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
Michael Jackson Beat It Gong Virtual Instrument Plugins
Old 4th January 2011
  #31
Gear Head
 

Quote:
Originally Posted by acreil View Post
Actually a bit crusher effect applied to the output won't quite cut it. The Synclavier's waveforms are 8 bit per voice, before any amplitude envelope. The result is harmonic distortion rather than the broadband quantization noise you get from a bit crusher.
Yes, I'm aware of that.

I'm just talking about an approximation of it. You would need separate envelopes for each individual harmonic, i.e. AT LEAST 48 separate instances of Absynth (or simple sine oscillators), each with its own bit crusher, and then triple or quadruple each set of 24 for detuning.

In my opinion, a more impressive trick of the Synclavier is reducing the level of the harmonic oscillators so that the quantization "noise" becomes audible, which creates pulse waves. Increasing the level by stepping through timbre frames creates a filtering effect so that fewer harmonics are actually audible as the strength of the sine wave becomes digitally stronger. Digital filters on a Synclav--who knew? Although I haven't tried it yet, I'm concerned there might be a "stepping" effect when crossfading from the quantization pulse to a full sine wave. But since the Synclavier always uses two voices for timbre frames, that might be a non-issue.

But that's just my opinion.

As to my homegrown Absynth FM Gong, the basic recipe is to create classic bell sounds and layer them. It's a basic tubular bell sound. One of those is a simple FM, glassy kind of sound at 5:1 (mod:carrier). Take THAT as your carrier wave and modulate it at 12:1.

On a DX7 or in FM8, you can get pretty close by using 2 modulators and 4 carriers. On the DX7, that would be algorithms 21 and 22 (22 gets my vote). In FM8, you might do better with 5 carriers and 1 modulator, since that would be the simplest setup, and it's not available on the DX7. On the DX, your carriers would be at 1, 4, 6, and 9 with both modulators set to 12. FM8 using 5 carriers you'd just add 14, and a single modulator at 12. The mod index would need to be fairly high. That's about as basic as I can make it, and be generous with chorusing effects. I have a TX7, a DX7II, and a TX802. I'd use the DX7II with two programs using this recipe, slightly detuned, and I'd use the TX802 for more detail at the cost of voices since I could layer enough over 8 oscillators to fill out the rest of the sound. But that would take more time than I have. I'm estimating it would take 5 instances of FM8 to do the job. To get the chorusing effect, you'd want to repeat that between 6 and 8 times.

Or, you can get a Synclavier and do it with just two partial timbres or Absynth with just two channels.

The above recipe only accounts for one Absynth channel. I got the other sound more through trial and error, initially using more harmonics but cutting it down to 24, just as you would on the real thing. On that wave, I set the single sine modulator to 5 and the carrier (of course) to 1. They're set slightly higher than that, actually, but I used a pitch envelope to bring it back down in tune. I left the amp EG as a sustaining sound, and I figured anyone wanting to use it would tweak that for whatever they wanted. I'd personally like it better with a slight filter vibrato on the end--but again, it's just an attempt at recreating it. If I can find the actual Synclav program, it should be interesting to compare.
Old 5th January 2011
  #32
Here for the gear
 

AngelRho: Just tried your Absynth patch. Very well done! Just hit a low Ab then down to an Eb -- it'll probably fake out most ears.
Old 30th March 2011
  #33
Gear Head
 

"Beat It" gong update--Good news!!!

OK... It's been a LONG time, and I absolutely REFUSE to give up on this. I really do hope someone out there is still reading!!!

First the bad news:
I looked ALL OVER my Synclavier for the gong sound. It is NOT there.

Now the GOOD news: I'm 98.7% sure that I have in my possession SCREEN SHOTS of ALL the original FM sounds!!! Now, by that I'm talking about what I THINK were 4 original floppy disks that came with the Synclavier II. Now, technically, the Synclavier PSMT IS a SyncII with more advanced features (such as sampling) and larger footprint. Therefore the FM sounds should be IDENTICAL. But since I did get the PSMT, not the SyncII, naturally I have none of the SyncII disks.

Which means I don't have the "Beat It" gong.

My understanding is the big blue vinyl LP demo was made using sounds from "Disk 3." I am in the process of getting screen shots of those and other factory sounds. Once I accomplish that, I can manually input the EXACT numbers to, I hope, perfectly recreate the sound.

Now, I'm not at my PSMT at the moment, so I'm going through some of these screenshots to see if I can guess which one it is. There is a timbre called "Giant MetalSheet." I'm purely guessing, but I think that's it. I plugged some of the numbers into Absynth to confirm. All I have so far are two partial timbres--only half of the complete sound, but it's enough to get a clue as to whether it's the real deal or not. The Absynth sound is only approximate. I wasn't able to plug in the exact settings in the spectrum window, but I got as close as Absynth would let me. Now, I have two different values for FM index in the patch window. Synclavier "harmonic volume" and Absynth "mod index" do not translate very neatly, at least not as far as I can tell at the moment. So I pushed both to what sounded good to me. If you're loading this patch, I'd suggest setting the mod index to something like 4 on both of them. With a little bit more work, I could probably get something on Absynth virtually indistinguishable, which I very well might do. But in the meantime I need to rebuild this sound on my own machine.

