Which parameters on DX7?

[frompuggle]

I'm reading Howard Massey's great book "The Complete DX7" to better learn FM. It's really great! But I'm confused on what some of the parameters are called on a DX7 versus a PreenFM2.

On the DX, he keeps talking about coarse freq (op freq on Preen, but it goes to 31 rather than 16?), but then he talks about operator amplitude which goes to 100 on the DX. There's IM, which seems to be modulator amp, which goes to 16 on the Preen. But then there's the envelope, which I think modulates this same parameter between 0 and 1, and I'm not really sure how these interact.

Massey says when you have a 1:1 freq ratio between carrier and modulator and 100 modulator amp (full IM?), you should get a saw wave, and 2:1 and 70 amp (about 10 IM?), you get a square wave. I've been able to get the square wave and something saw-like, and the pulse mod exercises he explains, but more when IM is closer to 6 and envelopes are static. He then describes detuning particular operators in particular algorithm configs as creating beating, phasing, wah-ing, and (if you use audio level fixed freq carrier) as phasing. I'm not sure I hear all the differences between these. I'm curious how much these differences are due to those between FM on the Preen and PM on the DX, or that I'm doing something wrong. 

Any help on translating between these two interfaces?

Btw, the Preen sounds AMAZING, such a cool synth, which is why I want to understand all this better. Thanks so much for a great synth!

[Xavier]

Hi,

You cannot map the DX7 parameters to preenfm2 with precision.
Freq on the DX7 from 0 to 31 are internal integer values of the DX7.

With the preenfm2 i tried to make the values more relevant.
The freq is the ratio of the op freq / the note you play. Ex : freq is 2.0 : you play a A 440hz, the freq of the op will be 880 Hz.

The IM is only relevent to FM, the DX7 is phase modulation not real FM.
And the IM is what you'll read about if you read generic FM manual.

The DX7 preset import was hard to do and is limited because of that.

Reading DX7 manual is very interesting but it's not easy to map that to the preenfM2.

Xavier

[frompuggle]

So the IM on the Preen roughly corresponds to something like modulator amp on the DX?

More importantly, how do IM and the operator envelopes for modulators interact on the Preen?

(and I must admit, I don't really know what he's talking about with the differences between phasing, chorusing/flanging, wahing, beating techniques).

[Xavier]

So the IM on the Preen roughly corresponds to something like modulator amp on the DX?

Yes.
Except for the lower level of operator (the ones that produces the sound / carrier operators). Then you have volume and Pan on the preenFM.

More importantly, how do IM and the operator envelopes for modulators interact on the Preen?

Enveloppes are inside an operator.
IM is a multiplier between operators depending on the frequency of the operators. The spectral impact is so the same for all note you play.

Xavier

[frompuggle]

I am still confused. On the DX, frequency ratio is timbre, modulator op amp level is brightness, and envelope simply modulates modulator op amp (brightness).   

The Preen seems to work really differently. If you set IM to 0 and mod envelope sustain to 1, you hear nothing from the modulator, so the envelope is not merely modulator amplitude with 0=nothing and 1=full modulator amplitude. But when you raise the IM it seems to change not only brightness (like on the DX) but also timbre, yet differently from that done by the freq ratio. 

So when you say that "the IM is a multiplier between operators depending on the frequency of the operators," what does it exactly multiply, and how does it do this in relation to the envelope?

Also it takes MANY turns of the encoder to go between 0-16 on the Preen, which makes me think it unusual that you'd use the higher values, and looking through the presets seems to confirm this. How to best think of what these values are doing?

Thanks!

[frompuggle]

Looking at the Wikipedia entry on FM and Chowning's original paper, it seems that the basic pattern of the sideband distribution is determined by the freq ratio, but the number of sidebands increases with increase of IM. Since this is done by increasing the amp of the modulator, it controls both brightness and the number of sidebands, hence the depth but not shape of the new timbre, since the shape is done more by freq ratio. And the envelope would then control the slope between minimum of 0 and maximum of 1 of the IM. Yes?

If this is right, it would seem a division into timbre and brightness on Yamaha PM in the DX line would be a simplification for programming purposes. I read something about pitch stability and self-oscillation signal dropout as well.

I also noticed that when I modulate the IM by larger numbers by an LFO or envelope, it sounds phasey and strange, but less so at small modulations, is this because of phase cancellations?

I am still confused as to why it takes so many encoder turns to cover the full domain of IM from 0-16 starting from 0, is IM not meant to be in high ranges?

[Xavier]

The easiest way to understand how it works is to start with a very simple exemple :
Use sin waveforms only.
Set a simple algo : ALGO1 and start with all IM* to 0.
Set a fast attack and high sustain for OP1. OP1 enveloppe in ALGO1 control the volume of the sound.
You should get a simple sin sound.

