Bert, you’re comparing the SSP with the Capybara('s Motorola processor). That’s not really fair, since the Capy is out of production for quite a while, and has been replaced by the Pacarana & Paca systems, since 2009 or thereabouts. Though nobody seems to know which DSP’s are inside the Paca[rana], I’m pretty sure it isn’t ye olde Moto’s anymore… Carla & Kurt have been doing this since 1989, so I reckon they know what they’re doing.
I didn’t say that they don’t know what they are doing. You asked for an objective factual comparison and I gave you one.
If you read the kyma website you will see they specify the processing power of the 3 products they sell (paca/pacarana/wormhole) in terms of capstones where 1 capstone is defined as one DSP from the Capybara-320. It’s public what DSP is used in that box so that is all that matters. The 56309 is also one of the latest DSPs from the now obsolete 56K motorola family. So one capstone = one 56309. And we know the MIPS of that processor, that is also public info.
I am quite sure they still use the same motorola DSPs because the DSP code was written in assembly specifically for that family of processors and you just don’t move over to a different architecture that easily.
Furthermore, the Wormhole isn’t a system an sich. It’s just several Paca[rana]'s hooked up to each other via ethernet. So, if you’re chaining 8 Pacarana’s – like they do in the Lucasfim studios – you’re looking at 16.000 MIPS (*), blowing the SSP straight out of the water. Admittedly, my wife wouldn’t be happy if I’d spend 32k$ on some small boxes. 
I don’t understand your logic regarding the wormhole. You are comparing an investment of $32,000 in an 8x pacarana system with one SSP of $2,000?
You can put 3 SSPs in your eurorack system, then you’ll have 21,600 MIPS which is more than the 16,000 MIPS you get with 8 pacarana’s, and you will have spent $6,000.
Your comments make me think you just came here trying to prove a point rather than asking for info to make a decision on what is best for your workflow and needs.
(*) This is assuming their processors do 100 MIPS @ 100MHz. I’m guessing they’re a lot faster now.
They are not. Read the datasheet of the 56309, I gave you the link.
Then there’s the philosophy behind the Kyma systems. There’s a reason why it’s “just a simple box”, and it’s a very important reason. All that box needs to do, is process audio. Nothing else. It doesn’t have to drive a fancy OLED display (and losing ticks in doing so), it doesn’t have to care about IO protocols nor does it
It’s best not to voice opinions about complex technical topics if it’s not what you deal with on a daily basis. I’m a computer scientist and I know what I’m talking about. This is my full time job.
There are no computing cycles “lost” by driving a big display because the CPU we uses has dedicated graphics hardware on board to do all that. The same goes for all the I/O.
have to convert A to D and back (since that’s what the external Audio Interface deals with), etc… All audio processing is always sample accurate. Everything it needs to do – and there’s really a lot of it – it does within that sample.
See above. Moving data from converters into and out of memory ready for processing is done by dedicated hardware in the CPU.
On top of that, literally, sits the Kyma software. It’s been around for nearly 30 years, which could be translated to “being an adult”. Compared to such vast history & experience, the SSP’s software is not even “a toddler”, I’m sorry to say. Having just some basic modules like an LFO, envelope generator, step sequencer, sampler, wavetable oscillator, granular synthesis, filters and a bit more… well, is not really a lot nor does it carry a lot of weight, compared to Kyma’s thousands and thousands of building blocks (“Prototypes”).
The development of the synthesis engine of the SSP started about 10 years ago.
Yes, we don’t have ALL the modules in the kyma system yet but we already have the most important building blocks necessary to build something useful and impressive, and we’ll be adding more modules as we go forward. I think I made this clear in my post as well.
And herein lies the biggest difference between the Kyma and the SSP. Kyma’s Prototypes can be regarded as quarks, whilst SSP modules are more something like a complete atom, or even a molecule. SSP modules are “set”, or fixed. “Here’s your VCO”… which is fine by itself, but what if I want to build my own VCO (or 16.000 of them running in parallel) ? Or what if I want to build my own formant filter, but I want it to process only the “a” vowels of an incoming live audio stream ?
If you want to build your own VCO or a VCO that has 16.000 oscillators in parallel you can use our SDK (which will be based on the VST standard, for which there is a massive amount of free source code and help available online, starting with the kvraudio community). You can also build whatever synth architecture you want in the patching grid on the SSP.
I’m guessing your answer will be "but sure you can, with our SDK."
Alas, I’m not a programmer. Nor do I feel the need to be one… So I’ll be stuck with whatever the SSP team or its user base-with-programming-skills has to offer me.
OK, so explain to me the process by which you build your own VCO in the kyma system without programming? Using smalltalk for writing DSP code is extremely inefficient. Are you sure you are not talking about controlling DSP modules? That would make more sense.
Let me add one more thing in favour of the Kyma system… indeed, you need a computer to “program” it {initially}, but once you’ve done that, you can disconnect the computer, and take just the Paca[rana] with you on stage. Apparently it’s also very reliable, and will never ever crash. There’s a famous example out there of a performance which had a laptop hooked up to a Pacarana, which (of course) crashed, but the Pacarana happily went on working, and the audience didn’t even notice anything.
You seem to imply the SSP is not reliable. Based on what info?
The SSP, in simple terms, is a digital modular synth/sampler. Granted, a very sophisticated & powerful one, but still kind of a closed system, like synths generally are. If I wanted to change the innards of one of its modules – akin to modifying the hardware of a VCO to adapt it to my needs (which I can actually do, btw) – I’d be stuck {because I ain’t no programmer}. And just like the SSP’s analog counterpart, a modular synth, I can’t really compose with it; I can only try to come up with a clever patch which might disguise itself as a whole composition, but really isn’t it.
The SSP is a modular synthesis system just like the Kyma system with the differences I outlined. There is nothing closed about it, you can write your own modules for it and you can even replace our software if you want and use the hardware for your own software.
You say you can build a VCO in software without programming, so I would like to hear in detail how you do that on the Kyma system and how you make it run efficiently.
Your statement that you cannot compose with a modular synth doesn’t make any sense to me. The kyma system is a modular synthesis system.
The Kyma system is like having a whole laboratory of resistors, caps, etc available, without limitations. I could use these to build my own modules, but that’s not even necessary; there’s plenty of modules (“sounds”, in Kyma speak) available. Then I’d build a “patch” with these. After that, I’d save that patch as a new sound (comprised of n modules, made with n resistors etc). Then I’d repeat that, and again and again, until I have something like 24 “full synths” running simultaneously. But I’m not done yet. I take all of that stuff, and put it in a linear timeline and/or non-linear multi-grid… which really is a simple way of saying that I’m actually creating a fully-fledged composition there.
In the kyma system you use software on the computer so obviously there you can work with compositions like you would do in any other DAW or sequencing software. The kyma hardware does not have a display so you cannot edit the sequence on the hardware itself.
The SSP does not have a fully fledged multi-track sequencer like is available in DAWs, but it could have one, which could allow you to edit directly on the SSP, because it HAS a big display. To play back sequences on the SSP all we would have to do is support loading MIDI Files in a MIDI playback module.
So again, I think you mostly wrote this reply because you wanted to prove a point, rather than asking for an objective opinion to make up your mind …
