Korg RADIAS: Cooked Meat


My reasons for buying a RADIAS are a bit iffy, but not entirely unreasonable. Right in the middle of the hilarity when I denounced the MS20 Mini, I was actually seeking a positive – what would Korg themselves consider their progress from that point – where had the MS20’s ideas gone after it was retired for the first time?


Straight after the MS period was a muddle – amidst a bunch of random ideas only the Poly6 and Mony/Poly were careful responses to the American keyboards of the time, but considerably cheaper. Korg took on this budget area with a succession of ever more hideous digital keyboards bottoming at the loathsome Poly800. It wasn’t until the M1 that they seemed to have any clear idea of where to advance themselves, and once they found it there was no shifting them from this ‘high end workstation’ identity. The baroque OASYS physical modelling system and the many splinters of it were the capstone of an edifice that Roland has tried to invade ever since.


Suddenly in 2000 we have the MS2000. Why? The full OASYS keyboard failed to sell, probably because Korg are good at potential but not good at designing its daily use. The big idea was sliced into workloads, and one of these (having belatedly reached the mainstream buyer) was the dance club; thus the Electribe, the Kaoss pad, and the 4 note MS2000. On the evidence it’s not so much an evolution as a de-evolution of an overblown idea and unless there’s been a hero in Korg ruining every design meeting with ‘What are we doing about aceeeeeeeed!?’, my interest is misplaced. Too bad.


No really, on to the RADIAS


Despite the complex history the RADIAS does present some kind of connection with the MS2000 and the MS system before it. There are some quality differences that our Polish friend can explain better than me; I’m really more interested in the oddness, and here it is.


Delete all the factory presets. They’re not terrible, but they make you think the wrong thing right away.


There’s a conscious attempt to lay out a MS style panel, with the 2 oscillators, 2 filters and so on, including a SQ10 down the bottom. I found it easy to set up my test MS sound which is a fake formant made by a hi pass and low pass combination. Not the same sound as the MS20 but pretty nice. To answer GearSlutz question number one: your tendency is to turn up oscillator 1 with your tricky sound then try to put a sine from oscillator 2 under it to add the bass. This will fail. Put the sine in first, then add the tricky sound gently into that. Now use the drive before the filter to warm up this mix and you are done. If you need it, turn up the bass EQ, that’s why it’s there. You now have the phat sound you always use for everything, so go away.


The rest of us can notice useful features spread all over the signal path in unexpected places. There’s a sub oscillator in the amplifier as part of the wave shaper, which must have been where it was needed technically, but logically should be visible alongside the others. Also in this section is ‘punch’ which adds a square wave to the attack only. Osc 1 is the fancy one with PCM samples and supersaws and the like – you’re not going to use the string sound much but the inharmonic spectra are pretty good when detuned as are the various waveform distortions. Osc 2 is less fancy but ready to detune, ring and synch.


While you can ring modulate the two main oscillators, there’s a stereo ring modulator in the effects which adds another LFO and saves you the voice. And in general the three effects sections are all part of the noise making circuit – while something like reverb is obviously for the master effects, the grain shifter seems part of the patch. I’m not dead sure but it also looks like using the FX doesn’t cut into the voice budget, whereas most knobby treatments will halve your polyphony. The strategy is not that of an analogue signal which overcomes circuits downstream, you instead work back and forwards across the whole flow to get the sound you want.


The phrase sequencer is new to me, as I was too poor to buy anything at the time the Electribe was in style. You arm a recording and then tweak a knob to make a ‘phrase’. Now the tweak is part of the sound patch – you can have it do the tweak on each key press, or keep doing it in a loop. Much easier than my SQ10, although the bottom row of knobs can be twiddled in the old way, and I have done so to create burblings that remind me of 1985.


Behind this Vocoder is a door to another dimension.


Right. Vocoder. Shit idea that restarted with the MS2000 and infected everything after with a bad case of Robot Concept Album. But there’s some strange things going on here as well. The first is Formant Motion, something that records up to 7.5 seconds of you yelling YO EVERYONE IN DA HOUSE but only as a modulator for the vocoder. Why wouldn’t you just make it a sound sample? Simpler, more versatile, would still work? I don’t get it, except the Korg engineer guys thought it was extra cool to convert your voice into phonemes. I have to get some time to throw a bunch of stupid sounds in there to see what it does.


Better still is that with much reading and re-reading of the manual I found that you can send a voice’s output to an internal bus, and then have that bus sent to the vocoder such that one voice vocodes another, which is what vocoders were always about in the good old days. So far it sounds horrible, which is a good sign for the future. It was only after I’d turned off the gear for the night that it struck me – what happens if you feed a voice back into itself? Will it explode? It was that moment I knew that the RADIAS is my kind of machine. If you don’t hear from me again – it was a glorious success.

We designed OASYS, the world’s most advanced synthesis system and you didn’t want it? Fuck you, have this crap instead.







It was either this or a red car and I think I chose wisely.