Let me know what you think!
Attached Files
File Type: zip Beat It (SynclvScrnsht).ksd.zip (9.4 KB, 206 views)
Old 30th March 2011
  #34
Gear Head
 

I have a patch from an old Akai cd called "Vintage Time Traveller" which has the Beat It sound on it - PM me and I'll dig it out for you...
Old 30th March 2011
  #35
Gear Nut
 

Hi,

I play a cover version with my band, I took a sample from the original and pitched/melodyned the notes .. made the release a little longer with some verb. They come from a high res mp3, but live the quality is more than enough. As far as I remember it is only 2 notes (D and E), ooops .. there's the intro too ofcourse

Don't mind sharing the samples .. I have them in the studio, can send them tomorrow if you want them! pm me if needed.

Good luck!

Or decrackle or denoise the sample from the file from the link mentioned earlier .. chop it up and melodyne and create samples/patch done! (“The Incredible Sounds of SynclavierÂ*II”, and otherÂ*hits « faunæ or automat?)
Attached Files

synclavier-demo45mj.mp3 (537.2 KB, 1857 views)


Last edited by chielio72; 30th March 2011 at 11:41 PM.. Reason: forgot the intro :-) .. added sample
Old 31st March 2011
  #36
Gear Head
 

Meh... Sampling is too easy! dfegad samples.

I'm just kidding, of course. My issue is that I have the actual synthesizer that originated that sound, yet I don't have the program itself. Now it is very soon to be ALL MINE!!! I'm interested in looking at the Synclavier timbres and recreating them on other plugins. I have Absynth and really enjoy it, though it seems a bit harsh every time I've tried copying a patch from another synth on it. I prefer designing sound on Absynth for its own sake. But at least with Absynth you can make Synclavier-esque sounds, which would be difficult by comparison with Yamaha DX7 or FM8. On the other hand, it's difficult to emulate a lot of DX sounds on the 'clav--I just made a DX-ish sound, but it took 8-voices to make it--all I have are 32 voices. For the kinds of sounds you'd make on a DX7, the algorithm system is much more efficient than the Synclav. BUT you can't use additive synthesis the create your own waveforms and THEN modulate them with FM. Suffice it to say the kinds of sound you'd make on a Synclav you wouldn't want to attempt on the DX7, and vice versa.

Anyway, I'm rambling again... There is a sort of "sport" value to building the patches either from a parameter list or just completely from scratch. You learn a lot about sound creation that way, things you might not have known otherwise. You can go to a grocery store to buy any kind of meat you want, and you don't have to work for it at all other than the exchange of money. So why is it people still spend hours out in the woods with a bow waiting to take down a prize buck? Ingenuity and know-how are valuable. Lifting the notes you need from a sample doesn't really give you any insight into the sound itself or the work that went into creating it. I'm very fortunate to actually own a Synclav, not something I'll soon take for granted.

If you want to get the samples "in the wild," might I suggest this:

Something that might work better than pulling the samples from the original recording is find a recording of the blue LP demo. I read somewhere that MJ straight copied it from that record anyway, note for note. And if you find it, you'll note there's not really much difference. I would think the blue record might be easier to take a sample from. When I get a chance, when I'm not punching in numbers, I'll see if I can find that recording online if someone else doesn't find it before I do. The one I'm thinking of was (I think) on a French website dedicated to Synclavier, but I honestly haven't seen it in a while and couldn't find it all last time I searched for it. You MIGHT have luck on 500sound.com. I just can't remember exactly where.
Old 31st March 2011
  #37
The sound is in my new Fairlight CMI Legacy II+ refill for reason, in the new bonus bank (which is free for all customers).
Old 31st March 2011
  #38
Lives for gear
Quote:
Originally Posted by Yoozer View Post
And that's a perfectly good reason. It was aimed at Shy's "take someone else's work". If it was a factory preset anyway, the original did exactly that - take someone else's work (the Synclavier employee who programmed it) and get away with it (and gain millions )

I'm not sure about the source, but I recall reading (was it VSE?) an explanation of a Synclavier owner who elaborated on how the FM exactly worked - instead of Yamaha's 6 operators in various configurations you'd have several 2-op pairs that you could stack at any pitch; each stack would reduce polyphony. Kind of like algorithms 5/6 on the DX7, only with as many pairs per voice as possible. I can't find the write-up, though.

The original already sounds layered - the percussive attack and the almost unison-like FM sound.
Real funny considering how huge the thing was....

Now you could have the equivalent of 16 synclaviers in a 2 pound unit and an apogee duet that sits on your lap sampling and all.
Old 31st March 2011
  #39
Gear Head
 

Quote:
Originally Posted by chielio72 View Post
Hi,

I play a cover version with my band, I took a sample from the original and pitched/melodyned the notes .. made the release a little longer with some verb. They come from a high res mp3, but live the quality is more than enough. As far as I remember it is only 2 notes (D and E), ooops .. there's the intro too ofcourse

Don't mind sharing the samples .. I have them in the studio, can send them tomorrow if you want them! pm me if needed.

