VSTI world hate MIDI GUITAR?

[AdmiralQuality]

More evidence... the Portamento only works when it's MIDI driven. If it was tracking, it would be a simple matter to fade from one CV to the next as new notes were detected. In fact, it would be the EXACT same "code" used for the MIDI note driven sounds.

I'm not even going to get into the awful aliasing this puppy makes.

Again, I'm glad you like it and have fun with it. More power to ya. But I'm pretty damn sure at this point, after forming hypotheses and testing them with experiments, that I'm right in my guess as to how Tripp Lead works.

[guitarzan]

You'll find out soon that the FS Synths track WAY better than Roland MIDI out is ever going to get you. Axon is better - but there's a reason nobody REALLY likes MIDI guitar ,you know.

These FS synths are made to play like a guitar - so a lot of synthy stuff is probably left out. I've written to FrettedSynth, maybe he'll bother, maybe not. Why not try some of the other three synths? The Phase Distortion Synth v1.1 is the newest.

[AdmiralQuality]

Again, they don't "track" at all. They're effects in the guise of synths. So of COURSE they "track" better, they're just effecting the input exactly as entered. Not tracking then syntheszing from scratch, as a true guitar synth does. (I'm not putting this method down, just saying they're different things. People love their Roland VG's and they probably do something not dissimilar.)

I'm going to assume that the other 3 synths use the same architecture, and not waste my whole night analyzing something that isn't a product I want, nor competition for a product I'm making. But it was fun though. And yes, lets wait to hear what FS has to say on the subject.

Cheers.

[arakula]

(This is actually probably not unlike some of the stuff the roland VG's do! Of course, the VGs can do it for each string, polyphonically. And their pickup goes as close to the bridge as possible, as they have smarter fundamental frequency determination logic.)

Yeah well, from what I know, the VGs don't even use pitch to CV but deal with the individual pickup signals straight ahead (so it's indeed just a 6x FX unit).

That is correct. I happen to own both VG-8EX and VG-88V2. They are absolutely great FX units in the style of a Line 6 VariAX. Some sounds that you can produce with them:

http://www.hermannseib.com/audio/hail128.mp3 is all done with a Stratocaster and the VG-8 (except for a little Boom-Chack from Steinbergs LM-7). Never mind the playing (I'll never be a really good or fast guitar player), but all the sounds are VG-8 factory sounds.

http://www.hermannseib.com/audio/ghosts.mp3 is also Strat and VG-8. This bears little resemblance with a guitar any more.

Anyway - neither of these devices converts anything to MIDI or some other kind of pitch/loudness information. They are polyphonic digital effect devices with a latency close to zero. Just the pitch changer lags a bit - this one has to do the "what the *** is he playing?" calculations.

[Sascha Franck]

arakula, that "ghosts" thing is nice. Reminds me why (at least in principle) I always wanted one of the VG models, so I could use such sounds live without the troubles a MIDI guitar is giving me. I played around with them a lot in shops and it's just amazing that you don't even have to alter your technique at all. Responds great to slides, slurs, bends and whatever.
Seriously, it's always been the fact that you have to use a specialized guitar to be up and running. I couldn't imagine throwing additional electronics into my main live guitars (basically a Tom Anderson, a G&L and a very old Ibanez "pre-Artists" 335 clone). So I may have to buy one of the Godins or so - which deliver way better pickup performance anyways, due to them build straight into the bridge saddles. Unfortunately I can't use those replacement bridge saddles from Fishman and the likes either as I already have Graph Techs on the mentioned axes (no more string breakage ever!).

And uhm, the GK pickup... time for a little rant. It might be that they changed a bit in the design of the GK 3 (I own a GK 2a) but it still looks pretty bad. Why in the world do you *have* to have the cable coming out at the top (at your low E string that is)? That's where all the forearm and wrist action is taking place. It'd be WAY better if they allowed for the cable to come out of the bottom end as well. If they allowed for such a thing, they could probably also shorten the damn device. I mean, when you try to dampen the bottom strings with your wrist, there's no way you don't touch the sort of sharp-ish end (at least in case you mount it properly, so it's as close as possible to the bridge.
I truly hate those GK pickups - and they also make each and every guitar look bad. I would never put them onto any of my fine instruments (I'm using mine on a fugly looking old Ibanez Sabre, which I never use live).


