written by: Bjoern
Hi 0beron, hi all,Well, I played around with the "by accident" working setup "alternative 2" a bit - this is fun and doesn't suffer from hanging notes!
"alternative 5" or 0beron's sax setup are also nice, they are not really playable for a longer time atm. though because of hanging notes on my side. But the concept definitely heads into a very promising direction (many possibilities I didn't even think of), so as a "proof of concept" this was a big step forward!
I managed to play a tune using the grey code fingering on my Pico! There's potential here.
The biggest concern I have at the moment is that the instrument is responding extremely slowly to breath articulation - there's a very noticeable lag which really impedes the ability to articulate anything. I can slur around to my heart's content, but one cannot play everything with slurs.
As a test I pulled out the breath pipe and blew directly into the hole, and found no observable difference in lag, so perhaps it's a threshold value or something. I did muck about with the breath parameter threshold on the fingerer and it's not that.
I really hope that this is something to do with the early software or my setup, not an intrinsic thing to do with the breath sensor and the way it's constructed.
Hi Matt
Do you get the delay when you use a different oscillator? The clarinet oscillator is a physical model and treats the breath pressure as a driving energy, so (in the same way an actual resonant tube/reed does) it takes a while to respond to changes. I'm interested to see if you experience the same effect when using an oscillator that doesn't work that way, maybe you could have a play with the sawtooth oscillator and let us know your results?
The breath sensor is low latency by the way, and high bandwidth. Sam and Dave (both horn players by training and the first Eigenharp players and testers) both used to get very expressive notes from it and used a variety of wind techniques with it, including very rapid tounging. Fingerer does not introduce any additional latency in the signal chain, in case you were wondering.
And could you let us know which instrument you're playing?
John
Thanks John, it's reassuring to know it's not a hardware limitation.
I'll try it with another oscillator, although I can't believe a real clarinet takes this long to respond to breath changes. I have no actual experience there though, as my wind experience is limited to recorders, crumhorns, EWI and WX5, all of which are pretty much immediate in response (except for the larger recorders and crumhorns).
My instrument is a Pico.
Hi Matt
Real tube/reed systems do actually take quite a long time to start resonating. Wind players compensate naturally for this - it's just part of the learning process they go through (often when quite young) and most pick up the need to anticipate the timing a little without ever being aware of it. I've not heard anyone who's used to traditional wind instruments even comment on the clarinet behaviour, so it's quite possible that this is the issue you are experiencing. Please do let us know what happens when you use a more straightforward oscillator, I'd like to be sure this isn't a more fundamental issue somewhere. I'd also like to hear from any wind players out there as to their opinions on the clarinet oscillator - it's possible it could be made more 'lively' though this might come at the expense of veracity.
You can also try playing with the shaper (usually wired in front of the oscillator) parameters - the shaper acts as a kind of compressor and will help you to make things more sensitive if you want it that way.
John
I do find the clarinet oscillator is quite slow to start in its default configuration, and when I found the controls for the shaper (if you load up stage then there are some controls in the Clarinet tab), adjusting the compression seemed to help a lot to make it more responsive. I guess you are sacrificing some of the dynamic range of the breath sensor by doing this, but I find it much nicer to play this way.
On the WX5, there are several patches on the associated physical modelling synth ( the VL70m ) which are very responsive indeed, and equally there are others that are unplayable except for really slow legato phrases or sound effects since the attack time is so long.
I did some experiments, thinking about what John said and what 0beron said, and the cause seems to be the Shaper. Playing with the Shaper parameters in Stage was fairly unsatisfactory in outcome (although I didn't do it for very long and applied the 'ham-fisted' approach). It certainly didn't do much for the response time.
Then I hooked up a sawtooth oscillator - scaler frequency out to sawtooth frequency in, scaler pressure out to sawtooth pressure in, sawtooth audio out to summer audio in - and played that.
Wow. Instant response. Very nice. Well, not very nice as a raw sawtooth wave is not the most pleasant thing and the attack was very harsh indeed, but instant noise. Like.
So I thought what if it's the Shaper's fault? Hooked a Shaper up between Sawtooth and Scaler, and there was the lag again.
Hooked up the Clarinet without the Shaper and that's now playable too. It's not immediate, but it's not glacially slow either so it feels like a usable model, something akin to a bass or maybe a great bass recorder. I wasn't able to do anything odd to it, so my question would be, why would I need a Shaper?
By the way, I also upgraded to 2.0.42-experimental and it's much nicer! So much more stable and just feels better manipulating Workbench and everything. Well done!
