Kyma APU (Ristretto)

[thetechnobear]

not really expressive controller related, though I know there is an overlap between the kyma and expressive controller community…

so this is the new generation Symbolic Sound hardware…
(its been in the works for some time, but I think announced over last day or so?)

https://kyma.symbolicsound.com/pacamara-apu/

I like that they are marketing this as an APU (so audio version of GPU)…
of course, nothing new here, and you could argue UAD are much closer to this analogy…
(as kyma you patch directly for… so more like OpenGL + GPU?)

anyway, Kyma is well respected, so good to see it being updated…
with USB3, wifi , faster / smaller - all good

non-pro - $3300 for 2 gb/2 cpu
pro - $3800 4gb/4 cpus

pricing feels a little odd… Im not sure many will bother with the non pro (2 cpu) version, half power, and save ‘only’ $500.
don’t get me wrong, from symbolic sound side… the units are going to be very similar in cost… just not sure how many will go for it.

theres also no mention (I could find) of the possibility of upgrading non-pro to pro.
e.g. with plugin board.
as, if that was possible (at say $500) then I could see it as a way to ‘deep your toes in’… though again, for $500… is it worth the hassle? you are already pretty committed with $3k.

bit too pricey for me… (as a hobbyist)
but perhaps this’ll means a few of the older Pacaranas will turn up, as the pros upgrade

does anyone have any more details? in particular… have they stayed with the same DSP chips?

I wondered if for cost/ease of updating, if they might move to more ‘generic’ ARM SoCs… no musician would care, as long as performance is there… and might be easier for SS in the long run…and also might give SS the option of doing cheaper versions… to get a wider user base.
(that said, the real investment is in the software (kyma)… so I can see why ‘cheapening’ it, might be unattractive to SS)

[Photon]

Thanks for sharing this. While I can offer no insight, I’ll comment that I was excited at first at the prospect of an “AuPU” that can offload arbitrary audio tasks without latency issues (which is the bane of working on a random laptop). I guess this is once more an addition to DSP units which handle proprietary plug-ins and sound generators rather than a general solution. Too bad.

[Kai]

If Kyma is most naturally thought of as a sound source, rather than an audio processor, I suspect I might like it. But I want everything related to the mixing/mastering process to be VSTs.

I used the Sonic Core (initially Creamware) Scope platform for a very long time. It is flexible and great sounding. I still miss it, but… as DAWs became more powerful, the hassle of having to deal with two “environments” on the same computer became more and more of an issue. Initially, I had all audio in Scope, and just used Cubase for midi sequencing. But as technological progress made it natural to do more and more audio work in Cubase, routing audio/midi back and forth, the extra latency introduced every time I did so, etc became a significant workflow nuisance.

My hardware sound sources gets recorded into Cubase at some point. And from there on, everything is contained within that single environment. Introducing DSP based processors at that stage means only real-time mixdowns, which then means a risk of getting clicks in the audio, etc. For me, it made the process of iterating on a nearly finished mix cumbersome enough that I decided it wasn’t worth it.

[thetechnobear]

yeah, these ‘audio processors’ are really their own environments… rather than a ‘generic resource’.
of course, there are some chips with integrated DSP chips, but they are pretty niche.
(I think for most audio developers the integrated fpu is enough)

I don’t see this likely to change any time soon…
if we look at GPUs , their success was because there was widespread requirements for lots of users.

so, whilst the gpu’s all have their own ‘api’ , operating systems added supported to abstract them…
in the gaming world we got DirectX from Microsoft , and more serious applications had OpenGL
(also Microsoft/Apple wanted more visual in their OSs so that became important)

this means ‘application’ developers, dont need to code to specific graphics cards… this is an essential step…
thats pretty much missing in the audio world… and unlikely to happen since the music making market is tiny compared to gaming and the likes.

also there has been no consensus as far as I can see about what would be needed… music tech companies have focused on proprietary environments to build audio apps…
(VST vs AU, Intel vs Arm, Windows vs Apple)

the only effort we have see to even try to address this was SOUL by Roli.
But thats just a kind of proof of concept, which appears to have had limited take up… and the repo, has not been updated for about 18 months… so looks to have stalled.
(looks like the split of Juce/Roli may have also had an impact)

It’s a shame, SOUL did kind of tackle this ‘head on’, recognising the requirement to have a way to split the DSP code from the target… so that specialised hardware could be supported.
(no comment on its approach, as thats a different topic)

---

anyway, for now… I see it staying the same…

I do think Kyma is appealing, but like all of these ‘environments’ (Supercollider/Max/PD) it does require commitment (and investment) - thats probably off putting to general musicians/hobbyists.

however, if you’re a specialist… e.g. doing somethings like movie fx, learning a specific tool is not unusual (not different from a learning a daw, or unity for 3d graphics, etc etc) … which I think is Symbolic Sounds primary market.

of course, as a developer, it does also interest me a lot…
whilst I agree with @Kai its a hassle to have extra hardware, and more tech,
I also like dedicated hardware for tasks, the idea that you can just plug it in without a computer (for use/performance) is appealing, also off-loading the ‘load’ to is useful (*)
But I can see that , this trade-off is pretty personal… and depends on if you want to do the time investment.
(I’ll say the long running nature of Kyma, makes this a bit more appealing… which is why Im pleased to see it being updated, and kept relevant… as it could easily fall in to obscurity by tech advances)

---

(*) that said, my M1 Mac is powerful enough, that I never have issues

[keymanpal]

@thetechnobear

that said, my M1 Mac is powerful enough

eheh, yeah already a pretty high investment
The old “laptop-culture” is winning me back big time… sure, starting with the speed/load, the whole thing seems more refined/solid.

[MWAU]

Nice to see this get posted here. I’ve been a long time lurker (as a Continuum owner) but never actually posted…hi!

I’m a hobbyist who dipped my toe in Kyma five years ago. It does take an investment of time but of all the different systems I’ve used (inc Max, Reaktor etc) it’s by far the most creative, inspiring and rewarding I’ve found. That it plays so well with the Continuum is an added bonus.

The Pacamara is a lovely little box. So glad they’ve future proofed the system!