Good luck!

Or decrackle or denoise the sample from the file from the link mentioned earlier .. chop it up and melodyne and create samples/patch done! (“The Incredible Sounds of SynclavierÂ*II”, and otherÂ*hits « faunæ or automat?)
That's the one I was talking about. It would take a lot of cleaning up, but honestly that's the one I'd use.

I'll be in my home studio most of the day, so I'll have it programmed on my 'clav by noon. W00h00!!!
Old 31st March 2011
  #40
Lives for gear
Quote:
Originally Posted by acreil View Post
Yeah but it's not just simple 2 operator FM. The modulator is a sine wave but the carrier is an arbitrary waveform with 24 definable harmonics. Software can smoothly crossfade between pairs of voices for time variant additive synthesis (or wavetable, if you want to think about it that way).

And it's quite low resolution, but since it uses variable sample rate playback and analog amplitude scaling, this manifests as harmonic distortion and image frequencies rather than aliasing and quantization noise.

I think stacked 2 operator FM can get "close enough" for this case, but a really good emulation will be very difficult.
I think this is a case for Zebra 2.

Zebra 2 could/would probably nail this sound...
Old 31st March 2011
  #41
Gear Head
 

Quote:
Originally Posted by systematika View Post
I think this is a case for Zebra 2.

Zebra 2 could/would probably nail this sound...
Hmmm... It probably could IF the FM module will allow you to create an additive waveform to use as a carrier wave. Do you have Zebra 2? The thing that's holding me back on Absynth is that it will only let you use 3 channels. It looks to me like Zebra uses 4, which would be perfect.

I do know with Absynth you can get around this somewhat by using "Double" oscillators in each channel. You could simulate the Synclav envelopes by setting up the additive waveform and applying the right amount of FM in the wave editor instead of in the patch window. You'd copy the additive source waves and paste them into the morph waves window. That's 1 morph wave per Synclav partial timbre for a total of 4 morph waves, and that would occupy 2 Absynth channels quite neatly. But here the problem is the Synclav partials aren't neat and tidy. The FM ratio is not a whole number, and the transform function in the Absynth wave window will only use whole numbers (there's a reason why). You'd have to listen to the effect that a detuned modulator wave has on the carrier and simulate the beat frequencies it creates by slightly modulating the pitch and amplitude with either another envelope or an LFO. This is where it gets really complicated. But, of course, I'm sure Absynth was designed with Synclavier 4-partial architecture in mind, right? heh

If you can create an additive waveform in Zebra to be used as a carrier wave in FM, and that x4, then you'll have no problem coming up with it. I don't have Zebra, so I'll leave that to the experts.

I bet this would also sound really nice in Alchemy, which I also don't have (drool). I've been burning more money in hardware, which has left me to make some painful choices like declining to get stuff like Alchemy and even upgrades to stuff I already own. I'm still using Absynth4 and Reason4, and I haven't even upgraded Finale since 2009. My priorities are: current MacOS, current version of Logic.

More to come...
Old 31st March 2011
  #42
Gear Nut
 

this sound is from old EMULATOR sample library

i'm 100% sure

probably you can find it on newer emu III, IV sample cds as well
Old 31st March 2011
  #43
Lives for gear
Quote:
Originally Posted by AngelRho View Post
Hmmm... It probably could IF the FM module will allow you to create an additive waveform to use as a carrier wave. Do you have Zebra 2? The thing that's holding me back on Absynth is that it will only let you use 3 channels. It looks to me like Zebra uses 4, which would be perfect.

I do know with Absynth you can get around this somewhat by using "Double" oscillators in each channel. You could simulate the Synclav envelopes by setting up the additive waveform and applying the right amount of FM in the wave editor instead of in the patch window. You'd copy the additive source waves and paste them into the morph waves window. That's 1 morph wave per Synclav partial timbre for a total of 4 morph waves, and that would occupy 2 Absynth channels quite neatly. But here the problem is the Synclav partials aren't neat and tidy. The FM ratio is not a whole number, and the transform function in the Absynth wave window will only use whole numbers (there's a reason why). You'd have to listen to the effect that a detuned modulator wave has on the carrier and simulate the beat frequencies it creates by slightly modulating the pitch and amplitude with either another envelope or an LFO. This is where it gets really complicated. But, of course, I'm sure Absynth was designed with Synclavier 4-partial architecture in mind, right? heh

If you can create an additive waveform in Zebra to be used as a carrier wave in FM, and that x4, then you'll have no problem coming up with it. I don't have Zebra, so I'll leave that to the experts.

I bet this would also sound really nice in Alchemy, which I also don't have (drool). I've been burning more money in hardware, which has left me to make some painful choices like declining to get stuff like Alchemy and even upgrades to stuff I already own. I'm still using Absynth4 and Reason4, and I haven't even upgraded Finale since 2009. My priorities are: current MacOS, current version of Logic.

