Sound designers - new programs please, NOT combi's

[Daz]

Despite the perceived difficulty of the SY range at the time, only 5-6 parameters were critical to EFFECTIVE sound design, where as with MOD-7 I still cannot even figure out which parameters, among hundreds are most valuable.

I can't imagine a system where only 5 or 6 controls can make effective changes to an FM program unless those 5 or 6 controls are able to control multiple parameters, in a similar way to the macro parameters in MOD-7.

Funnily enough, one of the great MOD-7 programs is a single FM op running into a Pickup waveshaper whose depth is modulated by an envelope to create a very cool EP sound base. Effective sound design with the Oasys doesn't always need to be complex. Similarly some of the more complex looking Programs are just layering the same carrier/mod setups, with some tweaks. Again a lot of great programs can be made that way, and those programs are easily understood (not by Tone Adjust though).

Daz.

[Daz]

Forgive me ... another post ... it's just so exciting to be talking here about something which is actually on topic and about synthesis

Two things :

1) What I like about FM is that it tickles that same spot that things like PWM or Wavesequencing do. It's so great to be able to get the oscillator itself to change timbre in far more interesting ways than you can by just filtering a static waveform or sample. Approaching an operator pair as a "Time Variant Oscillator" (as Roland might have called it), is a great way to think about this model and get somewhere with it.

2) Clavia seem to be going bonkers with their PM'ing of samples, and that kind of stuff can be done with MOD-7 too. As has been mentioned using PCM as modulator often leads to just crunchy results, but using some of the simpler waves (loads of which are available in Wavestation category) or prefiltering PCM before using it a modulator (how cool that you can do that!!) leads to less bizarro results. Of course I am focusing here more on the creation of abstract sounds rather than things acoustical.

Daz.

[ski]

Funnily enough, one of the great MOD-7 programs is a single FM op running into a Pickup waveshaper whose depth is modulated by an envelope to create a very cool EP sound base. Effective sound design with the Oasys doesn't always need to be complex. Similarly some of the more complex looking Programs are just layering the same carrier/mod setups, with some tweaks. Again a lot of great programs can be made that way, and those programs are easily understood (not by Tone Adjust though).

Daz,

That's a great example. Good sound design does not necessarly have to = complexity. I think you've also brought up some other really salient points related to FM programming OASYS, including the idea that you don't have to be tied-in to a preset algorithm a la DX7 etc. You can use any OSC as a modulator or carrier. You can patch them up as needed. This is precisely the approach I took when programming MOD-7 --- I never used the preset "algorithm"s because I wanted control over each step of the process of building up a sound.

Look, at the end of the day, MOD-7 blows away the programming capabilities of a DX7, yet the fundamental principles of FM synthesis haven't changed or been compromised. In fact, they've been enhanced.

Let's leave PCM out of the picture entirely for the moment... All sounds must start with an OSC, yes? Assuming you're starting with the configuration that lets you patch your own "algorithm", patch an OSC to the mixer and you have sound. Actually, that's already preset for you, isn't it? It may not be very interesting to listen to but you have sound. Ta da! If you set that OSC to sine, you'll hear a sine wave. If you set it to square you'll hear square. Very straightforward.

Next, sticking with the simplest of FM models, you need a second OSC to modulate the first one to create some kind of a timbral change in the first OSC. At this point, the OSC you're hearing is considered the Carrier and the one doing the modulation is the Modulator.

MODULATOR---->CARRIER
(sine wave)       (sine wave)

So patch that up and turn up the output level of the modulator (if it's not already up). Bingo, you have an immediate change in timbre. Couldn't be easier.

Now, do you want that change in timbre to be dynamic (i.e., change over time)? Simple --- set the modulator's EG shape and depth to control the amount of modulation:

MODULATOR--->EG--->CARRIER
(sine wave)                (sine wave)

This is absolutely no different from the DX7, except that you have the freedom to choose which OSC you use to be a carrier or modulator. And of course you're not limited to using sine waves. And heck, the OSC's don't even have to be designated "carrier" or "modulator" by way of the Macro parameters. The OSC's just do what they do per your configuration. Couldn't be simpler.

Of course, if you want to designate each OSC as carrier or modulator in the Macro parameters, you then provide yourself with appropriate control of each "class" of OSC by means of the Macro controls. "Brightness" will only affect those OSC designated as modulators.

As mentioned, unlike the DX7, you're not limited to using only sinewaves for the OSC's. You have choices of waveshaping, triangle and square waves, and PCM modulation. It's up to you, the programmer, to experiment with these options and see if they appeal to you sonically. Some will, some won't.

