Aftertouch problem

Karamel1 · 6 September 2017 15:14

Hello ,

I am trying to implement the aftertouh function with csound… The opcode aftouch behaves strangely.

I get from time to time an error message :undefined sys_common msg f4 do you have any idea of what it is ?

Also , the opcode aftouch seems to be badly updated : I compare it with Pure data : when pure data starts at zero and respond very well to the pressure on my keyboard, the Csound opcode can stay on a value it has received previously.

Does anyone know how to implement aftertouch correctly with Csound ? or is this opcode buggy ?

Thank you for help.

rorywalsh · 6 September 2017 15:43

I think @iainmccurdy is the best one to comment on this. I’ve never used after touch with Csound.

iainmccurdy · 6 September 2017 17:37

I haven’t used aftouch in a long time but I give it a go when I get to a keyboard with aftertouch.
An alternative approach would be to read the data in using midiin and filter into the aftertouch data by using its status byte and channel (I can’t remember what the status byte for aftertouch will be but you could easily test to find out what this is by using midiin and print messages.

The status byte value appears to be 208 so you can try this and see if it works any better:

 kstatus,kchan,kdata1,kdata2	midiin
 if (kstatus==208 && kchan==1) then
  kAftouch	=	kdata1
 endif
 printk2	kAftouch

Karamel1 · 7 September 2017 07:56

Thank you @iainmccurdy . I encounter the same trouble : whn the note is released, then the aftertouch remains sometimes at the previous value and also some error messages : undefined sys-realtime msg XX XX being a value : fd, f4, f9 etc…

Also, from time to time, a note remains playing on for ever… which is annoying.
I am wondering how much compatible is midiin with the flags on csound option from cabbbage -n -d -+rtmidi=NULL -M0 -m0d --midi-key-cps=4 --midi-velocity-amp=5 . Is there a risk of double messages ?

Here is the code, what do I do wrong ? Thank you for your help :

<CsOptions>
-n -d -+rtmidi=NULL -M0  -m0d  --midi-key-cps=4 --midi-velocity-amp=5
</CsOptions>
<CsInstruments>
; Initialize the global variables. 
sr = 44100
ksmps = 32
nchnls = 2
0dbfs = 1


instr 130
 iNum notnum
 kstatus,kchan,kdata1,kdata2	midiin
 if (kstatus==144 && kchan==1) then
    if kdata2==0 then 
        kAft=0
    endif
 elseif (kstatus==208 && kchan==1) then
    kAft = kdata1  
 endif
  
 if kAft !=0 then 
   kAft_LFO lfo  kAft/128, kAft/12.8 
    else
    kAft_LFO =0 
 endif 
  
 kPitch = cpsmidinn(iNum+kAft_LFO)
 kEnv madsr 0.1,0.2,.85,0.3

 asig	oscil  p5, kPitch, 1
 outs  asig*kEnv, asig*kEnv
endin

</CsInstruments>
<CsScore>
;sine wave.
f 1 0 16384 10 1
i 130 0 z
</CsScore>
</CsoundSynthesizer>

Karamel1 · 2 October 2017 16:01

Hello Iain, @iainmccurdy
Did you have the opportunity to dig into the aftertouch problem ?

I checked it today again : it seems that there is something wrong with the way the message are interpreted/updated at k-rate . The following code is included into cabbage instrument

 ; -- aftertouch --
kAfttch  aftouch 0,0.9               ; read in aftertouch (range 0 to 0.9)
printks  "aft : %f \n",0,kAfttch

First key is pressed : here is the output with aftertouch :
then
Cabbage console shows the following until the key is released :

aft : 0.900000 
aft : 0.900000 
aft : 0.900000 
...
aft : 0.900000

then pressing a new key deeper to unlock the aftouch shows :

aft : 0.900000 
aft : 0.481890 
aft : 0.481890 
aft : 0.481890 
aft : 0.481890 
...
aft : 0.481890 
aft : 0.481890

Comparing with midi-ox log clearly shows that aftertouch is either wrong or I can figure how to use it (most likely):