Ah, the Fretted Synth issue: While it does indeed feel as if it'd do tracking (with all the typical glitches when you don't play carefully), AQ (even if presented in his own, shall we say "special" way) made some observations that do indeed seem to prove it's not actually tracking. If it was tracking, apart from probably better or worse tracking, changing the timbre of the input signal wouldn't change the timbre of the synth output. But it clearly does. The same goes for the release parameter. If it was a real OSC being triggered, the release would actually follow the parameter setting rather than what you're playing. But it doesn't. When you stop a note, the thing is producing no output anymore. A tracked synth however would.
Further, even if that's no true prove, I always thought that a tracking synth would work best in mono mode. With the Fretted Synth you can easily play chords as well.
After all, it really looks more like an FX unit than a synth. Not that it wouldn't have it's uses, but I don't see any CV tracking to happen either.

[nix808]

Hey again midi guitarists(and interested parties),
Allans Music in Melbourne gets a plug. They were very nice about exchanging a GR20 for a GI20. So now I can try to get this unit tracking adequately. The first thing I have found was that the
chromatic type was set to 2 originally. Using this mode I was able to do slides, although a new attack was introduced for every semitone slid. This mode was introducing a second note after I had played a first, also, without sliding. I set the chromatic to type 1, and this made playing a lot more similar to the GR20. I haven't been playing it and changing settings long enough to know whether I'm happy with it, but it seems quite good. I like the versatlity of midi guitar, being able to play whatever. Anyway, so far so good. I didn't notice any difference running through USB or midi.
Kind regards, Nix

[krraqk]

OH and please note, the FrettedSynth things AREN'T tracking. They're filtering, distorting, and pitch shifting the ORIGINAL WAVEFORM. Then applying envelopes (basically triggered envelope "followers". All stuff you could do with regular guitar effects! I don't have it, but I bet you could do most of what they do with G.R. They have octavers, pitch shifters, envelope followers, don't they? Sure, it might not be as integrated and "to the point", but again, don't be fooled, they're NOT pitch tracking. There is no "CV" stage in them, the output wave is just a distortion of the input wave, and then pitch shifted to get whatever interval you select (which again is the process that I think is introducing the bad latency I hear. Without the pitch shift this technique SHOULD be very responsive! )

And who cares about what is the system or technique used? I dont for sure. But you can play a totally different sound (a synth, better or worse, but a synth with many typical synth parameters) and with a very natural playing technique. Feels, more or less and probably more than a midi guitar in some situations, like an instrument.

Maybe you can do some radical sound modifications with lots of effects, but normally you don't get a defined synth sound (in the classic sense of the term) wich can play melody, etc, in general that's for other tasks I think, at least there is not a product created and advertised for this purpose that I know.

So GOOD for this approach, please developer do more in this area, we are a LOT of guitarist in the world waiting for tools like that.

[nix808]

Yeah, there are a plenitude of guitarists. If Fretted Synth made a retail plug, on the strength of their free ones it would be something to behold. I've been messing around with Synth Maker, you can run an audio signal through it's oscillators, but it surely doesn't approach what FSA has done. Still so far so good, the GI20 might be my tool of choice.
Cheers, Nix

[kevink]

??

[kevink]

Ah, the Fretted Synth issue: While it does indeed feel as if it'd do tracking (with all the typical glitches when you don't play carefully), AQ (even if presented in his own, shall we say "special" way) made some observations that do indeed seem to prove it's not actually tracking. If it was tracking, apart from probably better or worse tracking, changing the timbre of the input signal wouldn't change the timbre of the synth output.

Unless, that is, changing the timbre coincided with changing the average signal amplitude, and the input signal amplitude was in turn being used to modulate filters, as some synths can with MIDI velocity. Far from certain, but plausible.

But it clearly does. The same goes for the release parameter. If it was a real OSC being triggered, the release would actually follow the parameter setting rather than what you're playing. But it doesn't. When you stop a note, the thing is producing no output anymore. A tracked synth however would.

Not necessarily. If it's an ADSR-type envelope, and the virtual "gate" is active so long as you've got a note coming in to the device, you'd get the signal at the "S" level until you released the gate, which sounds like what you're describing.

Further, even if that's no true prove, I always thought that a tracking synth would work best in mono mode. With the Fretted Synth you can easily play chords as well.

