PlanetZ forums - scopeusers.com

It's DSP's after all?

I mean, programming up from scratch. Possible using the Scope / Xite hardware?

What about a S/C and symbolic sound link up??? - S**T...it's unimaginable............but possible?

(Yeah, I know they have new stuff out)

Ableton and Cycling 74 have done it...I am really excited about Max integration

If someone made the right DSP code, something like that could be done. morphing resynthesis would be pretty cool.
quite a lot of new low level DSP code would be needed though. lets wait and see what happens in the new SDK!

If you are only looking to morph sounds, yes that could be done by Scope. In fact I would guess (but admittedly haven't tried it) you could get decent morphing-like effects with the Mod3 Vocoder Analysis & Synthesis modules, or the Orbitone Vorb.

But if you are looking to have a Kyma-in-a-Scope, then no, Kyma and Scope are very different animals. Essentially Kyma is a sound design laboratory, while Scope is a music creation studio. Different universes, and using one is nothing remotely like using the other.

well said, John

cheers, Tom

johndunn wrote:If you are only looking to morph sounds, yes that could be done by Scope. In fact I would guess (but admittedly haven't tried it) you could get decent morphing-like effects with the Mod3 Vocoder Analysis & Synthesis modules, or the Orbitone Vorb.

But if you are looking to have a Kyma-in-a-Scope, then no, Kyma and Scope are very different animals. Essentially Kyma is a sound design laboratory, while Scope is a music creation studio. Different universes, and using one is nothing remotely like using the other.

Yes, I know. It would need a completely different OS but just use the raw DSPs of Scope.
I wonder what the new Kyma hardware uses for DSPs? If its the Analog Devices chips it could be ported over?
(I know it will never happen though.... ...just thinking if it is possible......)

The lady genius would be happy to explain what she's created.
It's an ingenius platform that has 2 things in common with Scope, it's a real time application, and it uses powerful DSP chips and programming.
Carla Scaletti actually will answer the phone when you call, and it is a priveledge to have her explain things better. She grew up with ARP2600's instead of Barbie Dolls & Betty Crocker.
She actually talked me out of buying Kyma and buying Scope as I explained what I wanted to do live back in 97'. SInce we use to play all over Illinois and especially Champaign where she's from, we had a great discussion.
My good friend was playing keyboards with Jason Bonham ( Jon Bonhams son ) and John Paul Jones was helping out and offering ideas. That is when I first heard a CD he recorded using Kyma to treat audio tracks especially, and yes even DSP synths. They do have some and can do great renditions of subtractive analog, additive and other forms of synthesis.
The ultimate set up I would think is a great VST hosted audio/MIDI sequencer of your choice, a new Kyma and XITE-1. 10 years ago having both would have cost 20-25,000 USD fully loaded, now instead of 8U, you get 2U and can have both for around 8200 USD.
KymaX users are as finatical as Scope users, and harshly defend and worship Ms. Scaletti and others who continue to develope this platform.


http://www.wolfram.com/technology/guide ... ndSupport/
http://www.symbolicsound.com/cgi-bin/bi ... s/Pacarana

Just a couple factual updates. I have used Kyma for about 10 years, Scope (starting with a Pulsar card) for about 8 years, and I have written extensive MIDI control modules for both, to link to my software. So I'm pretty familiar with both the differences and similarities.

The Kyma team is a twofer: Carla Scaletti who created the original Kyma software, and Kurt Hebel who is more on the hardware end. Both are brilliant sound engineers, and Kyma is by any measure an amazing sound design tool.

But it is not (as Scope is) a music creation tool, a musical instrument. There is probably nothing better for creating complex new sounds or merging and modifying existing sounds in more ways than you can imagine. But, no matter how many DSPs you have it essentially can do only 1 task at a time. It can do a very complex task with lots of sub parts, but your Kyma patch is set up as sort of a single horizontal tree, with the branches on the left all flowing into the final output at the root on the right. One data flow, no matter how many DSPs.

Also, Kyma is not a real time system. Yes, it plays and can be controlled in real time. But you have to set up the data flow links, then compile - then it plays. You want to run a different synth, you have to stop and compile it anew. Basically Kyma is run in Smalltalk on the PC, then compiled to DSP code and sent to the hardware.

This works well for a single task, even a complex, multi part task. But ultimately only one task. So you can build a very complex, very intelligent synth, for example, but you really can't have a stack of different synths as you can with Scope.

Also, while Kyma is very good at sculpting and modifying sound, it makes a very unsatisfactory musical instrument. The reason for this - at least in my opinion - is that Kyma filters are more mathematical than musical.

To illustrate, think about the sound of a Moog VCF. Engineering-wise, it's downright funky, and the later synths had VCFs that were provably better - mathematically. But none were able to touch the Moog as a musical instrument. Because it wasn't designed as a math function, it was designed as a musical instrument.