More to come...
Yes you can edit harmonic partials. I think I stopped counting at 32 points before I got bored and I was only about 1/3 of the way through. You can also draw your own waveforms free-hand or "vector" style with lines and curves.

You can also morph partials transwave style and sync that process to LFOs and envelopes, perhaps that will help you to get the subtle variation you're looking for. You can have 4 operators if configured for FM synthesis but it doesn't have to be. (You can use it as subtractive if you want) The FM operators act on their own along with the other oscillators. You can self modulate the FM operator, or you can use the oscillators as the modulation source. It also has multiple outs.

Zebra 2 is like a bastard child of the synclavier, a fizmo, and absynth. It's definitely it's own beast, which is why it's virtually the only thing I have that's *not* a "modular" style plugin interface. Then again, I guess it's still semi-modular but just with a mod matrix. :P

Last edited by systematika; 31st March 2011 at 03:30 PM..
Old 31st March 2011
  #44
Gear Head
 

Quote:
Originally Posted by systematika View Post
Yes you can edit harmonic partials. I think I stopped counting at 32 points before I got bored and I was only about 1/3 of the way through. You can also draw your own waveforms free-hand or "vector" style with lines and curves.

You can also morph partials transwave style and sync that process to LFOs and envelopes, perhaps that will help you to get the subtle variation you're looking for. You can have 4 operators if configured for FM synthesis but it doesn't have to be. (You can use it as subtractive if you want) The FM operators act on their own along with the other oscillators. You can self modulate the FM operator, or you can use the oscillators as the modulation source. It also has multiple outs.

Zebra 2 is like a bastard child of the synclavier, a fizmo, and absynth. It's definitely it's own beast, which is why it's virtually the only thing I have that's *not* a "modular" style plugin interface. Then again, I guess it's still semi-modular but just with a mod matrix. :P
Hmmm... Very interesting. Tell ya what: How about you try making this one? It's not a problem in Absynth, but to do it "right" you have to have two instances of it. You technically could get close in standalone mode, but it's really rough. I mentioned before that Absynth is really harsh sounding when you try to imitate other synths, but it's nothing some careful filtering can't fix. I'm going to try emulating this sound in Absynth using both 1 instance and 2 instances. I'm interested in hearing a comparison--part of which also includes being true to the original sound.

If I'm going to make a "new" MJ sound, what would anyone suggest for a multi-sample version? I'm going to post a screenshot of the numeric view in "Page F." You'll notice that compared to what your ears tell you, the actual numbers on this thing are remarkably simple. Here's how to interpret this page in case you have trouble following it:

You have 4 "partial timbres," which you'll see listed as 1-4 down the left side. For each partial, you have two rows of numbers. One is labeled "V" and the other is "H." V is the volume level, or what we might call the AEG. In FM terms, it's the carrier. H is the harmonic envelope, or what we'll call the modulator. You'll recognize the columns are a familiar DADSR. IDec is "initial decay," or what we think of as "decay; FDec is "final decay," or what we usually call "release." Peak is the loudest point. Sust is obviously the sustain volume. It really does have a good "analog" feel to it! All times are expressed in milliseconds. Peak and Sust are in percentage.

On the right side, you also have two rows of numbers. These are your harmonics, 1-12. I didn't bother making a screenshot of 13-24 because there's nothing there! The values expressed here are a percentage of the full volume. You need to be aware that the Synclav will not output these at less than full volume. If you make a Synclav timbre with only one harmonic and you drop the value here to <5, you will introduce harmonic distortion because the FM synth is only 8-bit. Really cool if you need a true pulse wave!!! You can use FM to get groovy phasing effects. With THIS timbre, though, you don't have to be concerned with that. So just take the percentages there and plug them into your own software, be it Zebra or otherwise. Just to clarify: the first row is harmonics 1-6, the second is 7-12. If you're a dedicated additive synth junky, you'll be disappointed that this isn't any more complex than it is.

So... That will get you started. V=carrier, H=modulator. The harmonics refer to the carrier, NOT the modulator. You cannot set harmonics for the modulator.

It is also possible to set the PHASE of each individual carrier harmonic, also possible in Absynth, but this info does not appear anywhere on the Synclavier terminal. You can only access this from the keyboard. There's no need to worry, though, because I happen to know that no harmonic has had its phase altered.

There are also performance settings that you can only access from the keyboard, typically portamento which is normally turned off by default. Rather than worry about that, though, I'd suggest you look at the performance controls in your own software and manipulate the sounds however you want. What I can tell you is that there are no modulation routings at all for this patch. Any pitch artifacts are due to the effects of FM! What you could do is change the FM sustain level and control the decay with aftertouch, mod wheel, or whatever.