===> MIDI-OX Version: 7.0.2.372
 ===> Log Opened: Mon 02-Oct-2017 17:57:34 ===>
 TIMESTAMP IN PORT STATUS DATA1 DATA2 CHAN NOTE EVENT               
   2373046  1   4    144    77    30    1  F  5 Note On               
   2373085  1   4    208     8   ---    1  ---  Channel Aft           
   2373085  1   4    208    22   ---    1  ---  Channel Aft           
   2373085  1   4    208    32   ---    1  ---  Channel Aft           
   2373086  1   4    208    37   ---    1  ---  Channel Aft           
   2373086  1   4    208    42   ---    1  ---  Channel Aft           
   2373087  1   4    208    47   ---    1  ---  Channel Aft           
   2373088  1   4    208    54   ---    1  ---  Channel Aft           
   2373089  1   4    208    58   ---    1  ---  Channel Aft           
   2373089  1   4    208    63   ---    1  ---  Channel Aft           
   2373090  1   4    208    70   ---    1  ---  Channel Aft           
   2373091  1   4    208    74   ---    1  ---  Channel Aft           
   2373092  1   4    208    78   ---    1  ---  Channel Aft           
   2373092  1   4    208    82   ---    1  ---  Channel Aft           
   2373093  1   4    208    86   ---    1  ---  Channel Aft           
   2373095  1   4    208    89   ---    1  ---  Channel Aft           
   2373097  1   4    208    93   ---    1  ---  Channel Aft           
   2373098  1   4    208    96   ---    1  ---  Channel Aft           
   2373100  1   4    208   100   ---    1  ---  Channel Aft           
   2373102  1   4    208   104   ---    1  ---  Channel Aft           
   2373104  1   4    208   108   ---    1  ---  Channel Aft           
   2373107  1   4    208   111   ---    1  ---  Channel Aft           
   2373109  1   4    208   115   ---    1  ---  Channel Aft           
   2373113  1   4    208   119   ---    1  ---  Channel Aft           
   2373119  1   4    208   122   ---    1  ---  Channel Aft           
   2373124  1   4    208   126   ---    1  ---  Channel Aft           
   2373125  1   4    208   127   ---    1  ---  Channel Aft           
   2373461  1   4    208   126   ---    1  ---  Channel Aft           
   2373462  1   4    208   121   ---    1  ---  Channel Aft           
   2373465  1   4    208   118   ---    1  ---  Channel Aft           
   2373468  1   4    208   114   ---    1  ---  Channel Aft           
   2373471  1   4    208   111   ---    1  ---  Channel Aft           
   2373473  1   4    208   107   ---    1  ---  Channel Aft           
   2373476  1   4    208   103   ---    1  ---  Channel Aft           
   2373477  1   4    208    99   ---    1  ---  Channel Aft           
   2373478  1   4    208    95   ---    1  ---  Channel Aft           
   2373480  1   4    208    91   ---    1  ---  Channel Aft           
   2373481  1   4    208    87   ---    1  ---  Channel Aft           
   2373482  1   4    208    83   ---    1  ---  Channel Aft           
   2373483  1   4    208    79   ---    1  ---  Channel Aft           
   2373484  1   4    208    75   ---    1  ---  Channel Aft           
   2373485  1   4    208    71   ---    1  ---  Channel Aft           
   2373486  1   4    208    66   ---    1  ---  Channel Aft           
   2373486  1   4    208    60   ---    1  ---  Channel Aft           
   2373487  1   4    208    54   ---    1  ---  Channel Aft           
   2373487  1   4    208    49   ---    1  ---  Channel Aft           
   2373488  1   4    208    45   ---    1  ---  Channel Aft           
   2373489  1   4    208    37   ---    1  ---  Channel Aft           
   2373489  1   4    208    28   ---    1  ---  Channel Aft           
   2373490  1   4    208    20   ---    1  ---  Channel Aft           
   2373490  1   4    208    13   ---    1  ---  Channel Aft           
   2373491  1   4    208     0   ---    1  ---  Channel Aft           
   2373533  1   4    128    77    30    1  F  5 Note Off              
   2381763  1   4    144    76    30    1  E  5 Note On               
   2382081  1   4    128    76    30    1  E  5 Note Off              
 <=== MIDI-OX Log Closed: Mon 02-Oct-2017 17:59:17 <===

iainmccurdy · 3 October 2017 17:38

I didn’t get a chance to look at this yet! I’m still in Uruguay at the conference but I’ll try to investigate when I get back.

iainmccurdy · 17 October 2017 10:11

I finally got a chance to test this. I found the data output by aftouch within Cabbage 2 (Mac) to match that shown by a MIDI monitor program equivalent to MIDI-Ox. The only issue I encountered was that until aftouch receives some information from the aftertouch hardware it outputs its maximum value (127). I suspect this is an issue with the opcode but a workaround is to not use aftouch at all but instead take in aftertouch data in a separate always-on instrument using midiin as I think I suggested earlier and as you were trying also. Channel aftertouch uses status byte value 208 and it then uses data1 as its value. I am assuming you are using channel aftertouch as keyboards with polyphonic aftertouch are like unicorns. Here is the code:

instr	AlwaysOn
 kstatus,kchan,kdata1,kdata2 midiin
 if kstatus==208 then
  gkAftouch = kdata1
 endif
endin

This also assumes you aren’t concerned about what channel it comes in on.
I am not getting any of the problems with “undefined sys-realtime…” errors that you mentioned previously, nor I am getting problems with stuck notes. The problem you mentioned with aftouch where is holds its previous value is slightly expected as new MIDI data is only generated when that type of data changes. Having said this, my MIDI notes always end on a zero when I use aftertouch, which is what would be expected as releasing a note means releasing pressure. Can you test your code that produces errors with alternative hardware?

Referring to a previous bit of code you posted, maybe printks2 is preferrable to printks with zero seconds between printings. Cabbage has always struggled a bit with printing to its terminal, although I think this has been improved recently. So instead of:

printks "aft : %f \n",0,kAfttch

use:

printks2 "aft : %f \n",kAfttch