hi david -

sounds like you could use a GEN16 table, and step through it. in such a 
table, you can change the slope very easy by the "type" parameter.

best -

	joachim


Am 16.04.2014 15:02, schrieb David Mooney:
> Hello--
>
> Not being a mathematician, this has me stumped. I'm using a calculation
> to increment a variable in a loop that is driving event_i for another
> instr. The basic calculation gives me a nice exponential progression of
> numbers so that the variable changes one step for each instance of the
> event_i:
>
> loop:
>    event_i "i",2,idel,idur,ifreq
>
>    idel = idel+idur
>
>    ;calculate exponential progression for ifreq
>    iprog = (icnt-1)*(icnt-1)/((inst-1)*(inst-1))
>    ifreq = istart +iprog*(istart-iend)
>
>    ;increment the counter
>    icnt = icnt+1
>
> loop_lt indx,1,inst,loop
>
> By trial and error, it seems that changing the incrementation of icnt
> is the way to change the slope of the resulting curve:
>
> ;increment the counter
> icnt = (icnt+icntincr)+islope
> islope = islope+islopeincr
>
> icntincr is an increments icnt in place of "1", for example "0.1".
> islope and its incrementer, islopeincr, when added to the icnt
> incrementation
> changes the slope of the curve. What I'm stuck on is some way to
> calculate the value of islopeincr for a given icntincr to get an inst
> number of steps from istart to iend, rather than having to go the trial
> and error route to hit upon a value for islopeincr that works.
>
> Will appreciate any thoughts, or alternative ways to get to the same end.
>
> --David Mooney
>
> --
> Opaque Melodies
> http://opaquemelodies.com

hi david -

what i had in mind was something like this:



giCurve ftgen 0, 0, 1024, -16, 0, 1023, 6, 100
instr 1
;read the giCurve table with regular increments
iIncr = .1
iIndx = 0
loop:
iVal table iIndx, giCurve, 1
prints "Table value at position %.1f = %f\n", iIndx, iVal
loop_le iIndx, iIncr, 1, loop
endin


i 1 0 0


which returns:
Table value at position 0.0 = 0.000000
Table value at position 0.1 = 0.203493
Table value at position 0.2 = 0.573630
Table value at position 0.3 = 1.255675
Table value at position 0.4 = 2.487460
Table value at position 0.5 = 4.757245
Table value at position 0.6 = 8.856521
Table value at position 0.7 = 16.312757
Table value at position 0.8 = 30.052217
Table value at position 0.9 = 54.865938
Table value at position 1.0 = 100.000000


not sure this is what you need, but it's i-rate. best -
	joachim


Am 17.04.2014 00:12, schrieb David Mooney:
> Joachim--thanks.
>
> I'll give this a try, but not sure this will work. What led me to the
> calculation was the  requirement of event_i to use i-rate rather than
> than k-rate variables. I had been using a transeg envelope to fill a
> table (GEN -2), using a phasor index--similar to the example for GEN 16.
> When reading the table at k-rate, I can get the values just fine, but
> trying to step through the table to extract variables at i-rate doesn't
> really work because the phasor index values aren't integers. I tried a
> conversion like this:
>
> kndx phasor ......
> kndx = kndx*100  ;to go from normalized to raw
> kndx = int(kndx)
>
> but this didn't get the integer steps into the table I expected---
>
> [Would the index steps be 1/(table size)?--just thought of this.]
>
>   ---and i(kval) doesn't work inside a loop for event_i.
>
> --David Mooney
>
>
> On Wed, Apr 16, 2014 at 10:30 AM, joachim heintz  > wrote:
>
>     hi david -
>
>     sounds like you could use a GEN16 table, and step through it. in
>     such a table, you can change the slope very easy by the "type"
>     parameter.
>
>     best -
>
>              joachim
>
>
>     Am 16.04.2014 15:02, schrieb David Mooney:
>
>         Hello--
>
>         Not being a mathematician, this has me stumped. I'm using a
>         calculation
>         to increment a variable in a loop that is driving event_i for
>         another
>         instr. The basic calculation gives me a nice exponential
>         progression of
>         numbers so that the variable changes one step for each instance
>         of the
>         event_i:
>
>         loop:
>             event_i "i",2,idel,idur,ifreq
>
>             idel = idel+idur
>
>             ;calculate exponential progression for ifreq
>             iprog = (icnt-1)*(icnt-1)/((inst-1)*(__inst-1))
>             ifreq = istart +iprog*(istart-iend)
>
>             ;increment the counter
>             icnt = icnt+1
>
>         loop_lt indx,1,inst,loop
>
>         By trial and error, it seems that changing the incrementation of
>         icnt
>         is the way to change the slope of the resulting curve:
>
>         ;increment the counter
>         icnt = (icnt+icntincr)+islope
>         islope = islope+islopeincr
>
>         icntincr is an increments icnt in place of "1", for example "0.1".
>         islope and its incrementer, islopeincr, when added to the icnt
>         incrementation
>         changes the slope of the curve. What I'm stuck on is some way to
>         calculate the value of islopeincr for a given icntincr to get an
>         inst
>         number of steps from istart to iend, rather than having to go
>         the trial
>         and error route to hit upon a value for islopeincr that works.
>
>         Will appreciate any thoughts, or alternative ways to get to the
>         same end.
>
>         --David Mooney
>
>         --
>         Opaque Melodies
>         http://opaquemelodies.com
>
>
>
>     Send bugs reports to
>     https://github.com/csound/__csound/issues
>     
>     Discussions of bugs and features can be posted here
>     To unsubscribe, send email sympa@lists.bath.ac.uk
>      with body "unsubscribe csound"
>
>
>
>
>
>
> --
> Opaque Melodies
> http://opaquemelodies.com