| Try/Modify these (the original soundfile was a 1 s marimba note at middle
C, but you can use whatever you like):
sr = 44100
kr = 44100
ksmps = 1
nchnls = 1
instr 1 ; vary pitch p4=pitch multiplier
iharm = int(44100/(262*p4)/2)-1 ;Chz max harm w/o alias
kenv linseg 0,0.01,1,1,1 ;get started right
kmovit linseg 0,p3,1.0029 ;1.0029 is duration of original file
krmsr,krmso,kerr,kcps lpread kmovit, "Marimba.lp"
axit buzz 1,kcps*p4,iharm,1,-1
araw lpreson axit
afinal gain araw, krmso*kenv,20
out afinal
endin
instr 2 ; play backwards & vary pitch p4=pitch multiplier
iharm = int(44100/(262*p4)/2)-1 ;Chz max harm w/o alias
kenv linseg 1,p3-0.01,1,0.01,0,1,0 ;get ended right
kmovit linseg 1.0029, p3, 100/44100,1,100/44100
;1.0029 is duration of original file, 100/44100 is to prevent reading past 0
krmsr,krmso,kerr,kcps lpread kmovit, "Marimba.lp"
axit buzz 1,kcps*p4,iharm,1,-1
araw lpreson axit
afinal gain araw, krmso*kenv,20
out afinal
endin
instr 3 ; hold pitch, shift formants p4=formant multiplier
iharm = int(44100/(262)/2)-1 ;Chz max harm w/o alias
kenv linseg 0,0.01,1,1,1 ;get started right
kmovit linseg 0,p3,1.0029 ;1.0029 is duration of original file
krmsr,krmso,kerr,kcps lpread kmovit, "Marimba.lp"
axit buzz 1,kcps,iharm,1,-1
araw lpfreson axit,p4
afinal gain araw, krmso*kenv,20
out afinal
endin
*******Score****************
f1 0 16384 10 1
i1 0 2 0.5
i1 2 2 0.75
i1 4 1 1
i1 5 1 1.5
i1 6 1 2
i2 8 1 1
i2 9 3 2
i2 12 1 1.3749831
i3 14 1 0.5
i3 15 1 1
i3 16 1 2
lpreson and lpfreson interpolate between frames but unless kr=sr you'll get
a buzz or noise at kr. I suspect that's your burble. I find the opcodes
very useful but the use of the parameters is a bit obscure.
At 12:23 PM -0800 2/20/99, Sean Costello wrote:
>Hi all:
>
>Just started experimenting with LPC. So far, every sound I have tried
>to produced has ended up sounding overly "liquid" - i.e. the filtering
>seems to change drastically between frames, resulting in a burbling
>sound that resembles a filter being modulated by white noise. Is this
>just the inevitable result of LPC? Or are there some tricks (different
>analysis parameters, different synthesis parameters) that produce
>smoother results?
>
>One other question: Does the Csound lpreson opcode interpolate between
>coefficients for overlapping frames, or does it just spit out the frames
>with no overlapping?
>
>Thanks,
>
>Sean Costello
>
>P.S. If anyone out there knows of any LPC tutorials that explain how to
>use it musically, I would love to seek them out.
-------------------------------------------------------------------------------
Erik Spjut (spyoot, rhymes with cute) - Associate Professor of Engineering
and Associate Director for Engineering Computing, Center for Design Education
Harvey Mudd College, Claremont, CA 91711-5990 USA
Erik_Spjut@hmc.edu Ph & Voice mail (909) 607-3890 Fax (909) 621-8967
|