Don't think for a second that us guys who programmed the stock sounds for MOD-7 were able to entirely predict in advance what the effect of, say, Wavetable #23 on the sine wave used as a carrier being modulated by two other OSC's. We didn't. Despite our collective FM-programming experience we too had to experiment to find what worked and what didn't (per our individual stylistic bents).

Ultimately, there is no prerequisite for a synth owner to need to know how to program a synth. None whatsoever. But if you want sounds that aren't already programmed in the factory patches, then your choices are:

a) hope that someone else produces a library of patches that includes that sound (no programming knowledge required, but perhaps an infinite wait)

b) learn something about programming so that you can tweak preset sounds (Tone Adjust is a good way to mangle sounds even if you don't know what you're doing LOL!)

c) dedicate time to learning various synthesis methods so that you can create your own sounds from scratch. It's addictive!

d) Finally, you can hire a programmer! You know, those guys on the street holding up signs saying, "Will program for food".

[Kevin Nolan]

There's a lot to respond to-

To ski-
Without trying to seem smart, consider the piano. That's an instrument where, with only one sound available, its full potential can be explored by simply performing with it. You don't need a computer science degree or 200 hours to program it to get a meaningful sound. Its already there. Your only option is to explore the potential of that sound through playing.

But the same is true of various other electric and electronic instuments - four examples worth mentioning are the Hammond B3, the Fender Rhodes and the Yamaha CS80 and the Roland Juno 106. I'm not saying these are special, they are just founr examples that come to mind of instruments whose full sonic potential is explorable. Take the CS80 - with no preset capability (apart from 4 mechanical memories) and with all parameters immediately available on an exquisite control surface, and with unique performance parameters such as an astounding ring modulator, a ribbon contoller and polyphonic aftertouch through a weighted keyboard, this instrument's formidable synthesis capabilities, involvong a lot of paramters, are presented to you in such a sophisticated way that you cannot but explore their full potential. Increadible thought was put in by those who designed it to allow it to be exploited incredibly well.

There's nothing theoretical about it - its totally possible. While in London I also tried out the Roland V-Synth GT and was blown away - despite its sophostication, its programs were very easily and effectively modified, and performable in hugely sophisticated ways, even with no knowledge. Now that's a performance synthesizer.

I believe the STR-1 synth engine is also along the lines I'm talking about. It was as stright forward to learn as the SY range, if not more so AND its provided presets are very strong, with many effective parameters in preset programs beautifully assigned to the available real time controls. Hence its programs are modifiable in radical ways easily, whether to create new programs or to perform in innovative ways. It's also quite stright forward to create a new sound from scratch.

So to me the possibility of providing a connection to the synthesis engine in musical, spontaneous and instant yet still hugely sophisticated ways and that are a pleasure, has been demonstrated over and over again in the past. I believe you are thinking in too much a 'parametarised' way, such as with the DX7 and the MOD-7 for sure, where it is clear that their potential is far harder to exploit.

The reason why I am forgiving to the DX and SY synths is not that they are good to program and create sounds - rather its because thousands of users did all the hard work which we can all benefit froml, and with such a vast array of created programs, the best that's possible from these instruments was actually realised. Hence DX and SY synthesizers are instantly playable through a very useful palette of sounds already available without each of us having to do but a small fraction of the necessary effort - power in numbers. But this does not forgive the absolutely dreadful progamming nature of the DX7 for example, and the SY77 and SY99 were not a lot better. Hence the MOD-7, while massively sophisticated, is very difficult to program. I could live with that if thousands of programs were available for it, but they are not. I have attempted over many long hours to realise some desired sounds though MOD-7 in recent months, but got absolutely nowhere and had to give up due to the long hours being wasted.

Why I don't simply move on, and why I want to harness the MOD-7 in similar fashion to the SY range is because of: MOD-7's potential to deliver hugely if actually programmed well, its greater polyphony and fedility, its interactivity within the OASYS environment; and also, because I find the SY sounds still to be so important that the hope would be that MOD-7 could potentially be even more significant for me in the long run.

Re. the SY range being a matter of taste - that's true, but their acoustic properties are not a matter of taste - that's what they are actually about, and likewise MOD-7's potential is in this department, but few of the available programs are harnessing that quality. MOD-7's power will be in acoustic natured programs that change harmonically though performance - because that's what acoustic really means to us. MOD-7 can be astounding in this manner - but you'd be hard pushed to find existing programs that respond to performance.


