Roger Linn Design

bhinton · Post by **bhinton** » Mon Oct 27, 2014 2:52 am

Hello Roger, Geert and Co!
I'm loving the linnstrument so far.

But here's an issue that I'm wondering about either a fix for or a pointer to something I'm missing:
On my linnstrument, pitch bend, regardless of the range setting on either the linnstrument or the synth used (usually, at the moment, Bazille), skips values with each slide across a gap between the squares. This occurs regardless of quantize being on, off, or its type, and regardless of technique (hard press slide, soft press, etc.) It's audible, and confirmed by looking at the values coming in via a midi monitor in bidule.
It IS sending 14 bit data - it just skips a ton of values between each square.
I know for many types of music this isn't a big issue, but in my picky postmusic experimental land of sound it usually means forgoing this feature, which would be an awful shame.
Advice appreciated!

- Brook
p.s. thank you for adding the advanced mode for CC assignments in 1.03!

gbevin · Post by **gbevin** » Mon Oct 27, 2014 7:33 am

Hi Brook,

Glad to hear that you're enjoying your LinnStrument and the ongoing software development!

I think I could improve the pitchbend jumps between cells in a firmware update, but it needs experimentation. One question though, do you mainly notice it when sliding slowly? Also, could you share a Bazille patch with me that clearly shows this for you. When experimenting with improvements, I like to test with real sounds that people work with.

Take care,

Geert

gbevin · Post by **gbevin** » Thu Oct 30, 2014 7:24 pm

Hi Brook,

We released v1.0.4 of the firmware which should reduce the gap between cell slides. It's not perfect yet, but it should be quite a bit less. Can you give it a try and let us know if it's better for you?

http://www.rogerlinndesign.com/linnstru ... sions.html

Take care,

Geert

bhinton · Post by **bhinton** » Fri Oct 31, 2014 2:04 am

Thanks! Will probably be sat/sun before I can test. Sorry I didn't get a patch to you. It's been a packed week, and not in the creative sense :/

gbevin · Post by **gbevin** » Fri Oct 31, 2014 6:29 am

Forgot to mention how to control the gap crossing. In v1.0.4 LinnStrument now uses the relative pressure on two adjacent cells to interpolate between the pitch on one edge of the gap and the pitch on the other edge. So by being conscious about your finger touching the edges of both cells, I find that you can very slowly and smoothly bring the pitch from one side to the other. As you move faster, you generally don't need the very smooth transition less since the gap will not be heard due to the rate of change.

bhinton · Post by **bhinton** » Mon Nov 03, 2014 5:59 am

It does work a little better... but it's still problematic.

Fast slides sound *much* better - but if they reverse direction within the same gesture, unless everything stays fast throughout the audible gaps are very distracting.
Bends across adjacent notes work better using the adjacent cell pressure technique you describe, but it's a VERY tough one to master, and when things are set so each cell is a half step (as I"m using using it) it's still very hard to get a fluid rise/fall or any kind of fine control to hit points between pitches. But definitely an improvement.

The big problem is slow slides - whether across two cells or across a wide range of cells (and the reverse direction problem is worse here too). Just no way around it, in sparsely arranged or solo passages it really sticks out. (again, I'm doing things like sustained solo bare sine/square wave-ish phrases - I can see how in the context of some music none of this would be all that noticeable.)

I guess I was expecting something like a violin string sweep or a slide guitar type of pitch bend between notes - this is kind of like teeny tiny frets dividing the pitch bend. Often sounds continuous in context, but alas, not enough, I guess, for some purposes. I can see how the design might make this just impossible, but poly pitch bend is part of the appeal in the marketing, so I suspect getting it as close as possible is probably worth the effort.

gbevin · Post by **gbevin** » Mon Nov 03, 2014 6:04 am

Thanks for the details Brook. One question, do you match the pitch bend setting the LinnStrument sends out with the pitch bend range that the synth accepts?

bhinton · Post by **bhinton** » Mon Nov 03, 2014 6:35 am

gbevin · Post by **gbevin** » Mon Nov 03, 2014 6:37 am

I guess you also performed the calibration procedure diligently and turn off quantize hold?

function · Post by **function** » Mon Nov 03, 2014 7:58 am

bhinton, Can you make a video - I'd like to see what you are talking about.

gbevin · Post by **gbevin** » Mon Nov 03, 2014 8:21 am

Btw Brook, I'm thinking through another improvement that might have very good results to fix this, but it will need quite a bit of testing to make sure it doesn't make matters works in other situations. I actually suspect that in the posts above we have a different definition of slow and fast. If I slide my finger extremely slowly, with an almost imperceptible pace, then as I hit the cell to the left or right there is indeed a little gap in the bend. This is due to the actual groove that's in between the cells and the fact that it's 1.5mm wide.

The improvement that I want to experiment with would essentially remove all the distances of the grooves using the calibration data, like virtually glueing the cells right next to each-other without any grooves. It might require a more elaborate calibration phase though, with two additional rows that need to be measured. I suspect that just the top and bottoms rows introduce too much deviation in the middle for this approach to work.

I'll keep you posted, I expect to have something for you to try out by next weekend.

Roger_Linn · Post by **Roger_Linn** » Tue Nov 04, 2014 11:23 pm

Hi Brook,
Sorry for jumping in late. Would you mind having a FaceTime or Skype video call with me? I'd like to know better exactly what you're doing and hearing? If that's OK, I'm here most of the day so please email me at support [at] rogerlinndesign.com with a suggested day/time to call. Write "To Roger" in the Subject so I'll answer it instead of Neal.
Thanks

gbevin · Post by **gbevin** » Sat Nov 08, 2014 7:30 am

Hi Brook,

I think a have found a good algorithm that to my ears completely removes any side effects of the gaps between the cells for the pitch slides. I'd like you to test a pre-release version of the firmware to see if this fixes it for your playing style also. Can you send me your email address in a private message, then I'll send you the firmware update?

Take care,

Geert

bhinton · Post by **bhinton** » Sun Nov 09, 2014 6:47 pm

Initial, brief test: this is miraculous. Completely smooth pitch bends, no gaps. Will report further when I have some more testing time. Thank you thank you thank you!

Roger Linn Design

KVR Audio

Pitchbend Question