Sound Design Stack Exchange is a question and answer site for sound engineers, producers, editors, and enthusiasts. It only takes a minute to sign up.
Sign up to join this community
I'm deeply interesting in purchasing a Pacarana. I use Max/MSP a lot, and while that's super powerful for tool creation, it's also super difficult to get something to sound awesome.
Oh, yeah, and super time consuming.
What can you say about your experience with Kyma? How does it fit into your workflow? Your answers will help me decide where to put this $3k I was otherwise going to spend on things like eating and rent.
(source: ssc-media.com)
I wrote this as an email response to someone asking for thoughts on KYMA, so thought I'd repost here for others benefits.
Ok, first off, here's a really good post on Kyma vs MAX/MSP vs Nord Modular by Anthony Bisset.
I too use MAX/MSP. I also use it for tool and workflow creation. Aka SevenUpLive 2.0 for the monome.
MAX is really good at connecting stuff and controlling stuff. I don't find it particularly well suited to designing sounds.
KYMA on the other hand has a primary purpose of being almost a definitive platform for sound design. Supporting everything from nice sounding patches and straight subtractive stuff to very esoteric and bizarre forms of synthesis. You pretty much can combine anything with anything so a lot of times you create mashups of a bunch of different higher level patches. This is far more productive at actually creating interesting sound design than MAX/MSP.
The sound quality is untouchable. It's immediately apparent when you start working with it. This all stems from the fact that all sound computations happen in realtime within one clock cycle or sample. So KYMA is sample accurate. You are not doing sound computations on vectors of samples, unless you want to.
Another aspect of KYMA is the ability to have hot values for outside control of the patch. This is very easy to do. These hot values can connect to controls or can be controlled via midi or now OSC which gives you a nice connection to the outside world. the new OSC support is great, especially to connect MAX to KYMA.
The flow of a KYMA patch is a directed graph of signal flow. This is of course very different than MAX but very familiar in synthesis since signals flow from left to right. The components in the flow can be very low level or very high level prototypes for synthesis.
It's ridiculously easy to put a filterbank of 32 bands for example in a flow controlled by other aspects of a patch. It's also easy to resynthesize a sound by using an oscillator bank of 1000 sine waves controlled by the spectra of another sound.
But then you might be using something very simple like a VCA or an ADSR envelope.
But to answer your specific questions
Everything I hear about kyma is really interesting to me. I love the semi-programming interface. What I wonder though is how "useful" it is. I mean, it's one thing for it to be an amazing tool for very infrequent occasions. It's another thing entirely to be a system that can contribute to every project I work on, and allow me to be more creative and think outside of the box in more than just the exceptional situations. That's what I'm looking for! Not that I don't love the lucid nightmares about buffer size and envelope shape etc. that Max/MSP gives me. So my question to you: How do you use Kyma? How valuable is it to you, creatively? How do you fit it into your workflow?
First of all, I'm not a sound designer by profession, I just have an interest in music and synthesis. If I was though I think I'd be all over KYMA for a few reasons.
http://createdigitalmusic.com/2010/03/15/osc-files-play-that-funky-music-hexagons/
Modular synth controlling KYMA http://www.vimeo.com/10943340
You can build your own library of building blocks and keep combining them into new things. For example, let's say you want to create crazy engine, mechanical drones. You could build a patch that uses various methods of using detuned oscillator banks to create you own drone machine.
Kind of a secret weapon for sound design. If that was my profession. I'd be all over it.
So my workflow for KYMA is like so. Firstly, I use it for effects. Granular reverbs. Pitch/Time shifting, Filterbanks etc...
Sometimes I use KYMA to recreate an idea I found on the modular synth. Maybe I was playing with some interesting FM concept and I want to build a patch that recreates an aspect of that for reuse later.
It's a learning tool for discovering new concepts. Someone mentioned Karplus strong algorithmns. I looked it up on Wikipedia, found a prototype in KYMA. Built a patch to recreate hitting metallic tubes. You can experiment for the full range of pretty much any synthesis technique ever developed, or maybe invent your own. That is just cool.