To Dan -
Despite my rediculouslty long posts, I am actually short on time at the moment (!) but will try in the coming few weeks to make a number of SY sounds to demonstrate what I mean. Regarding the vibrator and body feedback - what I mean is (and I'll be delighted to stand correccted or have my explanation imporved by any synthesis guru's including your good self) - in a violin, its sound is not just the sound of the string and the sound of the body - from what I understand its also a feedback from the body into the strings, where the body of the violin frequency modulates the string vibrations, constituting a vital part of the sound of the violin - and what FM synthesis is supposed to be about. Its not a direct algorithmic mapping in the Yamaha/CCRMA FM synthesis world, its more a general correspondance. But between velocity affecting the operator envelopes, and RCM synthesis, very useful 'acoustic like' results are obtained - particularly with acoustic guitars and bass's.

To Daz -

The 5 - 6 parameters I dound critical in the SY range are (from memory - it's been a while!)

Algorithm
Operator level
Operator Feedback
PCM Level
Velocity Scaling to Operator Level
Envelope settings

and that's it.

I know Envelope settings involves several parameters, but the envelopes in the SY range are an absolute breeze to manage - unlike the OASYS envelopes which I find bloody complicated. And you do not have to assign envelopes as you do in MOD-7 (and other OASYS synth engines) - you know immediately which envelope applies to which operator and its all very graphical. Looping envelopes on the SY range is also a breeze, and setting the all important (on some voice types) release level is also straight forward.

Again I do not want to be too critical - I am personally delighted that MOD-7 is so sophisticated, but it's on the far end of the scale to the piano example I gave above - it's got hundreds of parameters, offering more potential than any of can imagine - but that's the problem! Perhaps I'm remedial on these things and some of you are excellent at programming MOD-7, but I am not, though I have tried hard. i do accept the general arguement that time must be invested to reap rewards, and I will return to MOD-7, but I find it most frustrating of all because personally I'd love to exploit it, but currently cant.

To get back to my original point - the OASYS synth engines, in my view, coudl do at this stagewith some time invenstment in programming; and if the best programmers in Korg and on this forum spent the time on programs that they spend on combi's, the sky would be the limit for OASYS.


Cheers,
Kevin.

[ski]

I'm short on time too, so I'm going to be brief.

To ski-
Without trying to seem smart, consider the piano. That's an instrument where, with only one sound available, its full potential can be explored by simply performing with it. You don't need a computer science degree or 200 hours to program it to get a meaningful sound.

Anyone can bang on a piano. Not everyone can "play" it. That's the difference. So no, you don't need 200 hours to learn how to program it. But you need a lifetime to learn how to play it. But we're talking about sound design here, not performance ability, so I'm not sure I understand the analogy.

But the same is true of various other electric and electronic instuments - four examples worth mentioning are the Hammond B3, the Fender Rhodes and the Yamaha CS80 and the Roland Juno 106. I'm not saying these are special, they are just founr examples that come to mind of instruments whose full sonic potential is explorable.

Any instrument's full sonic potential can be realized provided that a person dedicates the time to it. The DX7 had one slider and 32 pushbuttons. That in itself was a gross simplification of the traditional multiple pot/switch approach. One parameter at a time... it didn't get much more difficult than that. It's probably one of the reasons there are so many tens of thousands of truly crappy DX7 sounds out there --- because not everyone had the aptitude to deal with such a severely limited "user interface" (as compared to, say, a minimoog).

I believe the STR-1 synth engine is also along the lines I'm talking about. It was as stright forward to learn as the SY range, if not more so AND its provided presets are very strong, with many effective parameters in preset programs beautifully assigned to the available real time controls. Hence its programs are modifiable in radical ways easily, whether to create new programs or to perform in innovative ways. It's also quite stright forward to create a new sound from scratch.

I'm surprised that you'd have trouble with MOD-7 if you find STR-1 easy to work with. I think that physical modeling is a much more arcane approach to synthesis compared to FM.

I believe you are thinking in too much a 'parametarised' way, such as with the DX7 and the MOD-7 for sure, where it is clear that their potential is far harder to exploit.

Well, that's not exactly a truism. MOD-7's GUI exposes FM synthesis to daylight quite unlike the DX-7's one-parameter-at-a-time, 2-line LCD approach to programming.

And regarding "thinking in parameters"? Every synthesist thinks in terms of parameters! Every synth is designed in terms of parameters! So no, I don't think "too much" in terms of parameters. There's just no other way to approach sound creation. In other words, when making a sound from scratch (which is what we do), we don't have the convenience of using any kind of macro to help shape the sound until that sound exists!