Now, at the bottom of the screen, you have the main parameters for each timbre. I personally found this part a little weird. "Partial" tuning refers to the overall tuning, not a fixed frequency. It IS possible to get fixed frequency on the Synclavier, which you'd recognize because it has a negative(-) value. A setting of exactly 440 would appear as -440.0. Same applies to FM ratio. On the DX7, FM8, and Absynth, this appears as a ratio, with Absynth giving you a few options. To get partial 2, which is displayed on the Synclav as 422.6, I reset the ratio until I got an A that matched the 422.6 fixed frequency on another channel. I wasn't able to nail it, but I came up with the ratio as 0.9605. You could also pitch it down (transpose) to -0.6977.

Volume and FM Rat are fairly straightforward. Chorus might be misleading. To get chorus (unison detuning), the Synclavier doubles the voice cards being used. A value of 1.000 is no chorus at all. The chorus tuning here is expressed as a ratio, so a value of 2 would be an octave. I'm not sure where it goes from there because I prefer not to use it that much. I would ignore the vibrato and stereo settings. They don't do anything!

Vintage Synth pr0n to follow:

[Edit: One minor correction I need to make. In the harmonic envelope, the Peak/Sust values go up to 1000. I have yet to figure out how this scales with other synths. I know that DX-style synths are able to get a lot noisier than the Synclav, while Absynth is a lot more conservative. How to translate these two values into any other synth is anybody's guess.]
Attached Thumbnails
Michael Jackson Beat It Gong-beat-.jpg  

Last edited by AngelRho; 31st March 2011 at 05:09 PM..
Old 31st March 2011
  #45
Gear Head
 

OK...

Just punched in the numbers. Yes, it IS a cool sound, but it is NOT--I repeat--NOT the "Beat It" sound. Personally I think my Absynth patch is pretty cool, but only the product of wishful thinking. There are just enough similarities that my Absynth patch fakes it pretty good.

Part of the problem is there are several chime/bell timbres. I was going for what appeared to be the most inharmonic and complex in that chaotic, noisy kind of way. It was a good guess, but nevertheless wrong. I'll work on more of these tonight and see if I have better luck tomorrow.
Old 1st April 2011
  #46
Lives for gear
Quote:
Originally Posted by AngelRho View Post
Hmmm... Very interesting. Tell ya what: How about you try making this one? It's not a problem in Absynth, but to do it "right" you have to have two instances of it. You technically could get close in standalone mode, but it's really rough. I mentioned before that Absynth is really harsh sounding when you try to imitate other synths, but it's nothing some careful filtering can't fix. I'm going to try emulating this sound in Absynth using both 1 instance and 2 instances. I'm interested in hearing a comparison--part of which also includes being true to the original sound.

If I'm going to make a "new" MJ sound, what would anyone suggest for a multi-sample version? I'm going to post a screenshot of the numeric view in "Page F." You'll notice that compared to what your ears tell you, the actual numbers on this thing are remarkably simple. Here's how to interpret this page in case you have trouble following it:

You have 4 "partial timbres," which you'll see listed as 1-4 down the left side. For each partial, you have two rows of numbers. One is labeled "V" and the other is "H." V is the volume level, or what we might call the AEG. In FM terms, it's the carrier. H is the harmonic envelope, or what we'll call the modulator. You'll recognize the columns are a familiar DADSR. IDec is "initial decay," or what we think of as "decay; FDec is "final decay," or what we usually call "release." Peak is the loudest point. Sust is obviously the sustain volume. It really does have a good "analog" feel to it! All times are expressed in milliseconds. Peak and Sust are in percentage.

On the right side, you also have two rows of numbers. These are your harmonics, 1-12. I didn't bother making a screenshot of 13-24 because there's nothing there! The values expressed here are a percentage of the full volume. You need to be aware that the Synclav will not output these at less than full volume. If you make a Synclav timbre with only one harmonic and you drop the value here to <5, you will introduce harmonic distortion because the FM synth is only 8-bit. Really cool if you need a true pulse wave!!! You can use FM to get groovy phasing effects. With THIS timbre, though, you don't have to be concerned with that. So just take the percentages there and plug them into your own software, be it Zebra or otherwise. Just to clarify: the first row is harmonics 1-6, the second is 7-12. If you're a dedicated additive synth junky, you'll be disappointed that this isn't any more complex than it is.

So... That will get you started. V=carrier, H=modulator. The harmonics refer to the carrier, NOT the modulator. You cannot set harmonics for the modulator.

It is also possible to set the PHASE of each individual carrier harmonic, also possible in Absynth, but this info does not appear anywhere on the Synclavier terminal. You can only access this from the keyboard. There's no need to worry, though, because I happen to know that no harmonic has had its phase altered.

There are also performance settings that you can only access from the keyboard, typically portamento which is normally turned off by default. Rather than worry about that, though, I'd suggest you look at the performance controls in your own software and manipulate the sounds however you want. What I can tell you is that there are no modulation routings at all for this patch. Any pitch artifacts are due to the effects of FM! What you could do is change the FM sustain level and control the decay with aftertouch, mod wheel, or whatever.

