 |
 |
 |
 |
 |
 |
 |
Forums - Synths & synthesis
Subject: Wavetable Synths vs Modeling Synths
Original Message 1/29 17-Nov-02 @ 05:00 PM - Wavetable Synths vs Modeling Synths
is there a difference? whats the difference? i know the XTk is based on wavetable technology, but how does that differ from modeling?
Message 2/29 17-Nov-02 @ 07:43 PM - RE: Wavetable Synths vs Modeling Synths
wavetables are basically samples of one or a couple cycles of a repeating waveform. These short samples are then looped and used as an oscillator in a synth. The waveforms are often more complex in nature compared to the oscillators of virtual analogs.
VA's (modeling) actually generate their sine, saw, and square waves digitally in realtime, which allows you to do things you can't do with a sample (like true pulse width modulation, PWM, or varying the gate time of a square wave). It's a digital model of a simple analog oscillator.
Now the thing that's tricky, and maybe someone else can help me here, but the XTk can modulate smoothly between its wavetables, right? Does that mean it atually generates those waveforms in realtime? Or is it a crossfade thing? XT owners?
psylichon
VA's (modeling) actually generate their sine, saw, and square waves digitally in realtime, which allows you to do things you can't do with a sample (like true pulse width modulation, PWM, or varying the gate time of a square wave). It's a digital model of a simple analog oscillator.
Now the thing that's tricky, and maybe someone else can help me here, but the XTk can modulate smoothly between its wavetables, right? Does that mean it atually generates those waveforms in realtime? Or is it a crossfade thing? XT owners?
psylichon
Message 3/29 17-Nov-02 @ 08:20 PM - RE: Wavetable Synths vs Modeling Synths
ok. so does that mean on a wavetable synth you cannot actually edit the wave itself, but the frequency(for lack of a better word) that comes out instead?
wouldnt that kinda hender your creativity a little bit? maybe im off base here... ??
wouldnt that kinda hender your creativity a little bit? maybe im off base here... ??
Message 4/29 17-Nov-02 @ 08:29 PM - RE: Wavetable Synths vs Modeling Synths
well, we're talking about single-cycle waveforms here, so any slight alteration of the wave will greatly affect harmonic spectrum. That's why there are usually 64 or so of them, to give that variety.
But think about VA oscillators... there are only so many things you can do to shape the actual waveform of the oscillator. It's basically gonna be a sine, tri, saw, or square, or something in between. As far as osc-level mods, ya got pulse width and skew and softness... about it. Most of the excitement in analog synthesis comes from interaction between oscillators... sync, ring mod, detuned unison... and from the filter section which has nothing to do with the oscillators.
So a wavetable synth with, say, 64 tables in it is actually quite flexible even without the addition of user-tables. Of course synths that offer true waveshaping are quite dope as well (a'la Absynth).
psylichon
But think about VA oscillators... there are only so many things you can do to shape the actual waveform of the oscillator. It's basically gonna be a sine, tri, saw, or square, or something in between. As far as osc-level mods, ya got pulse width and skew and softness... about it. Most of the excitement in analog synthesis comes from interaction between oscillators... sync, ring mod, detuned unison... and from the filter section which has nothing to do with the oscillators.
So a wavetable synth with, say, 64 tables in it is actually quite flexible even without the addition of user-tables. Of course synths that offer true waveshaping are quite dope as well (a'la Absynth).
psylichon
Message 5/29 18-Nov-02 @ 05:20 AM - RE: Wavetable Synths vs Modeling Synths
it is important to know the sound is extremely diferant. The wavetable is cabable of very buzzy digital sounds, as well as doing simple square, or saw waves that sound very analog. The waldorf digital filter is best sounding digital filter i have ever heard and used. If you were to buy a waldorf xt, and a waldorf q, the sounds would be like night and day.
The analog modeling synth will have a softer rounder and well, analog sound. Using a wavetable, wavesequence, or vector synth as they are all the same, will yeild some very diferant sounds.The wavetable is still warm because it's going through a analog filter. The original microwave has an analog filter, but the sounds can be made so much harsher from analog to totaly strange digital growling and shreeking by manipulating the start of the sequence and modualtiing the speed with with you go through the 64 cycles. The first 30 are the same as the waldorf ppg wave. You can download the demo of that VSTi and listen to the tones. But the XT has the nice filter and analog waveforms, but be warned, The knobs are velocity sensative. So the quicker you turn them they jump forward. It was almost impossible to tweek fast in kind of accurate way.
