On and on feasibility of implementing Kyma natively?

[soundmodel]

So I have a Capybara-320 and the Kyma library is pretty extensive.

However, what I dislike is having to have external boxes, firewire etc. and having to use separate software to do processing.

Thus I always think, why not have something Kyma-like as a VST?

It should be doable, right? But I don't know who would do it.

Perhaps something like the CSound library could be extended to cover the extra base in Kyma?

[Xenakios]

But I don't know who would do it.

Nobody is going to start wasting years of their time to reverse engineer the DSP algorithms and file formats, only in the end risking a lawsuit with Symbolic Sound.

[Michael L]

As a Kyma user, can you tell us:
What group of existing VSTs we should buy to get similar Kyma functionality, if we connect them in a modular DAW like MuLab or Bidule?
What Kyma functions are missing in the VST market?

[VariKusBrainZ]

Hehe

[Tj Shredder]

I worked for a short time on a kyma system. As with all tools, the specific workflow will influence what you are doing. I was deep into Max/MSP back then already, thus Kyma was a bit close to the model and mindset I was used to anyway. But there was a big difference with the sequencer. A one of its kind... Some aspects you could do in modular vsts, or max4live, but its very different and a very open environment...
Imagine Reaktors core would be simple and has a low learning curve...
To receate something like the Kyma language, you need to get into smalltalk, its a smalltalk derivate...
Symbolic sound might do it eventually themself when they are about to retire, with todays multicore CPUs there is less need for slower than that DSP chips...
But they are settled very well in the film industry - there money is no issue...

[soundmodel]

I'm still fond of this idea. Kyma is great, but it's closed so as to not being able to draw as large userbase as it perhaps could, if it was more accessible.

I've been thinking of starting to write plug-ins for Faust, inspired by Kyma.

[soundmodel]

As a Kyma user, can you tell us:
What group of existing VSTs we should buy to get similar Kyma functionality, if we connect them in a modular DAW like MuLab or Bidule?
What Kyma functions are missing in the VST market?

I'm not sure of this yet, because of limited Kyma experience. But one would be latency-free additive/spectral synthesis with "big number" of partials. Or also anything involving using FFT here and there. While modular environments may support, they don't support it to as high partial count as Kyma, which limits their fidelity.

The other thing I've noticed is that Kyma comes with several "production-ready" patches. Designing those requires some ingenuity I believe, so for "production-ready" it's not sufficient to have an environment that has the "blocks for it", but also have ready patches that offer interesting musical results.

There's sometimes an indifference between gear designers and musicians. Musicians tend to want "something ready-built to play with" and are not always "interested in building their own instruments".

[knobcore]

Why not just use cabbage?

https://cabbageaudio.com/

[whyterabbyt]

Why not just use cabbage?

https://cabbageaudio.com/

Probably because Csound.

(Dont get me wrong; Rory is a genius, and Csound is an incredibly powerful audio language. But its also a very old, and in many ways, primitive one, and the fact that its text-based is an immediate stumbling block for many people who'd be fine with a Kyma, MAX, Puredata etc.

IMO, Csound would really needs a visual instrument builder for that audience, which, unfortunately, Cabbage doesnt provide. Always thought it was a sort of pity Paul Barton wrote Ardour instead of pursuing his visual front-end for CSound.)

[knobcore]

Why not just use cabbage?

https://cabbageaudio.com/

Probably because Csound.

(Dont get me wrong; Rory is a genius, and Csound is an incredibly powerful audio language. But its also a very old, and in many ways, primitive one, and the fact that its text-based is an immediate stumbling block for many people who'd be fine with a Kyma, MAX, Puredata etc.

IMO, Csound would really needs a visual instrument builder for that audience, which, unfortunately, Cabbage doesnt provide. Always thought it was a sort of pity Paul Barton wrote Ardour instead of pursuing his visual front-end for CSound.)

I honestly didn't find it THAT hard to deal with CSound. It took about as long as it took for me to learn how to get around in PD. More of a matter of RTFM. And it's really not the orchestra part of CSound that is hard more than the score. The fact you don't gotta write scores anymore and can just use MIDI/OSC/float automation in DAWs with cabbage/csound~ really cuts down the headache time.

Like, you got i-rate (initialization signals processed at message time), k-rate (control signals processed at audio buffer time), and a-rate (audio signals). Then you got all of the gen routines for table generation/manipulation. Then you got instrument blocks.

Then once you have that understood, you can create oscillators and filters and things or know enough to rip the vast library of qutecsound that already has all that stuff available and save as VST. Then once you have your "library" of VSTs with your processing algorithms that can be handled with standard DAW automation, you don't even need to communicate with CSound anymore.

[whyterabbyt]

I honestly didn't find it THAT hard to deal with CSound. It took about as long as it took for me to learn how to get around in PD. More of a matter of RTFM.