Now, at the bottom of the screen, you have the main parameters for each timbre. I personally found this part a little weird. "Partial" tuning refers to the overall tuning, not a fixed frequency. It IS possible to get fixed frequency on the Synclavier, which you'd recognize because it has a negative(-) value. A setting of exactly 440 would appear as -440.0. Same applies to FM ratio. On the DX7, FM8, and Absynth, this appears as a ratio, with Absynth giving you a few options. To get partial 2, which is displayed on the Synclav as 422.6, I reset the ratio until I got an A that matched the 422.6 fixed frequency on another channel. I wasn't able to nail it, but I came up with the ratio as 0.9605. You could also pitch it down (transpose) to -0.6977.

Volume and FM Rat are fairly straightforward. Chorus might be misleading. To get chorus (unison detuning), the Synclavier doubles the voice cards being used. A value of 1.000 is no chorus at all. The chorus tuning here is expressed as a ratio, so a value of 2 would be an octave. I'm not sure where it goes from there because I prefer not to use it that much. I would ignore the vibrato and stereo settings. They don't do anything!

Vintage Synth pr0n to follow:

[Edit: One minor correction I need to make. In the harmonic envelope, the Peak/Sust values go up to 1000. I have yet to figure out how this scales with other synths. I know that DX-style synths are able to get a lot noisier than the Synclav, while Absynth is a lot more conservative. How to translate these two values into any other synth is anybody's guess.]
If I get some time, I'll start patching. I'm in the middle of an album mix atm and the release party is less than a week. I'm to death right now!

Can you give me some more information on how the harmonics work, ie starting/ending partial plus spacing, etc. If it starts at the base partial, no worries should be easy to transcribe into something useful.

Perhaps if we started with a blank synclav patch (any of the other guys who have this wanna do this?) and did a wavetable sample of each shaping parameter gradually we could measure the spectrum and/or table FFT numbers and have a better understanding of exactly what it's doing in a given state and the harmonic distortion involved.

Could turn into an interesting max patch/whole different beast and shared here even.

Last edited by systematika; 1st April 2011 at 05:55 AM..
Old 2nd April 2011
  #47
Gear Head
 

Quote:
Originally Posted by systematika View Post
If I get some time, I'll start patching. I'm in the middle of an album mix atm and the release party is less than a week. I'm to death right now!

Can you give me some more information on how the harmonics work, ie starting/ending partial plus spacing, etc. If it starts at the base partial, no worries should be easy to transcribe into something useful.
Well, it's really quite simple. The Synclav uses something called "timbre frames," which is just like what you get with, say, a Wavestation "wavesequence." You create the sequence by either dialing in what you want or by analysis/resynthesis of a sample (much more complicated). One frame crossfades into another according to parameters you set, though it's automatic for resynthesis.

If we were getting into wave sequences, I'd need MANY more screenshots!

But since the patch I posted, albeit the wrong one, doesn't use any timbre frames, we don't have to worry about it.

The Numeric display that I posted is all the info you need for THIS patch. The harmonics are in order, 1-12. You can get up to 24, and that's it. I only posted 1-12 because that's all you need. Not sure what you mean by the spacing. The Synclavier was the first of its kind, so the lingo is a lot different than what we're used to. "Partial" is analogous to a Korg "Program" and "Timbre" is like Korg "Combi."

The way the harmonics (harmonic coefficients) work is basically the way it looks onscreen. Harmonic 1 is the fundamental, and 2-24 just follow the overtone series. If that's what you mean by "base partial," then you get the idea. It shouldn't be a difficult patch at all.

I wouldn't worry about the more nuanced aspects of Synclav synthesis. It's 8-bit. So a "blank" Synclav patch is nothing but a sine wave. If I understand correctly, different pitches are produced by varying the sample rate. This means you're going to hear a LOT of harmonic distortion on lower notes, while the distortion gets "filtered out" of the high notes. Decrease the level of the lowest harmonics, and you'll introduce even more distortion until you get a square/pulse wave. The only thing I can think of that would simulate this effect is set up an individual waveshaper on every single harmonic, so what you'd have are 24 individual sine waves that can act as FM carrier waves, each paired with its own bit-crushing waveshaper. Now, there is NEVER any added harmonic distortion when the combined OUTPUT passes through the aeg. This is because the volume control on each voice is analog. A bit-crusher would only increase the level of distortion and noise floor as the sound fades. But this does not happen on the Synclavier. It's one of the great reasons to use one.

I bet you could simulate this in Max. I'd love to see that!

The guys left over from the NED days HAVE worked on a VST/AU plugin. They are just unsatisfied with the lack of commercial interest in it.
Old 2nd April 2011
  #48
Lives for gear
Quote:
Originally Posted by AngelRho View Post
Well, it's really quite simple. The Synclav uses something called "timbre frames," which is just like what you get with, say, a Wavestation "wavesequence." You create the sequence by either dialing in what you want or by analysis/resynthesis of a sample (much more complicated). One frame crossfades into another according to parameters you set, though it's automatic for resynthesis.

If we were getting into wave sequences, I'd need MANY more screenshots!

But since the patch I posted, albeit the wrong one, doesn't use any timbre frames, we don't have to worry about it.