I can do things in KYMA I just cannot do in other systems. So it complements other techniques. More fun to patch subtractive stuff on modular, but then I can do stuff like put the sound through a 32 band fixed filterbank whose bands are keying of notes in another sequence or something.
Cons: While you can use preexisting effects and sound out of the box. That would not be enough to justify KYMA. There is a learning curve for sure. If you have already learned MAX then I'm not worried but plan to invest 100 hours in this before you get over the hump. If you just want say an environment to do subtractive synth stuff. It definitely does this but it would be faster in an environment that was optimized to be a modular synth. The open ended nature means sometimes you have to discover which tool to use to get there. There are over 200 prototype components, so takes a bit to find the best path. There are usually a lot of solutions to the same problem.
For example, simple square wave, use an Oscillator that uses a square wave table, OR use two saws offset by a delay line and combined together. Now I can adjust the delay to morph smoothly from saw to square. I can put in a hot variable for Pulse Width, or I just grab a prototype that implements PWM. You can get to the metal or work at a much higher level than you can in MAX.
Use Tao editor to create smooth spectral morphs between two completely different sounds, human to cat to dog etc... Resynthesis and morphing? The are a half dozen ways to do this in KYMA that are different.
Last advise. If you are serious, buy the book!! KYMA revealed $35. It's a great read even without KYMA. If you feel drawn to the concepts there, you will likely enjoy the system.
Hi there
Got my Paca a few months ago too, and have been digging my way through the manual, and just tried out experimenting on my own.
As said earlier, this unit, and program can do things no other program can, but one has to beware of the fact that it takes alot of time to get into (At least it does for me. Im on to page 100 in the manual, and have not touched it in a few weeks because of a big project tht needed to be done, but Im eager to dig in again.
Its exiting to use the different sounds with the Wacom tablet, and it makes the sound tweaking so much more natural, then mousing does.
Up until now, I think I get most satifying results on treating voices. Here you can easily get weird starwars like accents. When I try to use the patches on winds, machine sounds synths, I have a tendency to go to far, and suddenly it all sounds bad, distorted, or non-character like (is that a word???)
On thing I really like, and find very well thought out, is that you can take a prebuild sound and look at each of the building blocks it consits of. This gives me some insight in how to get more different and exiting results.
Oh, and the TAU editor is nuts!!! When analysing a sample,( again a voice) you can get results you never imagined was possible!
Customer support is as mentioned earlier. Super. Carla and Kurt are great.
Im using the PAca on a macbook pro 2,26, and the Konnekt8 soundcard. I wish I had the money to get a bigger and faster mac, since the opening of timelines etc. can take a bit of time sometimes. The Konnekt8 soundcard is ok. Nothíng special. Using the Spdifs as inserts into my Protools setup.
I say go for it if you got time, but be prepared to spend months and months and then some, if you are using it a few hours a day. Its deep, really deep.
Best wishes,
Mikkel
Hi i'm bought my paca just few months ago... and had used only a little in two projects. but it's a terrific machine. You need to invest a lot of time (other than money) but it's a real pleasure: always find something inspiring. I use max as well, but I found that in kyma is much easier to come up with something exciting; and it sounds gorgeous. I used it to create some fill for production tracks (a trick that you can find in the kyma guide) and to process voices (making fire and waves "speaks"). but the incredibile thing is how fast you can create your library of processes that you can use and use again. I'm creating some patches to process sounds with doppler fx/ panning/realtime stretching and pitch, all controlled with the wacom tablet. Carla and Kurt, the owner and maker of kyma, are really helpfull and great at answer questions and sort out problem. it isn't the most user fiendly tool, but it's amazing... and i have just scratched the surface :)
Davide F
Preface: I currently use Audiomulch, Plogue Bidule, Reaktor and MaxMSP - pretty much in that order according to frequency. Not coincidentally, that's also the order I'd put them in in terms of how quickly and easily I can sit down and get some great results out of them when building from scratch. In reality, I do very little from scratch in either Reaktor or Max.
