bunch of variants of voice modeling that I can now use as starting point for new patches.

Anyways, on the http://www.VoiceComponentModeling.com website, there is a bunch more info, if you're interested. Also, for anyone interested in some example patches, shoot me an email and I'd be happy to share some patches with you. Also, follow the thread on Sequential's forum for additional info: https://forum.sequential.com/index.p...ic,3449.0.html

Cheers, Jason
Thank you for sharing this Jason I like the sound of the VCM far more than the slop on my Rev2 and OB6! Look forward to buying your patches!
Old 4 weeks ago
  #179
Here for the gear
 

Quote:
Originally Posted by TJT View Post
Those sound like two completely separate patches. The first sounds totally out of tune. The second sounds in tune. Is that the lowest that slop will go on the rev.2? Because, that's not very subtle at all. It's more of a "detune" patch than a voice variance effect.

Have you tried doing a comparison using very slow random lfos at very low amounts,key synced to reset to a different value on every key trigger? Low enough amounts to not get that overly detuned effect,but more the offset phasing effect? I think that would be a more direct comparison.
In this example, I set the slop at 30, which is relatively high, but gives approximately the same "range from zero" as the voice modeling example... ie: approximately 10-15 cents sharp or flat from centerpoint.

The biggest difference you're hearing (and dissonance) is happening when the two oscillators in each voice are at divergent values... ie: Osc1 might be -12 cents flat and Osc2 might be +12 cents sharp, just based on the current phase that the LFO is in. (Osc Slop is just multiple LFOs) And then that detuning is just multiplied by the fact that each voice has slop detuning going on.

As mentioned before there are two "categories of phasing / detuning"... You get natural phasing detuning motion just by juxtaposing two oscillators with some tuning offset between them. There does not need to be any motion in each oscillator.. just a stable offset, and it will produce "natural phasing / natural detuning" motion. Most people thing of phasing/detuning motion requiring actual motion in each oscillators frequency, but that is not the case. That is what I would call "oscillator drift" and is a more artificial motion above and beyond the natural phasing of just having two oscillators with offsets. If you analyze classic VCO synths (as I have), you'll notice that if you hold down a key/voice you will get a mostly stable frequency... there may be a tiny amount of drift, and likely some harmonic frequency jitter, but overall each oscillator will have a stable centerpoint of frequency with a unique offset.

But back to the example... at the end with the A/B test, I actually recorded it a few times with slop... not all of them sounded that bad / dissonant... I did pick the worst sounding one for comparison... but that's just the nature of using LFOs / Slop to get detuning... there is constant motion, and when two oscillators are at their extreme points based on divergent phase of the LFOs, it sounds pretty terrible. But you're always going to get that randomly when using LFOs/Slop at higher values.

In contrast, by just using a "Lookup Table" like in the Voice Modeling, you can set a wide range of detuning that is unique on a per-voice / per-osc basis, but each osc is stable, and the result is lots of "natural phasing" without the randomness and artificial drifting motion. Its much more realistic for modeling classic synth sounds. Also, much more realistic for modeling acoustic ensembles... ie: violin ensemble patches, brass, wind, etc... if you analyzed the tuning of each string on each violin in an ensemble, you'll find that for each player (each voice), their instrument has unique offsets in tuning per string (each oscillator). One player might have (+3 cents, +1 cents, -4 cents, 0 cents) for their four strings, and another might be further detuned... but there is no drifting motion on each string (performance modulation is a separate topic).
Old 4 weeks ago
  #180
Lives for gear
 

Quote:
Originally Posted by rockmanrock View Post
I'm referring to the use of modulation to emulate VCOs, analogue envelope shapes and as you've mentioned, general poor maintenance. With real VCOs we start with imperfection and the struggle is to get it in tune and keep it there. With DCO we start with near perfection and then try to artificially add life by adding these faults. If they're already there we don't need another 20 mod slots.
But that fails to account for the popularity of synths like the Roland Juno's, DCO-based with limited modulation, yet sound great.

Adding "life" is not done through "faults". My guitars don't sound great going out of tune or if the intonation is out, so why is this fetishised amongst these forums?
My SH101 sounds like **** when it out of tune because its out of tune. Artifically mimicking that or thinking that it's in any way a good sound is worrying to say the least.
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
otic / Electronic Music Instruments and Electronic Music Production
157
teashea / Electronic Music Instruments and Electronic Music Production
53
unicorn74 / Electronic Music Instruments and Electronic Music Production
18
dovecote / So Much Gear, So Little Time
2

Forum Jump
Forum Jump