The Numeric display that I posted is all the info you need for THIS patch. The harmonics are in order, 1-12. You can get up to 24, and that's it. I only posted 1-12 because that's all you need. Not sure what you mean by the spacing. The Synclavier was the first of its kind, so the lingo is a lot different than what we're used to. "Partial" is analogous to a Korg "Program" and "Timbre" is like Korg "Combi."

The way the harmonics (harmonic coefficients) work is basically the way it looks onscreen. Harmonic 1 is the fundamental, and 2-24 just follow the overtone series. If that's what you mean by "base partial," then you get the idea. It shouldn't be a difficult patch at all.

I wouldn't worry about the more nuanced aspects of Synclav synthesis. It's 8-bit. So a "blank" Synclav patch is nothing but a sine wave. If I understand correctly, different pitches are produced by varying the sample rate. This means you're going to hear a LOT of harmonic distortion on lower notes, while the distortion gets "filtered out" of the high notes. Decrease the level of the lowest harmonics, and you'll introduce even more distortion until you get a square/pulse wave. The only thing I can think of that would simulate this effect is set up an individual waveshaper on every single harmonic, so what you'd have are 24 individual sine waves that can act as FM carrier waves, each paired with its own bit-crushing waveshaper. Now, there is NEVER any added harmonic distortion when the combined OUTPUT passes through the aeg. This is because the volume control on each voice is analog. A bit-crusher would only increase the level of distortion and noise floor as the sound fades. But this does not happen on the Synclavier. It's one of the great reasons to use one.

I bet you could simulate this in Max. I'd love to see that!

The guys left over from the NED days HAVE worked on a VST/AU plugin. They are just unsatisfied with the lack of commercial interest in it.
Oh okay, thanks for the info! I just wanted to be more accurate, that is all. I'm pretty sure I understand fully now. I'll experiment over the next couple of days.

I just wanted to make sure that it wasn't like 1st, 3rd, 6th, or something crazy like that. (i've seen synths only use odd or even harmonics before) You never know with those funky engineers back east. Now that I've read over your post again I realize I'm blind and you said already haha my bad. I guess that's what I get for going 30+ hours with only 3 hours of sleep. Welcome to the world of music.

I've never used a synclavier before but it's always been an interesting device to read about history wise. I mostly know that it's FM, and the fact that it's analog but that's it. :P

As for the waveshaper, yeah that's how you'd have to do it.
Old 2nd April 2011
  #49
Gear Head
 

Quote:
Originally Posted by systematika View Post
Oh okay, thanks for the info! I just wanted to be more accurate, that is all. I'm pretty sure I understand fully now. I'll experiment over the next couple of days.

I just wanted to make sure that it wasn't like 1st, 3rd, 6th, or something crazy like that. (i've seen synths only use odd or even harmonics before) You never know with those funky engineers back east. Now that I've read over your post again I realize I'm blind and you said already haha my bad. I guess that's what I get for going 30+ hours with only 3 hours of sleep. Welcome to the world of music.

I've never used a synclavier before but it's always been an interesting device to read about history wise. I mostly know that it's FM, and the fact that it's analog but that's it. :P

As for the waveshaper, yeah that's how you'd have to do it.
Those funky engineers? lol Yeah!

The truth is that they set out to build a do-anything machine and succeeded. I only intended to get into the composing business, and I'm not seeing a lot of success at this late point in my career. But I do enjoy doing what I do despite being dirt poor. The way I look at it, there were hundreds, maybe even thousands, of composers active back in Mozart's day. So why is it Mozart is the only one you ever hear about, right? There are better reasons to work at it than money--but the money sure would be nice!!!

Anyway, without someone doing a LOT of work, there's never going to be a virtualized Synclavier. I wish I knew the trick to getting that HUGE sound! The patch I posted, you'll notice, is monaural. There's no panning or anything. But if you play a big arpeggio fairly quickly so that you get a wash of sound from the final decay transient, it really fills up the room! It's like it refuses to let any harmonics from one note to the next phase cancel or something. Too bad the Synclav doesn't have any other kinds of synthesis tricks! As primitive as it might seem, there'll never be anything that will do what it does.

Commercially, no one cares about it anymore. I could see NI selling a Synclavier product, but I bet it would only be a FM sample library and not an emulation. There's still a lot of mistrust because the NED guys screwed over so many people, and people in the business want nothing to do with it. There's no reason at all why someone couldn't miniaturize the voice cards and package it in something like Kyma, which would mean less CPU usage on the host computer. They already have the hardware to control the cards directly from a Mac. I personally wouldn't object to coughing up as much as $500 for a VST/AU version, or up to $2000 for the hardware. What's more, I still wouldn't even give up my Synclav, even if I had all that. The main problem is there's no money forthcoming in all this.

I'd been following the Synclavier.com website for a few years after I got out of college. Once a year, something new would be posted. I got my hopes up when it said there was a plugin in development. One year passed. Two years passed. So I finally emailed Cameron and asked him "what gives?" I think it was maybe about 3 months after I got a reply, which surprised me that I even got a reply at all, and the website was updated to omit any mention of the plugin. I almost cried, too! Cameron is a nice guy. But he's not getting any younger. The last I heard anything through the grapevine about Cameron was when the economy tanked, he ran off to Canada.

