[sdiy] Additive Synthesis - Wendy would say it works...

[Les Mizzell]

>>Ed asked (off list) why the K5K is close but no banana.

The K5 voice structure was actually *very* clever. It wasn't as additive 
as it really looked!

1. Basically, there were 4 digital oscillators, each with 1 EG.
2. On the main screen, you set the initial level of 64 (63?)
    harmonics individually.
3. Then, you assigned each of those to one of 4 EGs

Here's the "clever" part that made it look more complicated than it was 
- When you assigned a harmonic to an EG - it also assigned which of the 
4 OSC's were generating that specific harmonic, but that was part of the 
process you didn't "see" when programming.   So, you got a big graph 
showing the combined harmonic content of the 4 OSC's on one page. You 
never got to see an OSC anywhere - only only dealt with the individual 
harmonics and which EG was applied.

They could have done a completely different user interface like:

OSC 1 --> EG 1
OSC 2 --> EG 2
OSC 3 --> EG 3
OSC 4 --> EG 4

  - with a page for each OSC where you defined the harmonic content for 
each OSC and it wouldn't have been the same synth.

There was also a pitch EG, a master EG, EQ and some other nice stuff 
build in as well.

You could get an absolutely killer B3 out of it with a little work - but 
it seemed to excel in glassy sounding bells and the like. It would do 
interesting pads, but "warm" isn't something I would use to describe 
hardly anything that came out of it. I had a reasonable amount of 
success looking at FFT graphs in books and getting something that 
sounded sorta like what I was going after when programming, with 
emphasis on the "sorta".

It was a pretty cool idea, but the outputs had a rather low level and 
was pretty noisy. I'd love to see a more modern implementation of this 
with improved filters and such - but I doubt it ever happens in 
hardware. Software is the way to go for doing this, and a few folks seem 
to have done a pretty good job with it...



-- 
-----------
Les Mizzell