About once a year for the past decade I take another wanton look at Kyma. Only very recently did I finally meet someone who owns one and get a chance to spend some hands-on time with it. Bear in mind, I only had a weekend with it and it was immediately clear that I wouldn't even be able to begin to put a hairline scratch on the surface in 3 days time. So instead of making a feeble attempt at learning to program it from scratch, I instead just spent a few days digging through the library of patches the owner has built up over the years, as well as a bunch factory example patches. Here's my first impressions in no particular order:
Yup, I no longer want one. I simply don't have the time or left brain energy to sink into it anymore. Pretty much the same reason I've slowly regressed to using Max and Reaktor as "you program it, I'll play with it" environments for tweaking existing creations from others. I've kind of fallen madly in love with Audiomulch and I can spend all day in there without even noticing. It hasn't got the depth or complexity of Kyma, but combined with my other tools I can get results I am more than pleased with very quickly.
Kyma has probably been one of the simultaneously frustrating and rewarding devices I've ever used. At times it feels like it was purposely designed to make learning it far more difficult than it should be. But at the same time, it will make the most insane sounds ever with barely any effort (except for the 100-200 hours of figuring things out). I look at it this way. I remember reading somewhere that modern digital synths compared to analog, are more difficult to get to make a usable sound because there are so many options. But with this comes a more extended range. Analog on the otherhand is limited, but almost every sound that comes out is nice. Kyma is the extreme far end of the digital synth paredigm. The user has pretty much control over every single aspect of every parameter as well as access to all synthesis models that exist. The user needs to have under control every single last thing, much of which has been taken care of automatically with most commercial synthesizers. This includes gain staging at pretty much every transformation point in the signal path. Trivial once learned, but an absolute nightmare in the beginning because along with getting into this new mentality, you also need to figure out where everything is on the interface. And the interface is not a good one. The interface probably makes perfect sense if you have never used any piece of software ever.
Overall, this requires a lot of time to learn and I think knowing some software programming skills could be necessary as well to get it to do really cool things. I knew max/msp for several years before this and I don't think that has helped much other than the conceptualizing part. Technically, it has been annoying knowing how to easily do something in Max which then takes 3 days to figure out in Kyma, but ends up being just as easy, just conmpletely uninituitive. But maybe Max was the same way while learning too, it was just a while ago and now is second nature. Can't wait for kyma to be the same way.
A few things that I don't think are great are the overall process of learning this, the quality control, and the hardware. For 3K the hardware seems very cheap. The metal is thin, the fans are many and loud, and overall, it just seems made as cheap as possible. It gets the job done though so the processors are good, just the overall feel is cheap and flimsy. It gets super hot too which I don't think is good either. The manual is deep, but repeats often and isn't too organized nearly as good. But most importantly, the examples are way too complex for someone new. I went through something that looked like a tutorial and it took a few days to get through. And what I learned was so specific that I don't think I would directly use much of what I learned other than in a real general way. On the other hand though, there are sections and paragraphs here and there that will teach a whole world of synthesis in an hour. Amazing... but really only after spending so long to be able to understand simple things. I believe that if the manual was re-organized and some simple tutorials were setup for very basic familiar concepts, the learning curve could be reduced by 90%. Seriously, the chaos is that bad. Most material is written by someone who knows a ton about the machine and is really far off from someone brand new. It gets really musical, but it is the most engineering/computer science paradigm synthesizer out there. You will feel this while using the machine, but that's why it's so powerful.
I forgot what the original question was, but I would both recommend and not recommend Kyma. Even if you think you know your stuff, you probably don't and this will make you realize that you don't. And if you don't think you know your stuff, this could be so far over your head that it might be difficult to get it to make any sound. But then again, maybe not. At the least, I would say to be comfortable with the basics of subtractive, fm, sampling and have some clue about additive, granular and analysis. Again, knowing the basics of programming will also help more in deiphering what the manual is trying to say and the mindset of the UI, more so than actually programming the parameters of the sounds which is actually not too complex for some nice results.
I've had Kyma for a few months now. The sound quality is awesome. I have been using it for things like
Just the simple use of Kyma has upped my production in my mixes. Yes, I've made really bizarre sounds with Kyma. It kicks ass with basic effects to. It also sounds better than VST plugins. I still use VST stuff here and there but, Kyma has added that little bit of bright polish that I was missing.
