tvconv

A time-varying convolution (FIR filter) opcode.

An opcode that takes two incoming signals and interprets one of them as the coefficients of linear time-variable finite impulse response filter. This is implemented via direct convolution (for partition sizes of 1 sample) or DFT-based partitioned convolution. The signals can be 'frozen' (i.e. the filter coefficients are kept the same) at any point in time, at a-rate or k-rate.

Syntax

ares tvconv asig1, asig2, xfreez1, xfreez2, iparts, ifils

Initialization

iparts -- partition size, for sizes > 1, a DFT-based partitioned convolution process is used. Otherwise a time-domain delay line FIR is implemented. Partition sizes > 1 are rounded to the nearest power-of-two.

ifils -- filter size. For partition sizes > 1, filter sizes are rounded to the nearest power-of-two. With partition size = 1, since direct convolution is used, filters can be of any size.

Performance

ares -- audio output.

asig1, asig2 -- audio inputs.

xfreez1 -- freeze switch for asig1. Coefficients are only updated (ie. the signal is passing into the convolution) if xfreez1 > 0. This input can take an audio or a k-rate signal, or a constant.

xfreez2 -- freeze switch for asig2, similar to xfreez1 in operation.

Examples

Here is an example of the tvconv opcode. It uses the file tvconv.csd.

<CsoundSynthesizer>
<CsOptions>
-odac 
</CsOptions>
<CsInstruments>
sr = 44100
ksmps = 32
nchnls = 2
0dbfs = 1

instr 1

 asig diskin "fox.wav",1,0,1
 air diskin "drumsMlp.wav",1,0,1
 air buthp air/0dbfs,1000
 k1 linseg 0,p3/3,0,0,1,2*p3/3,1
 a1 oscili k1, 0.5, 1
 a2 oscili k1, 0.6, 1
 asig tvconv asig,air,1-a2,1-a1,256,1024
 asig clip asig,1,0dbfs
 outs asig, asig

endin

</CsInstruments>
<CsScore>
f1 0 1024 7 0 512 0 1 1 511 1
i1 0 30
e
</CsScore>
</CsoundSynthesizer>

See also

Convolution and Morphing

Credits

Author: Victor Lazzarini
2017

New in version 6.09