The analog modeling synth will have a softer rounder and well, analog sound. Using a wavetable, wavesequence, or vector synth as they are all the same, will yeild some very diferant sounds.The wavetable is still warm because it's going through a analog filter. The original microwave has an analog filter, but the sounds can be made so much harsher from analog to totaly strange digital growling and shreeking by manipulating the start of the sequence and modualtiing the speed with with you go through the 64 cycles. The first 30 are the same as the waldorf ppg wave. You can download the demo of that VSTi and listen to the tones. But the XT has the nice filter and analog waveforms, but be warned, The knobs are velocity sensative. So the quicker you turn them they jump forward. It was almost impossible to tweek fast in kind of accurate way.
Message 6/29 18-Nov-02 @ 02:24 PM - RE: Wavetable Synths vs Modeling Synths
most of what's been said is true, but ignores the original purpose of wavetables. you CAN use them as just 'static harmonically rich waves', but that's not what they were designed for.
wavetables were originally done as a way to implement in the digital realm what could only be done in the analog world. remember, wavetables date back way before VAs, to the PPG wavecomputer 360. there were no digital filters etc. at the time. (and the wavecomputer 360 had no filter at all! the later PPG wave did have an analog filter).
the idea of wavetable synthesis is basic: you create a series of waves that implement a sort of 'harmonic series'. then you scan through it. for example, think of a filter sweep (to keep it simple, we'll assume no resonance). when the filter is almost closed, you just get a sine wave. as you open the filter, more harmonics appear, until it's wide open and all the harmonics are there.
ok, so what we do is we create a series of samples, one with the filter all the way closed (sine wave), the next with it a little bit open, the next a little more open, and so on up until we have a wave for all the way open. now, by scanning across the waves as oscillators, we can 'simulate' the filter sweep - instead of using a filter to shape the harmonic structure of the oscillator, we modify the oscillator to do the harmonic morphing directly.
but we aren't limited to filter sweeps, i can choose any harmonic structure for each wave... i could sweep between a sawtooth and a square for instance, or create a series of vocal formants and make it 'talk'. basically, by choosing your waves correctly, you can implement just about _any_ series of waveforms... you could possibly implement a flute blown at different breath strengths... a DX7 patch with varying modulation amount... etc. you only have a SINGLE axis to work with though in most wavetable synths, so you couldn't for example layout one axis for frequency, and another for resonance (though, with a 2D wavetable, such a thing is possible). wavetable though is about making the sound 'move' in the way you want it to, and not be limited to just the changes a filter makes.
the downside is that the wavetable gets 'quantized'; depending on how fast you are doing 'harmonic morphing', i.e. how different each sample is from the next, this might not even be audible, or it might be nasty. of course that's part of the charm of wavetable synthesis
the XT uses 61 waves in its wavetable; the ensoniq ASR samplers use 128 (afaik); with more waves in the wavetable, theoretically it can be a lot more smooth (as well as the ensoniqs are 16 bit); but the XT has a lot of other advantages of the ASRs. and you can do so much with filters, sometimes it takes a bit of imagination to 'go outside the box' - it's a bit like additive in that regard.
different from VAs? even VAs sound different. but a wavetable synth SHOULD be a lot more powerful, programming-wise, than a VA. which one you like better though is a matter of opinion.
wavetables were originally done as a way to implement in the digital realm what could only be done in the analog world. remember, wavetables date back way before VAs, to the PPG wavecomputer 360. there were no digital filters etc. at the time. (and the wavecomputer 360 had no filter at all! the later PPG wave did have an analog filter).
the idea of wavetable synthesis is basic: you create a series of waves that implement a sort of 'harmonic series'. then you scan through it. for example, think of a filter sweep (to keep it simple, we'll assume no resonance). when the filter is almost closed, you just get a sine wave. as you open the filter, more harmonics appear, until it's wide open and all the harmonics are there.
ok, so what we do is we create a series of samples, one with the filter all the way closed (sine wave), the next with it a little bit open, the next a little more open, and so on up until we have a wave for all the way open. now, by scanning across the waves as oscillators, we can 'simulate' the filter sweep - instead of using a filter to shape the harmonic structure of the oscillator, we modify the oscillator to do the harmonic morphing directly.
but we aren't limited to filter sweeps, i can choose any harmonic structure for each wave... i could sweep between a sawtooth and a square for instance, or create a series of vocal formants and make it 'talk'. basically, by choosing your waves correctly, you can implement just about _any_ series of waveforms... you could possibly implement a flute blown at different breath strengths... a DX7 patch with varying modulation amount... etc. you only have a SINGLE axis to work with though in most wavetable synths, so you couldn't for example layout one axis for frequency, and another for resonance (though, with a 2D wavetable, such a thing is possible). wavetable though is about making the sound 'move' in the way you want it to, and not be limited to just the changes a filter makes.