Anyway, the Kyma filters (even the "retro" VCF they finally added) have a thin, distinctively "whiny" sound. It undoubtedly works fine for general sound design, but it just ain't musical. At least to my ears and given my skill, compared to the graceful and elegant Modular III, and other Scope based synths.

Scope, as you all know, was made for general studio work and for DSP based soft synths and effects. It allows you to set up multiple devices, not just one data flow. And, whether by design or accident (I'd say design, but probably accidentally blessed as well), it does work very well as a musical instrument. Sure, it's DSP based and therefore it's all based on digital math functions. But the VCFs and other components don't sound like math functions. They sound, well, musical.

So I'd say if you want to sculpt individual sounds, Kyma is the way to go. But for studio work and for the best DSP synths available, nothing touches Scope.

As for DSPs, Kyma uses Motorola 56000 series fixed point DSPs, Scope, as we all know, uses Analog Devices Sharc floating point DSPs, so even if someone wanted to port Kyma to the Scope platform, which is wildly improbable, it would require a fairly extensive rewrite of the compiler.

Thanks for the input, John! It's not usual to find someone who has hands-on experience with both Kyma and Scope... I have been looking into the new Pacarana, but listening to sounds and watching screenshots doesn't give you an idea of the workflow and a few things you said I hadn't realized before. Kind of thought one could have multiple parallel processes at the same time, like having a synth on one channel, and processing some tracks on another, etc.

Great info! Thanks again!


T

Great Info Indeed...
Basically the same thing explanied to me when I was interested.
Only 11 years later where I can understand John's explanation much better.
That's why you only hear certain synthesis rarely, and much more sould sculpting ala Autere, and many SFX for film, etc. DJ's who use it are the best " live " examples I ever saw.

BTW John Dunn ,
Are you considering development for Scope 5.0 or just an avid user?

Nice post John. I have a friend with a very basic Kyma system (Capybara with the just the stock DSP) and after 3 or 4 years he's still only scratching the surface. It does sound VERY VERY digital, but really good digital - in a totally different ballpark to other additive/resynthesis that I've heard. John is totally correct on its strengths/weaknesses - basically Kyma won't ever replace traditional instruments... it can only really excel at new ones.

Also regarding making something like Kyma for Scope - writing the code is the hard part, and is why it won't happen. And just because they both run on DSPs doesn't mean all that much - some DSPs are better suited to some things than others... AD make a huge variety of chips for example.

XITE-1/4LIVE wrote: BTW John Dunn ,
Are you considering development for Scope 5.0 or just an avid user?

Perhaps. Depends a lot on how the SDK5 falls out.

Kind of thought one could have multiple parallel processes at the same time, like having a synth on one channel, and processing some tracks on another, etc.

from the screen shots on the kyma site, it looks like their timeline lets you do multiple parallel processes...

not a Kyma expert at all, but the system is most likely to support paralell processing under 1 single 'control' instance.
But in Scope you have several independent items, like synths, mixers, fx-units under the (almost abstract) global SFP control.

cheers, Tom

does anyone know how many inputs and outputs kyma has? I would love it to have adat ports, unfortunately it hasnt. I would need realtime reshaping of cv signals, controlling modular or other analog gear...

if kyma could do that it would be awesome, but as it looks like it is really offline oriented, not really live...

is that true?

tomylee wrote:does anyone know how many inputs and outputs kyma has? I would love it to have adat ports, unfortunately it hasnt. I would need realtime reshaping of cv signals, controlling modular or other analog gear...

if kyma could do that it would be awesome, but as it looks like it is really offline oriented, not really live...

is that true?

Take a look at: http://www.symbolicsound.com/cgi-bin/bi ... ifications
the "Audio/MIDI Input & Output" section.

Kyma is a real-time platform just like Scope.

ah so you route it via the audio interfaces, but i guess you will get some latency through that

you'd get exactly the same latency if you'd perform those operations in Scope.
Kyma is well known to process entire 'time sections' of audio.
You cannot start to reverse playback something until the last tone is captured.

cheers, Tom

i dont think i fully understand what you meant, but I was wondering how much latency the mac interlal routing consumes, as far as I know mac osx can talk to different audio interfaces at the same time, but how much latency would that introduce? I am a mac noob but I am checking it out right now

tomylee wrote:ah so you route it via the audio interfaces, but i guess you will get some latency through that

The new paca(rana) talks directly to the audio interface, so latency is low and jitter is also low.

Macs have higher latency than scope or kyma systems, and believe it or not most equivalent pcs, too.

tomylee wrote:i dont think i fully understand what you meant, but I was wondering how much latency the mac interlal routing consumes...

very simple: the routing latency is neglectable, if you apply a (say) convolution process to a 16384 byte buffer.

cheers, Tom