Now *that* is what I'd call evidence that it's not really tracking. Pulling chords out of an audio signal is damned difficult and I have a hard time believing that a PC application could do it in real time, at least not with the CPUs that existed when those plugs were first published. To be clear, when you play a chord in, you hear a chord coming out played by the "synth", or do you here a mono voice that corresponds to the root or lowest note of the chord?

I'm not set up to do this myself, but another good experiment would be to run a vocal into it. If one hears formants and detailed consonnants, it's a very strong indication that the "synth" isn't synthesizing.

Which one of their synths are you doing this experiment on, anyway? There are four of them on the web page.

[Sascha Franck]

Not necessarily. If it's an ADSR-type envelope, and the virtual "gate" is active so long as you've got a note coming in to the device, you'd get the signal at the "S" level until you released the gate, which sounds like what you're describing.

Well, I tried with all ADSR settings, including S and R all the way up. As soon as the input stops, the sound stops as well

Now *that* is what I'd call evidence that it's not really tracking. Pulling chords out of an audio signal is damned difficult and I have a hard time believing that a PC application could do it in real time, at least not with the CPUs that existed when those plugs were first published. To be clear, when you play a chord in, you hear a chord coming out played by the "synth", or do you here a mono voice that corresponds to the root or lowest note of the chord?

You can clearly hear the complete chord (or what's left of it after being mangled so much).

I'm not set up to do this myself, but another good experiment would be to run a vocal into it. If one hears formants and detailed consonnants, it's a very strong indication that the "synth" isn't synthesizing.

Tried that and you can barely understand a word (due to the massive processing), yet, it's almost clear to me that there's no tracking going on.
And btw, so far I tried with the PhaseOsc only.

[guitarzan]

Further, even if that's no true prove, I always thought that a tracking synth would work best in mono mode. With the Fretted Synth you can easily play chords as well.
After all, it really looks more like an FX unit than a synth. Not that it wouldn't have it's uses, but I don't see any CV tracking to happen either.

You must be talking about the "synth" section in FreeAmp2 - that one is not tracking and is basically an effect. The four Fretted Synth Audio Triggered Synths are entirely different animals, and you absolutely must play clean mono lines - no exceptions, and they all have multiple oscillators.

EDIT: I read now where you used PhazOsc - all I can say is that pitch to CV is going to track differently than pitch to MIDI - better. To me that explains all the differences everyone has brought up - you're not indirectly triggering commands, the signal is directly converted into a virtual control voltage, so it'll trigger off any sound and it will only last as long as the signal is present (I'm sure it could be extended with some kind of manipulation). No note-on or note-off, much more direct than all that crap. It'll also contain much more of the detail of the original signal. NO MIDI, different, more like the good tracking old pre-MIDI Rolands.

I can almost see where AQ might think that the Tripp Lead synth is done with distortion and pitch shift because it is a lead synth and all the presets are distorted, but if you check out, say, the String Sculptor v2.0 I think it's obvious that it's tracking and using oscillators.

[Sascha Franck]

guitarzan, really, it seems there's no tracking happening in the PhaseOsc. As said, you can even play chords. Just try for yourself.

[guitarzan]

Read about control voltage, it works differenty. It's trying to play whole chords because it converts the whole chord into a virtual CV and does the best it can with it. SythEdit syths work internally with CV - not MIDI. It's a happy odd convergence for guitar player who want to play synth. Check out this page about the Roland GR-300, an analog pitch to CV synth. This is how guitar synths should be triggered. More developement is needed because pitch to MIDI just isn't going to cut it. I hope FrettedSynth jumps in here - I think he has a day job, though

[AdmiralQuality]

FS's "synths" are NOT bloody "control voltage" guitarzan. I spent an hour proving that last night.

THERE IS NO DIFFERENCE BETWEEN "MIDI" and "CV" (at least not in the digital realm). What do you think a VSTi gets from the host to tell it what note of the keyboard you're playing?

MIDI WAS INVENTED TO REPLACE CV and GATE signals that analog synths used to use -- because it was getting unweildy to have 16 signals required to hook an 8 voice synth up... and manufacturers were starting to make their OWN proprietary digital interfaces... Roland's was called DCB, was on the Jupiter 8... a big fuckin Centronix connector like a SCSI cable... in 1983 the manufacturers got together and made MIDI to standardize an interface.

Gah!