The world needs a synclav plugin, I think. But if you want it soon, you're going to reverse engineer it. Otherwise, it shows off your synth programming ability if you can emulate it on something we already have access to. I don't have Max, but that would be the perfect environment to get started on it!
Old 3rd April 2011
  #50
Lives for gear
 
acreil's Avatar
 

Quote:
Originally Posted by AngelRho View Post
I wouldn't worry about the more nuanced aspects of Synclav synthesis. It's 8-bit. So a "blank" Synclav patch is nothing but a sine wave. If I understand correctly, different pitches are produced by varying the sample rate. This means you're going to hear a LOT of harmonic distortion on lower notes, while the distortion gets "filtered out" of the high notes. Decrease the level of the lowest harmonics, and you'll introduce even more distortion until you get a square/pulse wave. The only thing I can think of that would simulate this effect is set up an individual waveshaper on every single harmonic, so what you'd have are 24 individual sine waves that can act as FM carrier waves, each paired with its own bit-crushing waveshaper.
It's a little simpler than that. You can sum the harmonics to a single cycle waveform, then quantize the amplitudes. The other issue is the image frequencies. The harmonics will repeat at higher frequencies, and these will be audible at lower pitches (distinct from the harmonic distortion resulting from the quantization). You can reproduce this just by using a low resolution waveform (I think it's 256 samples in this case) with drop sample interpolation if you have a high enough sample rate. The PPG Wave does this and doesn't have a lot of audible aliasing. To do this in software you'd just have to oversample. BUT FM makes this a lot harder since it generates a lot of non-band limited harmonics on top of that. I suspect you could synthesize high resolution, band limited multisamples containing the image frequencies, and then oversample for decent sounding FM.

It seems that if you can set the individual phases of each harmonic, and if low harmonic amplitudes have increased quantization error, that it's not doing any special normalization or scaling before it synthesizes the waveforms. But for 24 harmonics, that gives seemingly about 3 bits of resolution to each harmonic. Maybe the default mode somehow optimizes by shifting or inverting the phases of some of the harmonics (to reduce peak amplitude), and setting the phases manually either permits clipping or reduces total resolution.

Some really good measurements or samples would be helpful here. Maybe the patents say something about it...
Old 3rd April 2011
  #51
Lives for gear
 

Someone lend me a Prophet 5 and I will come up with it for you.
Old 5th April 2011
  #52
Gear Head
 

Quote:
Originally Posted by Franc View Post
Someone lend me a Prophet 5 and I will come up with it for you.
Meh... But if I ever need the "Jump" sound from 1984 and can't find an OB, I'll be sure to let you know!

Working on other things at the moment, but Thursday what I'll do is sample some sine waves. I'm not going to bother taking velocity multi-samples because the output envelopes are analog and thus it won't make a bit of difference. I was messing around with a new program just yesterday morning and heard an interesting artifact. Setting the FM amount, which would be in the harmonic envelope under "peak," I could actually "hear" the synclavier stepping through the FM levels. It's less noticeable at increased levels, which is no surprise. But on most other FM synths you'd find this annoying--well, I would, anyway. I'm wondering how useful this might be...

Anyway, I'll take several samples of just a sine wave so you see the correlation between the numbers and the actual result. The FM sampling becomes more accurate as the level increases--which is perfectly normal considering properties of 8-bit sound. Decrease the levels and the quantization error becomes obvious to the point it's not even a sine wave anymore. When only one harmonic is concerned, the output is full volume no matter how low you try to set it, except 0=0, hence some interesting and useful variants of pulse waves. This is unusual because you expect the volume would decrease in addition to quality. But rather the Synclav adjusts proportionally by the loudest harmonic. I made this really nasally clavi sound, and it really does sound like an unfiltered analog sound. If you need saw waves or triangle waves, just use a typical additive-synthesis formula for constructing them. I'll get some samples and screenshots up Thursday.
Old 5th April 2011
  #53
Lives for gear
 
acreil's Avatar
 

Quote:
Originally Posted by AngelRho View Post
Anyway, I'll take several samples of just a sine wave so you see the correlation between the numbers and the actual result.
What would also be helpful is a waveform with all harmonics turned up to maximum.
Old 6th April 2011
  #54
Lives for gear
I started on a maxforlive patch today, got the harmonics generator, and the envelopes done... Now all I need to do is get the parameter math right and add the other stuff now that I have the core generator working.

Two questions... What should the numbers reference to? Percentages? Do you know?

Should the decimator process the whole partials mix, or the individual operators?
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
- Michael P - / Electronic Music Instruments and Electronic Music Production
37
W-W-Int / Electronic Music Instruments and Electronic Music Production
8
AllAboutTone / For those we have lost. Memorials, RIPs and Obituaries
384

Forum Jump
Forum Jump