Forum rss-feed

Forum

General Discussion: Accelerometer in Tau and Alpha

Most Recent

written by: dhjdhj

I am drooling at the possibility of leveraging acceleration for live performance. I already wave the thing around on stage when I'm doing aftertouch such that it looks like the alpha is responding to movement (I've actually been asked that question) but I'd love that to work for real.

written by: TheTechnobear

Sun, 1 Sep 2013 22:24:26 +0100 BST

Hi John,

After a G+ Eigenharp hangout today, Antonio mentioned to me that the Alpha and Tau had accelerometers fitted, and perhaps I should have a look.

so being inquisitive i did :o) (see test data below)

I then checked the forums, and saw your comment on: http://www.eigenlabs.com/forum/threads/id/352/

Ok, so I understand its limitations on the Alpha, and that the Taus may be better placed to use it.

Anyway, Im thinking of creating an output port on the keyboard to expose it.

I already know how to get to the data (same way as breath and strip).

but I wondered if you could tell me what p,r,y (pressure, roll, yaw) represent exactly in this context?
id assume they are something like acceleration forces in each axis?
... do you have any details? is the data raw from the accelerometer?

Im sure I can figure it out myself, but if you have anything that might be useful it would be appreciated


Thanks
Mark
================

background for others who are interested, on what Ive found so far ...

defined in alpha2_active.h,
#define KBD_ACCEL (KBD_KEYS+5)
Its marked as a sensor, and handled in same way as strips and breath.
(its also defined for the TAU)

there are no further reference in the source, BUT excluded in the key_bundle.
so I chucked in some debug code... and then started swinging the alpha around like a crazy fool!

some sample data...
eigend: : kb_key acell t:58289325749k:137p:1314r:1495y:1722
eigend: : kb_key acell t:58289327749k:137p:1315r:1563y:1708
eigend: : kb_key acell t:58289329749k:137p:1319r:1603y:1770
eigend: : kb_key acell t:58289330749k:137p:1322r:1597y:1821
eigend: : kb_key acell t:58289336788k:137p:1350r:1573y:1769

k=137 = 132 keys + sensor number 5
p,r,y = presumably force data of accelerometer (pressure, roll, yaw - presumably x, y z)

taking away a 1000, and dividing by 10, makes it a bit more readable:
eigend: : kb_key p: 65 r: 88 y: 8
eigend: : kb_key p: 64 r: 89 y: 20
eigend: : kb_key p: 65 r: 87 y: 34
eigend: : kb_key p: 64 r: 84 y: 47
eigend: : kb_key p: 63 r: 84 y: 60
eigend: : kb_key p: 63 r: 80 y: 70
eigend: : kb_key p: 64 r: 68 y: 74
eigend: : kb_key p: 65 r: 61 y: 75
eigend: : kb_key p: 65 r: 65 y: 68
eigend: : kb_key p: 65 r: 67 y: 60
eigend: : kb_key p: 65 r: 76 y: 66
eigend: : kb_key p: 65 r: 78 y: 72
eigend: : kb_key p: 65 r: 85 y: 80
eigend: : kb_key p: 64 r: 92 y: 88
eigend: : kb_key p: 64 r: 95 y: 95
eigend: : kb_key p: 63 r: 98 y: 103
eigend: : kb_key p: 63 r: 101 y: 109
eigend: : kb_key p: 63 r: 107 y: 113
eigend: : kb_key p: 63 r: 100 y: 115
eigend: : kb_key p: 63 r: 101 y: 109
eigend: : kb_key p: 63 r: 95 y: 111

I think this snippet is turning the alpha upside down (length ways)


written by: TheTechnobear

Mon, 2 Sep 2013 00:00:36 +0100 BST

ok, a bit more analysis...
sensors seem to be in range of 500-2400, some variation on axis.
Id assume there is some 'acceleration' element but I dont see it, i.e its almost like I only get the rotation force.

p = pitch = turn instrument right way up/upside down, sensor runs vertical, i.e. same axis as column of keys

r = roll = rotation axis along centre line. e.g. have the instrument on your lap horizontally keys up, and rotate keys down. this is like the yaw of a key. sensor is across they instrument, i.e. same axis as a row of keys

y = yaw = with instrument upright, tilt forward and backwards away from you, almost like an extension of pitch of the key, sensor is through the instrument (front to back) i.e. same as as pressing keys.

i think this is correct, though really without plotting the data on a graph its a bit hard to say for sure.
... though ive suspicion pitch is a not quite defined properly, i seem to be overlapping it with yaw... when really it should (i think) give more data rotating in the 3rd axis, (imagine rotating around on the harness)

perhaps I need to do a bit more research on acelerometers and this may clear up misunderstandings :o)

oh, I also assume the values are some kind of ranged values, i.e. no real meaning, but just a linear (?) scale


written by: dhjdhj

Mon, 2 Sep 2013 03:17:36 +0100 BST

I am drooling at the possibility of leveraging acceleration for live performance. I already wave the thing around on stage when I'm doing aftertouch such that it looks like the alpha is responding to movement (I've actually been asked that question) but I'd love that to work for real.



Please log in to join the discussions