DSP Robotics Support

Hugh, maybe you could clarify something for me please:
All the Hammond drawbars go from left to right in ascending harmonic order, except for 5 1/3’ which seems like it should go between the 8’ and 4’, based on no more than what seems orderly to me.

Any idea why?

Cheers

Spogg

Here is something I put together along the lines described by Hugh. I misused the poly section to only generate a bit pattern corresponding to the the keys pressed, while the entire sound generation and overtone assignment takes place in (packed) blue. It turns out to be rather efficient in terms of cpu load, although the code is rather extensive (61 keys, each capable of producing up to 9 harmonics, to be assigned to 96 constantly running sine generators).

Quite remarkable Martin!

It works, it sounds great and the CPU barely wakes up. Unbelievable that you made this in such a short time too.

If you made this initially in DSP code, then optimised it, I'd love to see the code so I can stand a chance of seeing what you've done. I understand if you'd rather not release the "source code" though.

You never cease to amaze me Martin, you're a real hero to me. I wonder what Hugh will say...

Cheers

Spogg

Hugh agrees with Spogg and says, "Unbelievable!!" Pheww.

And seeing the tiny CPU hit means this clearly has immense potential. I'm about to go away for a couple of weeks - Android-land only (probably just as well ..) - so it'll be a while before I can remotely try and get my head round what Martin has done here, and hopefully take another step.

Inspiring as ever Martin!

On Spogg's 5 1/3' drawbar question, there's a short answer and a long answer ..
Short answer is that 5 1/3' is not part of the 8' harmonic set, so had it been placed between 8 & 4 it would likely have confused everyone who didn't grasp this fact. Clearly trying to set up tones at regular 8' piano-pitch, but including the non-harmonic 5 1/3' pitch in them would only have led to .. dissapointment.

For the long answer we need to ponder what the original Hammond designers in 1933 were having to contemplate. The 8' harmonic series goes 8, 4, 2 2/3, 2 etc., and they evidently decided to provide the first 8 harmonics, but leaving out the 7th. (1 1/7) whose pitch sits between A# & B, and which sounds horrid when 'borrowed'.
Ideally they probably also wanted to provide the 16' series of harmonics : 16, 8, 5 1/3, 4, 3 1/5, 2 2/3 etc. Note that every other one of these pitches also occur in the 8' series, so they already had them covered.
Ultimately they had to make economic decisions : each drawbar needs 61 contacts, another bus, and another mile of labour intensive cross-wiring, so there must have been an executive decision to settle for the compromise of the 9 drawbars we ended up with.
And the arrangement? - the 16' series (just 2 of them!) to the left, and the 8' series (7) to the right.

And just as a PS (- not a lot of people know this) ... have you spotted that 2 2/3 = 8/3 , and it's the 3rd. harmonic of 8'. Similarly 3 1/5 = 16/5, which is the 5th. harmonic of 16'. 1 1/7 = 8/7, it's the 7th. harmonic of 8'. And so on.
For that matter, 2 = 8/4, so it's the 4th harmonic of 8', and also =4/2, so it's the 2nd. harmonic of 4'.
It's kind of obvious once you spot it!

Cheers all.

Thanks for the detailed explanations and it makes some sense now.

Cheers

Spogg

Spogg wrote:If you made this initially in DSP code, then optimised it, I'd love to see the code so I can stand a chance of seeing what you've done.Spogg

No, I don't have a code version. Let me see if I can make one. It will be a mixed version, thoungh, because I use some asm features which are not available in code.

martinvicanek wrote:No, I don't have a code version. Let me see if I can make one.

OMG! That means you can actually write your stuff directly in Assembler. Another incremental increase in respect from me

Cheers

Spogg

Yeah, I am such a nerd you know.

Okay, here is te code version - still suprisingly lean.
I think one of the essential ingredients is that by design, the harmonics are not precisely integer multiples of the fundamental. Makes is sound less sterile.

Martin, you're brilliant. ( And the bonus is I can almost understand some of it now as well )

Sadly I have to go awol for a couple of weeks now, but I'll certainly get into this when I return.

Oh .. by the way .. small detail I'm sure, but here it's playing an octave low, i.e. 8' is playing 16'. Basso profundo.

Bye for now, and thanks again for your incredibly rapid effort.