2CAudio Kaleidoscope | It's A Trip | Latest Update 1.1

[Andrew Souter]

Andrew, this is very cool! Thank you for your informative posts and responses in this thread.

My pleasure.

I enjoyed reading the KS Manifesto last night in the wee hours (I went to make some tea after my first scan after I knew I wanted to give it a thorough read) which to some extent "resonates" with my interest in 3D fractal animation tours, space ambient soundpieces and a recent interest in Pythagorean mysticism.

You sound like an ideal candidate.

Simon (Sampleconstruct), I also appreciate your jam demos in this thread -- I think you are doing 2CA a service with those even or especially as they are jamming and ludic rather than tutorial and I'm sure this will also help the KS ecology grow in general.

Aboso F-ing Lutely... Simon rocks.

I have to consider my personal economy but this has definitely piqued my interest and is on my list.

Best wishes for a successful launch and development!

thanks.

[TwoToneshuzz]

I've downloaded and experimented with the Kaleidoscope and it works okay in Logic X Yosemite, I'm fascinated, and I'm very happy that the intro period is good and long as I think this might end up on my short list..

[Sampleconstruct]

Simon (Sampleconstruct), I also appreciate your jam demos in this thread -- I think you are doing 2CA a service with those even or especially as they are jamming and ludic rather than tutorial and I'm sure this will also help the KS ecology grow in general.

Thank's Havran, others can do the tutorials later, as I've spent a lot of time with this beast since last August, jamming and improvising with KS more and more feels like a musical journey, not at all scientific - once you know your way around, you instinctively know more or less what to do to achieve the sonic results you're aiming at.

[Hez]

It's not quite what you're after but I have had some joy so far "playing" Kaleidoscope by sticking it after Microtonic or Tremor playing very simple patterns on two or three lanes and then both tweaking the drum synths and the patterns on the fly. Mainly slow ambient sort of stuff though (20-60 bpm).

I too am interested in the best ways to get long evolving sequences. At the moment, as you note, it is not to be found, on my setup anyway, in tweaking the knobs in Kaleidoscope.

Good idea with Microtonic! I've just been dragging shitty Vengeance loops and other assorted vocal samples to trigger it so far but obviously pretty limited control there.

Happened when going either from 2 backwards or 64 forwards. CPU stayed high until switching the number of duplicates again. Haven't managed to reproduce it yet, unfortunately (fortunately?).

So I suspect what happened is you were using one of the String Resonator modes, and switching between 1 and 64 duplicates drastically changed the frequency distribution of the resonators. With 64 most of them were lower freq. With 1, many of them got into the more extreme high freqs? Extreme high freq settings for Strings cause more CPU usage than they should at the moment, as explained in the previous post.

You think this is what happened?

Yes, thinking back I'm pretty sure I was playing around with the string resonator modes at the time so that would make sense.

[BoerniH]

Many Thanks Andrew + Greg

that was the point !! )) ... now it works !!

and yes it's true, Kaleidoscope need a lot of power, i've a 6 Core i7-4930K ... the real-time peak goes to the max very fast ... and I'll eat, sleep and do some exercises

Cheers and all the best

[selvmarcus]

It's called Modal Synthesis. Reaktor and Max have special modules for this. See Reaktor Prism for application. Here combined with image scanning. Nice ICandy, but too much hype on the website and too expensive. Somebody could hack this in Max in a week, I guess.

[Sampleconstruct]

Somebody could hack this in Max in a week, I guess.

LOL - bring it on!

[selvmarcus]

'..time, frequency and space'. !!! Yeah right, heading directly towards the galactic center, correct?

[KBSoundSmith]

It's called Modal Synthesis. Reaktor and Max have special modules for this. See Reaktor Prism for application. Here combined with image scanning. Nice ICandy, but too much hype on the website and too expensive. Somebody could hack this in Max in a week, I guess.

Max is a workflow killer, and mainly a tool for academic masturbation. If someone wants to wank away making a Max patch, go ahead, but I won't be in line for their efforts...

[Gribs]

I hope I get my annual profit sharing check before the intro offer is over. I am tapped out until then.