I just thought I would mention an aspect of Kyma that you don't really read about. It's not as easy as using VST pluggins as far as the interface, but once you spend some time with it, it's not bad. It's Just different. I'm a new user and still have a lot of learning ahead of me.
I hope this is useful insight with a little different angle on Kyma.
I'm Kyma user too. There's always something to investigate and you always make new discoveries every time you dig it. It's a money investment for sure but the biggest investment will probably be the time you'll spend learning how to get good sounds out of it. I think it takes about 6 months to get a palette of really interesting sounds and to justify the money you spent, but after those 6 months you'll never regret it!
Realtime processing setup: - Pacarana with an Apogee Duet on a MacBook Pro with Wacom Intuos4 and soon an iPad with TouchOSC. (Analog In and Out From and To an iMac).
I'm always up to exchange sounds on Tweaky... You'll find me as JeanEdouardMiclot.
I've made a blog with already two articles including patches pics and soundcloud samples about Kyma at this page: http://jedsound.com/blog/
Have fun.
Kyma has probably been one of the simultaneously frustrating and rewarding devices I've ever used. At times it feels like it was purposely designed to make learning it far more difficult than it should be. But at the same time, it will make the most insane sounds ever with barely any effort (except for the 100-200 hours of figuring things out). I look at it this way. I remember reading somewhere that modern digital synths compared to analog, are more difficult to get to make a usable sound because there are so many options. But with this comes a more extended range. Analog on the otherhand is limited, but almost every sound that comes out is nice. Kyma is the extreme far end of the digital synth paredigm. The user has pretty much control over every single aspect of every parameter as well as access to all synthesis models that exist. The user needs to have under control every single last thing, much of which has been taken care of automatically with most commercial synthesizers. This includes gain staging at pretty much every transformation point in the signal path. Trivial once learned, but an absolute nightmare in the beginning because along with getting into this new mentality, you also need to figure out where everything is on the interface. And the interface is not a good one. The interface probably makes perfect sense if you have never used any piece of software ever.
Overall, this requires a lot of time to learn and I think knowing some software programming skills could be necessary as well to get it to do really cool things. I knew max/msp for several years before this and I don't think that has helped much other than the conceptualizing part. Technically, it has been annoying knowing how to easily do something in Max which then takes 3 days to figure out in Kyma, but ends up being just as easy, just conmpletely uninituitive. But maybe Max was the same way while learning too, it was just a while ago and now is second nature. Can't wait for kyma to be the same way.
A few things that I don't think are great are the overall process of learning this, the quality control, and the hardware. For 3K the hardware seems very cheap. The metal is thin, the fans are many and loud, and overall, it just seems made as cheap as possible. It gets the job done though so the processors are good, just the overall feel is cheap and flimsy. It gets super hot too which I don't think is good either. The manual is deep, but repeats often and isn't too organized nearly as good. But most importantly, the examples are way too complex for someone new. I went through something that looked like a tutorial and it took a few days to get through. And what I learned was so specific that I don't think I would directly use much of what I learned other than in a real general way. On the other hand though, there are sections and paragraphs here and there that will teach a whole world of synthesis in an hour. Amazing... but really only after spending so long to be able to understand simple things. I believe that if the manual was re-organized and some simple tutorials were setup for very basic familiar concepts, the learning curve could be reduced by 90%. Seriously, the chaos is that bad. Most material is written by someone who knows a ton about the machine and is really far off from someone brand new. It gets really musical, but it is the most engineering/computer science paradigm synthesizer out there. You will feel this while using the machine, but that's why it's so powerful.
I forgot what the original question was, but I would both recommend and not recommend Kyma. Even if you think you know your stuff, you probably don't and this will make you realize that you don't. And if you don't think you know your stuff, this could be so far over your head that it might be difficult to get it to make any sound. But then again, maybe not. At the least, I would say to be comfortable with the basics of subtractive, fm, sampling and have some clue about additive, granular and analysis. Again, knowing the basics of programming will also help more in deiphering what the manual is trying to say and the mindset of the UI, more so than actually programming the parameters of the sounds which is actually not too complex for some nice results.
.JPG){kind=link}