[sdiy] Alternatives to Faders

[Edward King]

Comments added in-line

Regards

Edward


----- Original Message ----- 
From: "Tom Wiltshire" <tom at electricdruid.net>
To: "Edward King" <edwardcking2001 at yahoo.co.uk>
Cc: "Synth-DIY (list)" <synth-diy at dropmix.xs4all.nl>
Sent: Wednesday, March 07, 2007 7:40 PM
Subject: Re: [sdiy] Alternatives to Faders


> Edward,
>
> On 7 Mar 2007, at 15:08, Edward King wrote:
>
>> you know the pitch-bend and modulation wheels you normally get on synths?
>> Thats the way Im going at the moment...although the wheel Im using is 
>> okay, Im having problems finding wheels / dials that are similar to 
>> modulation wheels so if anyone knows of any, I would be very grateful. 
>> They need to be roughly 10mm wide and the same diameter (or close to it) 
>> as modulation or pitch-bend wheels.
>
> I don't know of any, but I recently looked into buying 10mm acrylic sheet 
> with a view to cutting circles out of it for exactly this purpose. I was 
> only going to need two, so thought to do it by hand, but if you needed 
> lots, you could probably get someone to machine them for you.
>

This is probably a good idea. I have actually found some wheels. But the 
difficulty is getting ones with the grooves cut into them. Im not sure how 
you would machine these grooves into acrylic?


>
>> The only issue Ive found is that the level display is inadequate for fine 
>> control. Consequently, Ive come up with the following idea:
>> for each track, 2 x 3 digit 7 segment displays, one above the other. The 
>> upper one should be the level (i.e 0 to 255 or 0 to 100) with the lower 
>> one displaying "fine tune" (again 0 to 255 or 0 to 100). Using both 0-255 
>> gives us 16 bit level control and coupled with a rate button, this gives 
>> us a reasonable amount of control with enough to response rate to make 
>> them practical.
>
> 0-255 makes sense to electronic engineers and computer geeks, but doesn't 
> make much sense to people who don't count in hexadecimal.

I agree. Thats why I said 0-255 or 0-100
The actual value displayed doesnt have to be the value that gets as far as 
the audio generation though because the input passes through the MPU on the 
way.
Interpolation and LUT's can lend a hand here.

>
> I have also struggled with the problem of how to represent high resolution 
> digital values (even 10 or 12 bit is hard) using LEDs or such like. One 
> thing that is worth remembering is that analogue controls don't really 
> display the position with the theoretical infinite resolution they're 
> capable of. Consider: My Korg Polysix has scales marked 0-10, with 5mm 
> between tickmarks. Assuming the human eye can differentiate positions down 
> to 0.5mm, I could recognise only 100 different positions on the scale. 
> Even assuming that I could recognise positions only 0.05mm apart (highly 
> unlikely) only gives me around 10 bit resolution.
>
> One solution that I wondered about was to use bicolour LEDs for the level 
> display, with one colour representing the coarse value, and the other 
> colour representing the fine adjustment. Some rotary encoders include 
> switches so that you could push-and-turn for fine adjustment (or coarse, 
> whichever way round you prefer). Still, you'd need probably 32 LEDs per 
> control to get decent resolution this way.


This idea (coarse and fine sets of LED displays) is the same as the one I 
posted above. In my case I have played with using Red 3 digit 7 segment 
displays on the upper  - for coarse - and green 3 digit 7 segment LEDs for 
the lower - or fine.


> A final problem that your recycled-mouse idea might have avoided is that 
> commercial rotary encoders often don't have that many positions per 
> rotation - 24 pulses per rotation is a common value, although by no means 
> the only one available. This makes doing large changes of high resolution 
> controls quite difficult. If the control has a 16 bit resolution and one 
> full turn takes you from 0 to max (65536) then each of the 24 clicks 
> represents 2730 units. How then to make a parameter value jump 2730 units 
> without an abrupt change?
>
> Sorry not to offer more positive suggestions, but this is definitely still 
> a problem for which a really good solution is yet to be found.

I think theres some confusion here about position sensing and rate. Ball 
mice feed rate data back (the amount that their position has changed since 
the last 3 byte value) not absolute position.
Because you're measuring rate of change rather than absolute position, 
moving the wheel slowly gives you fine control and moving it quickly gives 
you coarse control.
If you really want the parameter to jump from 0 to max, just flick the wheel 
forward and you get the abrupt change you desire.

As I said in my last post, Im happy to send pictures, research material and 
experimental software and data to any who would like to try this out for 
themselves. I would invite others to do so because it performs pretty well.

I did look at using rotary encoders, but they have a limited lifespan and 
arent really conducive to what Im trying to do.


>
>
> ++++ Electric Druid ++++
> Web Design & Development
> http://www.electricdruid.net
>
> 


		
___________________________________________________________ 
Try the all-new Yahoo! Mail. "The New Version is radically easier to use" – The Wall Street Journal 
http://uk.docs.yahoo.com/nowyoucan.html