And then we DO provide "performance controls" in terms of the controllers themselves (knobs, ribbon, VJS, etc.). Do they not provide enough variation? And if not, what kinds of things are you looking for? Of course, we'd have to discuss this in terms of parameters, i.e., "a macro that raises the pitch of all modulator frequencies simultaneously" (which actually exists in MOD-7 in both macro and do-it-yourself form). It's entirely possible to scale the EG times simultaneously in MOD-7 (as with AL-1, HD-1, etc.). In fact, those controls exist too (see Tone Adjust).

Algorithm -- exists -- just change it!
Operator level -- there are macros for these, or, you can roll your own
Operator Feedback -- feedback for all OSC's can be tied to a controller or AMS source and controlled (and if I'm not mistaken, there's a macro for this too)
PCM Level -- piece of cake (modulate PCM level with a controller or AMS source)

Having read your post I understand why you want to try to incorporate MOD-7 as a source of your SY-type sounds. But like I said in my previous posts, there are four ways to get sounds, and there's a point at which you have to decide for youself if it's worth spending the time to pursue the goal. You can benefit from the work of others, or you can explore the synths yourself and discover your own creativity. But ultimately, no one's fault if the sounds you are after haven't been created.

[Daz]

And you do not have to assign envelopes as you do in MOD-7 (and other OASYS synth engines) - you know immediately which envelope applies to which operator

If you're talking about creating a Program from an initialized patch, then the envelopes are already assigned (EG1 thru EG6 to OSC1 thru OSC6, and EG7 for the PCM OSC). Funny that this feature of assigning envelopes isn't one you're fond of, but is one of the great features of MOD-7 for me. It's great to be able to control two modulators with the same EG for example, rather than setting individual params, or even tying your carriers to main AMP EG. It simplifies things and as a bonus frees up EG's to be used elsewhere. In the AL-1 and HD-1 the envelopes are also pre-assigned more or less (Pitch, Filter and AMP EG's). Admittedly it's not quite like the V-Synth where you have an envelope on pretty much every page, that is the most convenient envelope arrangement I've come across (accompanied by probably the least convenient LFO arrangements LOL).

Interesting that the list of parameters you gave doesn't include operator frequency ratios or velocity sensitivity. Anyways I am downloading the SY manual, yourself and Mike have piqued my interest

Daz.

[Mike Conway]

I am downloading the SY manual, yourself and Mike have piqued my interest

You go, booooy!


One trick that the SYs had was turning a 6 osc FM program into a "waveform." That is an AWM (sample) program could be assigned the FM wave, so that the ROM engine is playing the FM sound, plus you had 2 new filters to apply. In essence - a 4 element sound, 2 AFM + 2 AWM programs could combine to become a 24 oscillator program, with 8 note poly. Considering the AFM poly for 1 element was 16 voices (notes), it was an interesting option.

The segment looping envelopes were very cool, but I'm equally glad that MOD-7 has so many re-triggerable options, as well as step sequencing and MIDI sync LFOs. Different paths to animation, but both powerful.

As I said, ROM was such a premium, at the time, that the emphasis was on making samples expressive. Still a good goal, but less crucial, now.


One thing that the SY99 had that I would like to see on the OASYS is some of the tuning table ability. Each note could be any pitch from the whole keyboard spectrum, whereas the OASYS is limited to +/- 99 cents, per note. It's like having a "fine", but no "coarse" value.

[sirCombatWombat]

One thing that the SY99 had that I would like to see on the OASYS is some of the tuning table ability. Each note could be any pitch from the whole keyboard spectrum, whereas the OASYS is limited to +/- 99 cents, per note. It's like having a "fine", but no "coarse" value.

You have my vote on this!

For an interesting side note the MIDI Tuning Standard calls for 0.0061 cents adjustability.
The SY99 has micro-tuning steps of 1.171875 cents, and human ear detects a pitch change only after about 5 cents.
The resonance of two simultaneous notes is more easily heard.
For example if I play minor third with a saw wave and try to tune it to be harmonically perfect I find my self in need of something between +15 and +16 cents.

More on my adventures in microtuning here:
http://www.korgforums.com/forum/phpBB2/ ... hp?t=32817

[thekeymaster]

I know this swayed off topic a little with the FM thing but I'm gonna say,yes please Korg a few more programs would be welcome.You can never have enough in my eyes.My only concern now,with all the EXi etc etc and my own programming is space.

Are we gonna see more bank access anytime soon.I would welcome that so much.

But yes more top quality programs would be ace,a bit like the batch that was just released for the M3....heard some recently and they sounded very impressive.....not that i'm dissatisfied with the OASYS ones I have by any stretch.