Just finished a demotrack for the synth. The idea was to present as many sounds in as brief track as possible, yet in a form with enough resemblance to a dance track production.
All instruments and sounds are made with Kairatune, except for the drums and percussion. The Kairatune tracks were mixed very conservatively – no additional processing per track, although some reverb sends were used. The final mastering was performed without restrictions to bring the end result closer to what one might hear on a real production.
So far the feedback has been mainly positive and encouraging. Most of the negative comments have been about the user interface, nonstandard or unintuitive parameter naming or layout. Also the documentation for some of the more unusual parameters has been insufficient. These are the kind of shortcomings that are very hard to spot from the designer/developer point of view, which makes the feedback especially valuable. I’ll do my best to fix these issues before the release. Huge thanks to the testers!
I’ve also received a number of ideas and suggestions for improvement. I take that as a good sign, though i don’t have the time to implement much of that stuff on the first release.
A couple weeks ago i received a dozen very creative high quality presets designed by users. They’re quite unlike my own designs and very interesting. Things are moving on just great!
Kairatune’s sound engine consists of
- generator section,
- pitch section,
- spread section,
- amp section,
- filter section and
- effects section.
The generator section has just one multi-oscillator. The basic waveform generator is very simple adjustable mix of sawtooth and square wave. The multi-oscillator generates five waveforms whose relative pitch, phase and amplitude may be configured and modulated enabling a rich set of tones. The multi-oscillator uses quite unusual technique that generates a single fundamental signal independently of the (multiple) overtones, preventing the beating effect that arises where several oscillators are output in slightly detuned unison, yet preserving the full and fat sound on the overtone spectrum. The method enables building very rich and full sounds on a perfectly solid bottom and is especially useful in creating tight bass and low register lead sounds.
The pitch can be modulated by a traditional vibrato, by a trill modulator and by a shared (sinewave) LFO source. The trill modulator is basically a square wave LFO. The modulation depth of all three modulators is independently controllable by an envelope. This enables the momentary use of quite extreme modulation. With the modulation envelopes the exact time and duration of the modulation can be set to produce some quite interesting effects and timbres.
The spread section makes the sound appear to oscillate in the stereo field. It uses a variable amount of high frequency damping and delay to create the illusion of movement.
The amp section modulates the amplitude by an AHDSR envelope and the shared LFO source and velocity. The asymmetric waveshaper can handle quite extreme driving without audible aliasing.
The filters section has a resonant lowpass filter and a resonant highpass filter. The cutoff frequency can be modulated independently by an AHDSR envelope and the shared LFO source. The lowpass filter’s cutoff frequency can be modulated by a spread modulator, which creates interesting effect in the stereo image, especially on high resonance configuration.
The effects section runs two delay units in parallel. One has a pre highpass filter and the other has a bandpass filter in the feedback chain. The cutoff frequency of the BPF is modulated by an dedicated LFO. The BPF delay is capable of some really interesting sounds on high feedback configuration. The effects section contains also a six stage phaser and two EQ units with selectable mode.
At the beginning i didn’t think too much about how and when i was going to release something. At first it seemed like an option to sell the synth for a few bucks and maybe even make some out of it. However, reading some posts on KVR Forums written by serious professionals, convinced me of the fact that making profit out of plugins as an independent developer is very difficult. So, freeware it is, less business – more pleasure!
So i started playing around with a bunch of sawtooths and it sounded quite horrible. I’m not a DSP-guru of any sort and had the bitter sweet sensation of learning about aliasing the hard way. I learnt what it takes to generate better sound, but even the simplest thing such as generating a sawtooth waveform became complicated as hell. Later i learnt that unintended noise, distortion and aliasing lurk everywhere in audio DSP, just waiting to drive a poor engineer insane.
Frankly, i’m an electronic music enthusiast who’s been making noise out of computers all his life. I’ve been fascinated by the math behind sound and music. I’m also fairly experienced programmer. So, here we go.
A few months back, i had a good deal of audio plugin code lying around, a couple of hobby projects that i never got round to finish. They weren’t really made to be published, but the fact that i never released any of that work was kind of irritating, so i decided to put a synth plugin together.
I’ve always thought of Roland Juno-106 as the ultimate music making machine of all times. It’s just so inspiring to tweak and it’s next to impossible to spend more than five minutes on a 106 without cracking a laugh or two. A 106 can produce quite impressive screeks and screams when mishandled correctly. So i got some inspiration from that gem of synth hardware.