[Andrew Souter]

It's called Modal Synthesis. Reaktor and Max have special modules for this. See Reaktor Prism for application. Here combined with image scanning. Nice ICandy, but too much hype on the website and too expensive. Somebody could hack this in Max in a week, I guess.

This is bit like saying, well algorithmic reverbs are made with comb filters and all-pass filters, so anyone who can type:

y[n] = x[n] + a * y[n - M]

and

y[n] = c * (x[n] + y[n - M]) + x[n - M]


can make Aether, B2, Breeze etc. Well, there you go, the above are our magic formulas. Go ahead.


Sure, Modal Synthesis which is a branch of Physical Modeling is part of the idea, and we did not invent it just as we did not invent additive synthesis, comb filters, all-pass filters, or any other fundamental build-ing block of audio DSP...

In fact I say on the web:

In the real physical world, all musical instruments use acoustic resonators that filter sound waves at specific frequencies. Resonators are a fundamental building block of the mathematics and physics of physical sound generation and filtering. Every musical instrument has resonators! Some generate the sound directly, such as the Strings in Stringed instruments, the head of a drum, the wooden bars in a xylophone, and the pipes in an organ. Some modify the sound by enhancing particular frequencies, such as a piano soundboard or the sound box of a guitar or violin. Organ pipes, the bodies of woodwinds, and the sound boxes of Stringed instruments are also examples of resonators. There is in fact an entire branch of musical instrument synthesis research and development called Physical Modeling that deals specifically with these topics and related ideas. Even real-world acoustic spaces such as anything from a small closet, telephone booth, or automobile, to large spaces such as concert halls and cathedrals can also be considered a type of resonator called an Acoustic Cavity Resonator. Kaleidoscope offers two primary types of resonators: Strings and Springs.

And I say in the manual:

In digital audio String models can be implemented as a simple Feedback Comb Filter. Frequency-dependent damping is achieved by putting a filter of some variety (generally a low pass filter) into the feedback path. With the proper parameter settings this simple model can create very realistic sounding String sounds. The first exploration of this general idea was the Karplus–Strong algorithm in 1983...

It's also a bit like saying anyone can code y = sin(2 * pi *x), so what is special about Galbanum Architecture Waveforms? Well... I have to assume NI is quite capable of coding such things themselves, yes? So why did they license waveforms from me for NI Massive? Why did Camel Audio do the same for Alchemy? Why did many others also do the same? These are VERY smart people/companies, and somehow they see the value in my research in these topics. And guess what? I did NOT invent wavetable synthesis or mathematical/procedural generation of waveforms either.

If you think that is all we have done with Kaleidoscope though, you misunderstand the idea greatly... I recommend to try the demo.

It is a very, very, very long journey from the white paper on one of many fundamental technologies involved in something to a fully developed retail product... it takes many years (371 builds and 3-4 years in our case to be exact...) If you know someone who can do it in days/weeks -- something novel that is, not just stealing an idea once it has been fully developed and displayed publicly: please send them our way: I will hire them tomorrow.

[synzh]

Any way you could point us to a usable graphics program that could aid in generating the types of images that might play nicely as images for KS? I know there are millions of control points there but the images usually need to have sweeping curves or repeated patterns in order to be useful (and brightness differences, the sweeps might be, say, darker on the edges like your preset images)?

[mandolarian]

This will be the start of a VERY long term platform for us and we are VERY committed to it. This idea will still be interesting in one or many decades and we hope to be there continuing to evolve it. 1.0.0 is just the start.

In the twitterverse that would be an eternity...

(shit, I think I just made myself obsolete... there goes my scoring and sound-design contract work... )

And then a short eternity later... in 2025 (OMG! That's so sci-fi) the few remaining KVRians, having migrated to the iBoard platform, might comment:

"Remember back when they actually hired humans for sound design?"

"Yeah, those were the days! Then KS Version 3 was released for iOS X and now there's an app for that."



I will download the demo soon, but having just listened to the demo tracks....80 hours might seem like a 10 minute trial period. So entrancing...

Congrats on the release Andrew! Even if it foreshadows the end of your contract work. And worse, may have the power to raise the clumsy mouse clicks of us aural butchers to the sonic genius level. Not sure I'm ready for automated artistry. But, willing to find out.

