DSP Robotics Support

I'm super-amazed you can just do stuff like that Martin

Plus, I'm surprised that the DSP code version is still very low on CPU.

Cheers

Spogg

I wish I could also play that thing, not only code it.
Anyway, it needs a Leslie cabinet to sound right. Anyone?

martinvicanek wrote:Anyway, it needs a Leslie cabinet to sound right. Anyone?

Ive been wanting to do a rotary speaker for a while now but never got around to it yet

watching this

Leslie is definitely on my future project list too.

Or at least something generally related to 'moving loudspeakers' .. over the decades there have been other examples such as Allen's 'Gyrotone', Compton's 'Rotofon' etc. These two both feature travelling loudspeakers with slip-ring contacts, whereas in a Leslie (as I'm sure everyone knows) the speakers are stationary and the sound moved using rotating horns.

But in every case the objective remains to generate, in 3D, modulated side-bands from what is an essentially mono, phase-locked (=rather boring!) sound source. Because that's all that electric organs could manage back then.

Actual Leslie simulations have been pretty effectively covered elswhere in recent years, both in hardware & software; they invariably try and copy the unique features of the original cabinets - horns, rotors, crossovers, cabinet resonances and so on so I'm tempted to try and move on rather than cover the same ground again.

Back in the late 80s I designed a hardware circuit to simulate rotating speakers, as I recall using 3-phase bucket brigades (don't ask..). It was quite effective and I've long been intrigued how the priciple could be expanded and improved upon. When I have a mo ...

Fascinating!

I guess the challenge is to make as accurate an emulation of a Leslie as possible, by whatever means is necessary.

I wonder if convolution could help, like with normal speaker enclosure measurements. Record white noise going through an actual Leslie and use that as the source. Not easy though!

Cheers

Spogg

Back to Martin's incredible proof-of-concept .fsm. I'm still baffled, Martin, how you managed to come up with TWO such neat & spectacular versions - assembler & code - at such speed Respect!

I've only just got back to studying it all, and alas my knowledge is coming up short in several areas .. but anyway I'll keep at it for a bit longer before I begin the stupid questions

A couple of observations, if the project's to be taken further :
The harmonic level addition will need attention; at the moment it's linear (2+2=4), which results in over-adding. You can hear this if you draw 16, 8, 4 & 2 drawbars, hold down C2, hold down C4, and then play up A#2, B2, C3, C#3 ... you'll hear C3 sounds much louder than it should, because on that note several drawbars are being added.
What's needed is 2+2=3! It's achieved in a real Hammond through resistive mixing - the more notes that are adding up the bigger the load, so the summing diminishes in proportion. It'll be pseudo-logarithmic or something. Sums.

Percussion needed - omg - which inevitably brings in attack & decay envelopes. Actually I'm very interested in expanding this principle to do other kinds of organ, more harmonics, multi-generators (particularly if they're so CPU efficient!) where slower attacks and decays would be required.

But I reckon everything else in clonewheel-land - chorus/vibrato, Leslie and so on are relatively easily achievable at the 'audio' end of things.

So I have spent some more time with the organ thing: added percussion and some modulation options plus a reverb and a distortion unit. It is not intended to be a Hammond clone, only inspired by that legendary instrument. To me it sounds OK, therefore I removed the "proof of concept" suffix. But then I am not a keyborder, so I don't know if it is any good at all.
Opinions?

Just a quick "Werrheeyy!!" What a big surprise this morning, wasn't expecting this!

At first glance this looks like it works really well, love all the added effects. Alas I have to go and do other stuff today so I'll have to come back to it tomorrow (or soon..) and then I'll check it out in more detail.

Well done indeed.

H

martinvicanek wrote:... I am not a keyborder[*], so I don't know if it is any good at all.
Opinions?

I love it, love it, LOVE it!!

I'm not an expert on organs like Hugh, but I think this is a commercial grade product Martin, up there with the best.

Well done Sir!

Spogg

[*] You are definitely NOT the surrounding edge of a key