Hi all, I've made some further updates to my WX5 Sax B layout. Here it is:
[electric sax]
finger 1 = 1,1 1,2 1,3 2,4 2,5 2,6 2,7 1,8 2,8 * +5.5 ; Bb
finger 2 = 1,1 1,2 1,3 2,4 2,5 2,6 2,7 2,8 * +5.5 ; Bb
finger 3 = 1,1 1,2 1,3 2,4 2,5 2,6 2,7 1,8 * +6.0 ; B
finger 4 = 1,1 1,2 1,3 2,5 2,6 2,7 1,8 * +6.5 ; C
finger 5 = 1,1 1,2 1,3 2,4 2,5 2,6 2,7 * +7.0 ; C#
finger 6 = 1,1 1,2 1,3 1,4 2,5 2,6 2,7 1,8 * +7.0 ; C#
finger 7 = 1,1 1,2 1,3 2,5 2,6 2,7 * +8.0 ; scale root (D by default on WX)
finger 8 = 1,1 1,2 1,3 2,6 2,7 * +8.5 ; Eb
finger 9 = 1,1 1,2 1,3 2,5 2,6 2,7 2,8 * +8.5 ; Eb
finger 10 = 1,1 1,2 1,3 2,5 2,6 * +9.0 ; E
finger 11 = 1,1 1,2 1,3 2,5 * +9.5 ; F nat
finger 12 = 1,1 1,2 1,3 2,7 * +10.0 ; F#
finger 13 = 1,1 1,2 1,3 2,6 * +10.0 ; Alternate F#
finger 14 = 1,1 1,2 1,3 2,5 1,6 * +10.0 ; Weird Alternate F#
finger 15 = 1,1 1,2 1,3 * +11.0 ; G
finger 16 = 1,1 1,2 1,3 2,4 * +11.5 ; G#
finger 17 = 1,1 1,2 2,5 * +11.5 ; G#
finger 18 = 1,1 1,2 2,6 * +11.5 ; G#
finger 19 = 1,1 1,2 2,7 * +11.5 ; G#
finger 20 = 1,1 1,2 2,6 2,7 * +11.5 ; G#
finger 21 = 1,1 1,2 * +12.0 ; A
finger 22 = 1,1 2,2 * +12.5 ; Bb
finger 23 = 1,1 1,3 * +12.5 ; Bb
finger 24 = 1,1 2,5 * +12.5 ; Bb
finger 25 = 1,1 2,6 * +12.5 ; Bb
finger 26 = 1,1 2,7 * +12.5 ; Bb
finger 27 = 1,1 * +13.0 ; B
finger 28 = 1,2 * +13.5 ; C nat
finger 29 = 2,5 2,6 2,7 1,8 * +13.5 ; C nat
finger 30 = 1,1 2,5 2,6 2,7 1,8 * +13.5 ; C nat
finger 31 = 1,2 2,5 2,6 2,7 1,8 * +13.5 ; C nat
finger 32 = 1,3 2,5 2,6 2,7 1,8 * +13.5 ; C nat
finger 33 = open * +14.0 ; C#
finger 34 = 1,4 2,5 2,6 2,7 1,8 * +14.0 ; C#
finger 35 = 1,1 1,4 2,5 2,6 2,7 1,8 * +14.0 ; C#
finger 36 = 1,2 1,4 2,5 2,6 2,7 1,8 * +14.0 ; C#
finger 37 = 1,3 1,4 2,5 2,6 2,7 1,8 * +14.0 ; C#
finger 38 = 2,4 2,5 2,6 2,7 * +14.0 ; C#
finger 39 = 1,1 2,4 2,5 2,6 2,7 * +14.0 ; C#
finger 40 = 1,2 2,4 2,5 2,6 2,7 * +14.0 ; C#
finger 41 = 1,3 2,4 2,5 2,6 2,7 * +14.0 ; C#
finger 42 = 2,5 2,6 2,7 * +15.0 ; Top D
finger 43 = 1,1 2,5 2,6 2,7 * +15.0 ; Top D
finger 44 = 1,2 2,5 2,6 2,7 * +15.0 ; Top D
finger 45 = 1,3 2,5 2,6 2,7 * +15.0 ; Top D
finger 46 = 2,6 2,7 * +15.5 ; Top Eb
finger 47 = 1,1 2,6 2,7 * +15.5 ; Top Eb
finger 48 = 1,2 2,6 2,7 * +15.5 ; Top Eb
finger 49 = 1,3 2,6 2,7 * +15.5 ; Top Eb
finger 50 = 2,5 2,6 * +16.0 ; Top E
finger 51 = 1,1 2,5 2,6 * +16.0 ; Top E
finger 52 = 1,2 2,5 2,6 * +16.0 ; Top E
finger 53 = 1,3 2,5 2,6 * +16.0 ; Top E
finger 54 = 2,5 * +16.5 ; Top F nat
finger 55 = 1,2 2,5 * +16.5 ; Top F nat
finger 56 = 1,3 2,5 * +16.5 ; Top F nat
finger 57 = 2,6 * +17.0 ; Top F#
finger 58 = 1,2 2,6 * +17.0 ; Top F#
finger 59 = 1,3 2,6 * +17.0 ; Top F#
finger 60 = 2,7 * +17.0 ; Top F#
finger 61 = 1,2 2,7 * +17.0 ; Top F#
finger 62 = 1,3 2,7 * +17.0 ; Top F#
finger 63 = 1,3 * +18.0 ; Top G
finger 64 = 1,3 1,4 * +18.5 ; Top G#
modifier 1 = 2,1 * +12.0
I think this is now a more or less complete set of fingerings from the sax B reference pages, plus a few combinations from Sax A that I know work in Sax B despite not appearing in the manual!