[rustman]

I can report that I'm up and running the demo in Cubase 8-64 in Win 7-64. All buffer sizes are at 2048. Task manager is showing CPU fluctuating between 18-25%, evenly distributed among 4 cores (8 threads). That's a good start. Now to tweak some knobs...

[Andrew Souter]

Any way you could point us to a usable graphics program that could aid in generating the types of images that might play nicely as images for KS? I know there are millions of control points there but the images usually need to have sweeping curves or repeated patterns in order to be useful (and brightness differences, the sweeps might be, say, darker on the edges like your preset images)?

For novelty/fun/experimental-use we intend to allow JPG format in an update soon, so that people can use their mobile phone etc to make images to use. But as explained earlier in the thread, using random pictures only gets you so far. It's fun, but it's not likely to win you a grammy.

Deeply understanding the meaningful organization in time and frequency (and space) for musical uses and how to create visual art that follows these rules is a skilled art form.

to answer your question succinctly with a web link:

U&I's Artmatic is a good tool:

http://www.uisoftware.com/artmatic/indexAMinfo.php

"easy to use" for such things is questionable. It gets HIGHLY technical... It is NOT what I used to make my images, but i do find it very fascinating, and would like to explore using it to make some images for it. If enough people are interested in it, I can see if we can get some deal or something...

Personally I have actually made my own tools to make such things that are targeted more specifically to make musically useful images. I have not decided yet whether or not to make this part of any product or not in the future. I want to make Architecture Volume 2 first.

I have found, like waveform generation, which I am rather well known for, that the waveform/image generation step is best done as a separate step from the higher level integration/design stage. When I am writing music, I don't think to myself, well what if I played this melody that uses y = sin(x + 17 * (exp(2 * cos(3x) -1)) as the fundamental oscillator shape/waveform. No. F no!! Instead I separate the processes and when I am creating waveforms I try to think scientifically to span the potential parameter space of various classes of waveforms in meaningful ways and generate a library of possibilities. Frankly I have no F-ing idea how cool they will be in application when I make them. So I make a lot more than I use and I try to cover the range of possibility as best as possible.

Then I move on to a higher level of development: preset development. Here I choose from these raw ingredients to create "virtual instruments" that have emotional content to some extent. This is a different mindset. Waveforms don't make me feel love/fear/awe whatever... When I use them in context to make a larger creation, an instrument per say in the case of a synth patch, they start to obtain human significance. It's a bit like "casting runes". We humans ascribe meaning and context to random events/organizations.... Waveforms themselves are meaningless data. Effectively random. Like runes. It is up to the preset designer to use them in purposeful ways and help ascribe meaning to them. The preset designer is the intermediate step of this process. The preset design ascribes meaning and human significance to build an inspirational sound with playable characteristics, and the artist manipulates these to tell a human story.

It's the same for images. You can think of images as 2D waveforms. By itself it is meaningless data (although it helps to be aware of various organizational principals). The pairing of an image with a tonality, and scanning time, and even a second image and a second scanning time, and its interaction with the input signal is what creates meaning in Kaleidoscope.

ONE given image can be used trillions of ways in KS via X-offset, Y-offset, 8-orientations, inversion... etc.... but resizing and brightness remapping... this can all make a single image sound quite different... even with the same tonality...

So, when it comes to complex images at least, the result is very hard to predict/control when put into context with all of the above. This is why I say it is bit like an imagination machine. You need to have the heart and soul of an explorer. Don't try to control it completely. Instead leave some elements effectively to chance. The statically possibilities are so astronomically huge that it can be considered random in some ways. Your job is to ascribe meaning to this randomness and find ways to help it tell your emotional story.

If you ask me what will happen if you combine XYZ tuning, with ABC Image and 123 timing, with alphabetagama input, my answer is probably "F if I know.". But that is what makes it fun.

This point is not to replace the 5 billion soft synths out there that allow you perform exactly what you want. The point is to make stuff that inspires you to exlore new forms and come up with things that your fingers don't have muscle memory for, or years of radio exposure has not trained your brain to expect...

the universe is a deep place. let's jump in a see what we find.