Its never posed me any problems either; Ive poked around with things like CSound and its precursors since I bought Hal Chamberlin's book "Musical Applications of Microprocessors" in '86. However, I think you're overestimating the level at which most people want to work.

[knobcore]

I honestly didn't find it THAT hard to deal with CSound. It took about as long as it took for me to learn how to get around in PD. More of a matter of RTFM.

Its never posed me any problems either; Ive poked around with things like CSound and its precursors since I bought Hal Chamberlin's book "Musical Applications of Microprocessors" in '86. However, I think you're overestimating the level at which most people want to work.

I don't think I'm overestimating at all. Might've been true for the CSound of 1986, even the CSound of 2006, but not the CSound of 2016 and beyond. I skipped CSound on a lot of years but watched as they added real-time support, then CSound~, and now with Cabbage and all the Qute CSound stuff. And now in 2019, I have a 60$ Raspberry Pi 4 that runs CSound within PD environment on low latency kernel Linux and it performs damn great for the cost. Type in a couple oscillators or a phase vocoder, pipe MIDI to it, and you're done in five minutes flat.

And if you don't want to program it, just get all the Max4Live instruments and effects I guess from http://csoundforlive.com/ and then you can drag and drop! How hard.

Besides, the OP specifically mentioned CSound, and nobody said that this stuff was even available and I'm not even sure you are underestimating how far it's come even in the past few years.

[whyterabbyt]

I don't think I'm overestimating at all. Might've been true for the CSound of 1986, even the CSound of 2006, but not the CSound of 2016 and beyond. I skipped CSound on a lot of years but watched as they added real-time support, then CSound~, and now with Cabbage and all the Qute CSound stuff. And now in 2019, I have a 60$ Raspberry Pi 4 that runs CSound within PD environment on low latency kernel Linux and it performs damn great for the cost. Type in a couple oscillators or a phase vocoder, pipe MIDI to it, and you're done in five minutes flat.

And if you don't want to program it, just get all the Max4Live instruments and effects I guess from http://csoundforlive.com/ and then you can drag and drop! How hard.

Besides, the OP specifically mentioned CSound, and nobody said that this stuff was even available and I'm not even sure you are underestimating how far it's come even in the past few years.

Well I know Im not underestimating it, since Ive been there at all the stages you skipped. Im well aware of what the state of play is, I have both Pi and Bela based modules for running it in Eurorack, and I have been saying that CSound has the potential to be an open 'software Kyma' for years, but I still think that's primarily going to be if visual instrument building became possible.

I was saying as much earlier this year:

viewtopic.php?p=7342637#p7342637

And here's when I first made the suggestion, in 2005.

viewtopic.php?p=1005605#p1005605

However, having actually tracked it all the way through, you're going to have to do more than say 'there's this list of things people could do' to pursuade me that more than a very small minority will do it. CsoundVST predates Cabbage by at least half a decade (and I was posting here about Cabbage in 2013), but stayed firmly off the mainstream radar, as did csound~/Csound4Live.

I'm not persuaded that that situation is going to significantly shift until the ease of use does. And from what Ive observed from people who've actually tried, that specifically means not having to build new instruments in an unfortunately dated programming language. I'd prefer it if that were different, but having been thinking about the possibility for at least 15 years, Im a realist about that.

[Michael L]

Have you tried the Csound environment, blue?
https://blue.kunstmusik.com
I have not yet but have been reading the (2019!) manual.
It runs on Java 8. From the website:

Blue is a music composition environment for Csound, written in Java, and available for use on Windows, Mac OSX, and Linux. It allows doing everything one can do in Csound as well as builds on top of it to offer the following features:
-- SoundObjects are the building blocks within Blue's score timeline. SoundObjects can be lists of notes, algorithmic generators, python script code, Csound instrument definitions, PianoRolls, Pattern Editors, Tracker interfaces, and more.
-- These soundObjects may be text based or GUI-based as well.
-- Score timeline allows for visual organization of SoundObjects
-- Polyobject are objects which hold other SoundObjects, and have timelines in themselves. Working within them on their timelines and outside of them on the parent timeline helps organize and understand the concepts of objective time and relative time between different objects.
-- NoteProcessors allow for modifying the SoundObjects score results, i.e. adding 2 to all p4 values, multiplying all p5 values by 6, etc. These NoteProcessors can be chained together to manipulate and modify objects to achieve things like transposition, serial processing of scores, and more.
-- The Orchestra manager organizes instruments and functions as an instrument librarian.
-- Graphical instruments and effects that help to increase musical workflow.
-- A mixer system and user-interface to easily connect instruments and effects together that follows a conventional mixer paradigm.
-- BlueLive - work with soundObjects in realtime to experiment with musical ideas or performance.

[whyterabbyt]

Have you tried the Csound environment, blue?

Its probably the best front-end for (non-plugin) CSound, but its all about the Score, not the Instrument.