I've only put on a single octave key since I've no idea really how to expand the range using octave modifiers without tying your fingers in knots. Can eigenlabs design and build the 'eigenharp femto' or something (just 5 or 6 eigenkeys in a little standalone stick with a USB port) so I tape one onto the back of the pico (I'm kidding)...
I've got a little python script that stores my fingering and means I can offset the key numbers and compact the representation a bit. This is designed to be connected to a scaler set to the major scale, and with the tonic set to D (+2.0 semitones).
Also, since moving up to version 2.0.44, the stuck keys issue seems to have gone away, the fingerer is now pretty playable!
For reference, I'm playing it into the clarinet oscillator, with the pressure signal sent through a shaper with the following settings:
Compression: 0.65
Sharpness: 0.23
Thanks for this, it will be useful
I suspect that sax fingering is not the best context for making a lot of use of the modifiers etc.
In contrast, you can get 16 notes out of 4 keys and a binary pattern which leaves the other hand free to add intervals to those notes and build chords using poly modifiers.
If poly modifiers supported multiple intervals (and negative intervals), you could define keys for commonly used chords (and various inversions)... So you could play G and add the Maj7+9 bit with one other key
Here's one I've been working on along those lines
The notes run in a circles of fifths, so you have C, F and G on neighbouring keys making a I-IV-V pattern really easy. Then by pressing two neighboring keys you can get another circle of fifths offset by a major third. You could extend this idea with three key combinations, or two keys 2 keys apart etc, but with each new circle of fifths getting shorter overall.
Another discovery - if you hook up a fingerer to a scaler, and then change the scaler's tonic, scale, base note etc, you won't hear any effect at all until you change the fingering mode on the fingerer. Seems like at the time of reading the fingering files some info is taken from the scaler and stored as state inside the fingerer? Does the upstream scaler need to notify the fingerer that its scale has been altered?
[accordion chord]
# Use with a scaler in Chromatic mode.
# Db 2
# Ab 9
# Eb 4
# Bb 11
# F 6
# C 1
# G 8
# D 3
# A 10
# E 5
# B 12
# F# 7
# C# 2
finger 1 = 1,1 * +9.0
finger 2 = 1,2 * +4.0
finger 3 = 1,3 * +11.0
finger 4 = 1,4 * +6.0
finger 5 = 1,5 * +1.0
finger 6 = 1,6 * +8.0
finger 7 = 1,7 * +3.0
finger 8 = 1,8 * +10.0
finger 9 = 1,9 * +5.0
finger 10 = 1,1 1,2 * +1.0
finger 11 = 1,2 1,3 * +8.0
finger 12 = 1,3 1,4 * +3.0
finger 13 = 1,4 1,5 * +10.0
finger 14 = 1,5 1,6 * +5.0
finger 15 = 1,6 1,7 * +12.0
finger 16 = 1,7 1,8 * +7.0
finger 17 = 1,8 1,9 * +2.0
polyphony 1 = 2,5 * +3.0
polyphony 2 = 2,6 * +4.0
polyphony 3 = 2,7 * +7.0
polyphony 4 = 2,8 * +10.0
polyphony 5 = 2,9 * +11.0
polyphony 6 = 2,4 * -5.0 ;Doesn't work yet
polyphony 7 = 2,3 * -7.0 ; Ditto
polyphony 8 = 2,2 * -8.0 ; Ditto
Hi 0beron, hi all,
> Can eigenlabs design and build the 'eigenharp femto' or something (just 5 or 6 eigenkeys in a little standalone stick with a USB port) so I tape one onto the back of the pico (I'm kidding)...
Well :-) woodwind instruments of course do have keys for one thumb, on occasion even both thumbs.
Making a femto or pico with keys on the back is a whole new design, but I was wondering about the possibility of putting simple (flat) touch contacts on the back of the pico, and feeding that into EigenD.
So that's 1. a little electronics project (touch contacts to usb) - any thoughts or design ideas?
but 2. also a programming task, to get additional information into EigenD - any thoughts on that?
Cheers!
(Actually, the 2nd part of my post raises the general question whether other interfaces / control surfaces can provide input to EigenD... )
Please log in to join the discussions
UK