Now turn up IM1 to something like 4.0. OP2 now modulates OP1. The sounds gets richer.
And play with OP2 enveloppe. In ALGO1, OP2 enveloppe controls the modulation of OP1 by OP2 over time.
Try to plug an oscilloscope VST and keep an eye on it (http://www.voxengo.com/product/span/) while doing all that.

I am still confused as to why it takes so many encoder turns to cover the full domain of IM from 0-16 starting from 0, is IM not meant to be in high ranges?

No.
IM gives very intersting results with a very small value. Depends on the type of sound you want.

But if you want to accelerate the steps hold x10 button (INST) while turning.
And if you turn fast, there's also an acceleration in the Encoder behaviour.

Xavier

[frompuggle]

holding INST while turning the encoder helps a lot, thank you! I think the other things are making more sense now too. thanks for the help! and it REALLY sounds amazing, such an incredible synth. thanks so much!

[frompuggle]

Another question. I really, really love how the Preen sounds, it has totally gotten me into FM synthesis. I am using it with a Korg 707, an old 4op with the same chip as the Yamaha Tx81z, just better interface. So I started to look for some other FM or digital synths to complement this, but I noticed some sound very flat next to these. The Yamaha Dx200 was very flat, even though it is 6op FM, also I played a Korg Kronos FM engine, also really flat, no depth to the sound. Both sound like soft synths (and since the Korg Kronos is a DSP running software rather than something using chips, I guess it is almost the same as a softsynth).

My question is what makes the 707 and the Preen sound so thick and 3D, and when the oscillators go in and out of phase, so rich, that the others do not do? I know the 707 is 16 bit and the Preen is 18 bit, so it does not seem like it is a DAC issue, so could it be clock rate? I hear the Yamaha SY series, which has some nice features, is 16 bit but a much slower clock rate for the lookup tables, and someone said this makes the sound have less 'punch,' maybe that is what I heard as flatness on the Dx200.

[SirPrimalform]

The Preen is also a soft synth from that perspective, just a program running on a processor. Am I right in thinking that with the old FM chips, they performed the FM as a direct function of their architecture rather than being a processor running a program?

[frompuggle]

If so, I wonder why the Preen sounds so damn good. It sounds anything but flat. I wonder, could it be in the amp stage that some wonderful distortion gets introduced?

[frompuggle]

So Xavier. I am SO very, very happy with your incredible sounding synth. It is truly amazing. It sounds huge. But I am unsure why. The other day I played the Korg Mod-7 engine in an over $2000 Kronos synthesizer, and I was sure to find each and every effect and turn it off. And to my ears, the Preen wins on all fronts. It sounds much more immediate, direct, and real, 3D, while the hyper expensive FM sounded 2D, software, fake, like a softsynth. But the Preen sounds like hardware. Why? They are BOTH just software on a single DSP, right? So why does the Preen sound good? Comparing it to classic Yamaha FM, the Preen sounds CLASSIC. But from a technical perspective, why? 

[Xavier]

So Xavier. I am SO very, very happy with your incredible sounding synth. It is truly amazing. It sounds huge. But I am unsure why. The other day I played the Korg Mod-7 engine in an over $2000 Kronos synthesizer, and I was sure to find each and every effect and turn it off. And to my ears, the Preen wins on all fronts. It sounds much more immediate, direct, and real, 3D, while the hyper expensive FM sounded 2D, software, fake, like a softsynth. But the Preen sounds like hardware. Why? They are BOTH just software on a single DSP, right? So why does the Preen sound good? Comparing it to classic Yamaha FM, the Preen sounds CLASSIC. But from a technical perspective, why? 

Thanks  I'm really glad you like your preen.
Your message gave me a big smile.
Everybody have different ears and are more sensitive to some aspect of the sound.

I don't know what you ear that you like so much.
The thing that make it sound different from a softsynth is to my opinion the limit of its hardware.
The distortion of the audio path, the not perfect waveforms, as well as the limit of the DAC can be the "Hardware" sound you hear.

Of course its internal flexibility and the panning per operator are also responsible to make it sound "huge"

[tsutek]

Hi,

first post here, just wanted to comment on the DX/fm conventions:

frequency ratio means the same thing on both systems.
the DX envelopes are logarithmic, so their values are a bit tricky to translate
modulation index vs DX operator level -> this is weird to translate as well!
IM of around 13 = op level 99 in DX, IM of 1 = op level 65 in DX
(DX uses op level so that the algo does not matter, usability decision?)

EDIT: Added image



Other tidbits IIRC

Yamaha DX uses lookup tables, the puny cpus didn't have enough grunt to calculate stuff in real time

Chowing claims yamaha DXs operate @60kHz ie. up to 30kHz signals without aliasing

PS looking forward to grabbing a preenFM2 in the very near future!


Cheers