the downside is that the wavetable gets 'quantized'; depending on how fast you are doing 'harmonic morphing', i.e. how different each sample is from the next, this might not even be audible, or it might be nasty. of course that's part of the charm of wavetable synthesis
the XT uses 61 waves in its wavetable; the ensoniq ASR samplers use 128 (afaik); with more waves in the wavetable, theoretically it can be a lot more smooth (as well as the ensoniqs are 16 bit); but the XT has a lot of other advantages of the ASRs. and you can do so much with filters, sometimes it takes a bit of imagination to 'go outside the box' - it's a bit like additive in that regard.different from VAs? even VAs sound different. but a wavetable synth SHOULD be a lot more powerful, programming-wise, than a VA. which one you like better though is a matter of opinion.
Message 7/29 19-Nov-02 @ 03:32 AM - RE: Wavetable Synths vs Modeling Synths
you can do it on my EMU too if you have time to set up the realtime layers of 128 diferant cycles, and then you use a lfo to modulate them or manulay pan between them .
Message 8/29 19-Nov-02 @ 05:40 AM - RE: Wavetable Synths vs Modeling Synths
Hey, nice topic!
I've got a few hardware VA's, and am also thinking of getting into wavetable synthesis...
What are your favorite software programs and hardware boxes for wavetable synthesis?
Software: Someone mentioned the VST PPG software... The FM samples sounded far more harsh to my ear than the hardware PPGs and microwaves. I prefer smoother sounds... Was it just the programming of the those VST PPG patches, or was it the inherent sonic character of the software?
Also, Absynth, Metasynth, Reactor? Are these any better than the VST PPG?
Hardware: I've heard demos of the microwave and fizmo and really like them. But the fizmo, in particular, is so hard to find.
I wonder if one of those software programs could fill my desire for a few wavetable-type patches?
Or will the hardware always outshine the software? (hehehe)
FWIW, my audio interface is a Delta 1010, so it hopefully won't be a gating factor in sound quality and smoothness.
Any advise would be greatly appreciated.
Thanks,
Scott
I've got a few hardware VA's, and am also thinking of getting into wavetable synthesis...
What are your favorite software programs and hardware boxes for wavetable synthesis?
Software: Someone mentioned the VST PPG software... The FM samples sounded far more harsh to my ear than the hardware PPGs and microwaves. I prefer smoother sounds... Was it just the programming of the those VST PPG patches, or was it the inherent sonic character of the software?
Also, Absynth, Metasynth, Reactor? Are these any better than the VST PPG?
Hardware: I've heard demos of the microwave and fizmo and really like them. But the fizmo, in particular, is so hard to find.
I wonder if one of those software programs could fill my desire for a few wavetable-type patches?
Or will the hardware always outshine the software? (hehehe)
FWIW, my audio interface is a Delta 1010, so it hopefully won't be a gating factor in sound quality and smoothness.
Any advise would be greatly appreciated.
Thanks,
Scott
Message 9/29 19-Nov-02 @ 02:25 PM - RE: Wavetable Synths vs Modeling Synths
microwave? smooth? i didn't know that was what people expected from it. the XT can do smooth though (if nothing else, it's got the second 6db smoothing filter; remember to set higher quality/no aliasing/etc).
that's pretty cool brett... i didn't know emu samplers could do that. it's different from wavetable but similar in function (has pros and cons). how many voices does something like that use (w/128 layers)? would it only be a few, or 128?
that's pretty cool brett... i didn't know emu samplers could do that. it's different from wavetable but similar in function (has pros and cons). how many voices does something like that use (w/128 layers)? would it only be a few, or 128?
Message 10/29 19-Nov-02 @ 02:37 PM - RE: Wavetable Synths vs Modeling Synths
BJT
Posts: 90
Link?: No link
File?: No file
Yeah but the wavetable synth doesn't create the 64 samples, it just uses them. That's right aint it?
Interpolating the samples - that's sound morphing, can somebody PLEASE tell me if anyone has done work in this area 'cause I'm thinking of doing so.
Interpolating the samples - that's sound morphing, can somebody PLEASE tell me if anyone has done work in this area 'cause I'm thinking of doing so.
Pages: 1 2 3
There are 29 total messages for this topic
Reply to Thread
You need to register/login to use the forum.
Click here to Signup or Login !
[you'll be brought right back to this point after signing up]
Back to Forum
