A while back I started a few threads asking about how Amon Tobin created the sounds for his album "ISAM". After loads and loads of research, the basic answer I got was mostly with Kyma and with bits an pieces from other synths and hand crafted modular synth patches.

At the core of what Amon Tobin did was something like spectral morphing. This basically means taking two samples and mangling them together. Kyma's TAU editor seems like the best tool for this job, but I've seen no demos of it and have no idea how it works.

There are other packages like Camel's Alchemy and NI's Kontakt which claim to do some type of spectral morphing. I don't want to disparage these products in any way, but I haven't been able to recreate the kinds of sounds Kyma can produce with these products.

To get an instant idea of what I am talking about, check out this video:
http://www.youtube.com/watch?v=eAVLrtOrcyc

The problem with Kyma is that it costs $3000-$4000 USD because you have to buy a hardware box to run it. I want to experiment with this sound so I'm tossing around the idea of saving up for one of these. But, before idea do, there's something else that it bothering me.

The thing that bothers me is that Kyma has been around since the 80s and seems to be a technology that has evolved rather than a technology that harnesses all the capabilities of modern computing. The physical hardware of a pacarana box (the hardware you have to buy) is quite old. And, I'm not convinced at all that the hardware would be necessary to run the kind of algorithms that Kyma is based on.

So, it seems to me that it would just take a good small team of programmers / sound designers to create something as good as Kyma but without the need for external hardware.

So, my real question is: is there an alternative to Kyma? After all my research, I haven't found anything which just grabs me like the Kyma sound. I could probably reproduce some of the sounds using Max or Reaktor or whatever, but I can't find any information on the way they create their sound. What are other people's opinions on this? Should I just bite the bullet and buy a pacarana?

csound or CDP.

unless you want it in realtime.

Harmor.

Yep. Looking in to CSound right now. Anyone used this? Any comparisons to Kyma?

Harmor looks great but it's just a synth - it's not a sound design platform. However looks/sounds great and I want to try it out but that means I have to reboot in Windows. Yuk.

Just bought Skanner XT from NI. It's great. Already getting some ISAM like sounds with the factory presets. Very impressed. Of course it's just a basic synth but at least I'm in the ballpark of what I wanted in the first place.

Awesome! Just finished this tune with Skanner XT!

http://soundcloud.com/syncretia/tempest

On the CSound front, I'm starting to think that this is going to be great.

I found some patches that allow you to load CSound patches in to a MaxForLive patch so I'm going to pursue that. I'm a programmer so I think that CSound will open up some avenues that are closed off in the MaxForLive and Reaktor world.

If you're on a Mac then Metasynth has MASSIVE sound mangling capabilities. If you're not on a Mac it's worth getting a MacMini for.

SuperCollider: real time synthesis and composition programming language.

http://supercollider.sourceforge.net/

Interesting audio examples.

CSound (CS), MetaSynth, and SuperCollider (SC) all look great!

I'm a little put off by MetaSynth costing $600 while the other two are free. But I'd be totally willing to pay that price if it offered functionality similar to Kyma.

Part of my problem at the moment is that it's hard t compare apples with apples. SC and CS are both programming platforms and when you download them, they don't really seem to come with anything out of the box. So, I've got no real idea of their capabilities or the kinds of instruments that other people have built. Has anyone got any resources on actual downloadable CS or SC instruments? Something with example code that is a really interesting instrument would really help to understand the capabilities.

As for Kyma, I've never seen this thing under the hood, so I can't possibly know why you would pay $4000 for the system when there are plenty of other systems out there that boast the same kind of functionality. All I know is that artists that use Kyma get great sounds out of it. That's the only fact I have to work with.

I'm very curious as to how the morph in Kyma works. You seem to get a kind of 'frequency bending' effect during the transition which helps to morph sounds of wildly differing pitches. Whether it does this out of the box or without spending some time tuning the algorithm's parameters to the different sounds you're using, I don't know.

This kind of thing has to be done manually via separate processes in CDP. E.g., if you have one sound with a fundamental frequency of 400Hz and another that's 300Hz, you have to apply a time varying transposition on the first file to bend it down to match the target's pitch before bringing the morph in to get an acceptable result. It doesn't really deal with unstable pitches terribly well either, but there are all kinds of preprocessing tricks you can use within CDP to smooth this over. Imposing amplitudes, sliding all partials to a fixed pitch and its harmonics then back again after the morph, making a 'bridge' from one sound to the other with your voice and doing a three-stage morph with the vocal bridge in the middle etc. The videos I've seen imply that Kyma does this kind of thing automatically, so you never get a result that just sounds like a grainy crossfade from one sound to the other.

I heard a few sound examples of the Loris opcode in Csound (which Whyterabbyt says has been removed in recent versions) that seemed to deal with different pitches in a similar way to Kyma. The morph was applied to very simple harmonic sounds, but I definitely heard a pitch slide from one pitch to the other in there.

The Pacarana hardware isn't that old. I think it came out about two years ago, and replaced the quite old Capybara hardware. As for whether the DSP could be replaced with native code, I couldn't say. My gut instinct is that most of the algorithms could run on a standard CPU, but your whole processor would end up being used for a single sound. Okay, perhaps it isn't that extreme, but external processing certainly has its place. I've debated about getting one of these for years, but concluded that I don't have enough free time to dedicate to it to justify the purchase.

Bottom line, if you want what Kyma does, get one. If you're content with tools such as CSound and Supercollider, no need to spend the money.

For Csound, there's the CSound Catalog, at the bottom of this page:

http://www.csoundforlive.com/purchase.html

Isn't most of the cool stuff in that video due to the Haken Continuum ? Only had a quick listen and it didn't seem impossible to recreate with something else than Kyma. The Continuum is a bit harder to replace it would seem to me.