opcode — Defines the start of user-defined opcode block.

Defining opcodes

The opcode and endop statements allow defining a new opcode that can be used the same way as any of the built-in Csound opcodes. These opcode blocks are very similar to instruments (and are, in fact, implemented as special instruments), but cannot be called as a normal instrument e.g. with the i statements.

A user-defined opcode block must precede the instrument (or other opcode) from which it is used. But it is possible to call the opcode from itself. This allows recursion of any depth that is limited only by available memory. Additionally, there is an experimental feature that allows running the opcode definition at a higher control rate than the kr value specified in the orchestra header.

Similarly to instruments, the variables and labels of a user-defined opcode block are local and cannot be accessed from the caller instrument (and the opcode cannot access variables of the caller, either).

Some parameters are automatically copied at initialization, however:

  • all p-fields up to and including the highest-numbered one referenced in the calling instrument

  • extra time (see also xtratim, linsegr, and related opcodes). This may affect the operation of linsegr/expsegr/linenr/envlpxr in the user-defined opcode block.

  • MIDI parameters, if there are any.

Also, the release flag (see the release opcode) is copied at performance time.

Modifying the note duration in the opcode definition by assigning to p3, or using ihold, turnoff, xtratim, linsegr, or similar opcodes will also affect the caller instrument. Changes to MIDI controllers (for example with ctrlinit) will also apply to the instrument from which the opcode was called.

Use the setksmps opcode to set the local ksmps value.

The xin and xout opcodes copy variables to and from the opcode definition, allowing communication with the calling instrument.

The types of input and output variables are defined by the parameters intypes and outtypes.

[Tip] Tip

You can create UDOs which take no inputs or outputs by using 0 instead of a string.

[Note] Notes

  • xin and xout should be called only once, and xin should precede xout, otherwise an init error and deactivation of the current instrument may occur.

  • These opcodes actually run only at i-time. Performance time copying is done by the user opcode call. This means that skipping xin or xout with kgoto has no effect, while skipping with igoto affects both init and performance time operation.


opcode name, outtypes, intypes


name -- name of the opcode. It may consist of any combination of letters, digits, and underscore but should not begin with a digit. If an opcode with the specified name already exists, it is redefined (a warning is printed in such cases). Some reserved words (like instr and endin) cannot be redefined.

intypes -- list of input types, any combination of the characters: a, k, O, P, V, K, i, o, p, and j. A single 0 character can be used if there are no input arguments. Double quotes and delimiter characters (e.g. comma) are not needed.

The meaning of the various intypes is shown in the following table:

Type Description Variable Types Allowed Updated At
a a-rate variable a-rate a-rate
i i-rate variable i-rate i-time
j optional i-time, defaults to -1 i-rate, constant i-time
k k-rate variable k- and i-rate, constant k-rate
O optional k-rate variable, defaults to 0 k- and i-rate, constant k-rate
P optional k-rate variable, defaults to 1 k- and i-rate, constant k-rate
V optional k-rate variable, defaults to 0.5 k- and i-rate, constant k-rate
J optional k-rate variable, defaults to -1 k- and i-rate, constant k-rate
K k-rate with initialization k- and i-rate, constant i-time and k-rate
o optional i-time, defaults to 0 i-rate, constant i-time
p optional i-time, defaults to 1 i-rate, constant i-time
S string variable i-rate string i-time

The maximum allowed number of input arguments is 256.

outtypes -- list of output types. The format is the same as in the case of intypes.

Here are the available outtypes:

Type Description Variable Types Allowed Updated At
a a-rate variable a-rate a-rate
i i-rate variable i-rate i-time
k k-rate variable k-rate k-rate
K k-rate with initialization k-rate i-time and k-rate

The maximum allowed number of output arguments is 256.

iksmps (optional, default=0) -- sets the local ksmps value. Must be a positive integer, and also the ksmps of the calling instrument or opcode must be an integer multiple of this value. For example, if ksmps is 10 in the instrument from which the opcode was called, the allowed values for iksmps are 1, 2, 5, and 10.

If iksmps is set to zero, the ksmps of the caller instrument or opcode is used (this is the default behavior).

[Note] Note

The local ksmps is implemented by splitting up a control period into smaller sub-kperiods and temporarily modifying internal Csound global variables. This also requires converting the rate of k-rate input and output arguments (input variables receive the same value in all sub-kperiods, while outputs are written only in the last one).

[Warning] Warning about local ksmps

When the local ksmps is not the same as the orchestra level ksmps value (as specified in the orchestra header), global a-rate operations must not be used in the user-defined opcode block.

These include:

  • any access to ga variables

  • a-rate zak opcodes (zar, zaw, etc.)

  • tablera and tablewa (these two opcodes may in fact work, but caution is needed)

  • The in and out opcode family (these read from, and write to global a-rate buffers)

In general, the local ksmps should be used with care as it is an experimental feature, although it works correctly in most cases.

The setksmps statement can be used to set the local ksmps value of the user-defined opcode block. It has one i-time parameter specifying the new ksmps value (which is left unchanged if zero is used, see also the notes about iksmps above). setksmps should be used before any other opcodes (but allowed after xin), otherwise unpredictable results may occur.

The input parameters can be read with xin, and the output is written by xout opcode. Only one instance of these units should be used, as xout overwrites and does not accumulate the output. The number and type of arguments for xin and xout must be the same as in the declaration of the user-defined opcode block (see tables above).

The input and output arguments must agree with the definition both in number (except if the optional i-time input is used) and type. An optional i-time input parameter (iksmps) is automatically added to the intypes list, and (similarly to setksmps) sets the local ksmps value.


The syntax of a user-defined opcode block is as follows:

opcode  name, outtypes, intypes
xinarg1 [, xinarg2] [, xinarg3] ... [xinargN]  xin
[setksmps  iksmps]
... the rest of the instrument's code.
xout  xoutarg1 [, xoutarg2] [, xoutarg3] ... [xoutargN]

The new opcode can then be used with the usual syntax:

  [xoutarg1] [, xoutarg2] ... [xoutargN] name  [xinarg1] [, xinarg2] ... [xinargN] [, iksmps]

[Note] Note

The opcode call is always executed both at initialization and performance time, even if there are no a- or k-rate arguments. If there are many user opcode calls that are known to have no effect at performance time in an instrument, then it may save some CPU time to jump over groups of such opcodes with kgoto.


Here is an example of a user-defined opcode. It uses the file opcode.csd.

Example 652. Example of a user-defined opcode.

See the sections Real-time Audio and Command Line Flags for more information on using command line flags.

; Select audio/midi flags here according to platform
-odac    ;;;realtime audio out
;-iadc    ;;;uncomment -iadc if realtime audio input is needed too
; For Non-realtime ouput leave only the line below:
; -o opcode_example.wav -W ;;; for file output any platform

sr = 44100
ksmps = 32
nchnls = 2
0dbfs  = 1

/* example opcode 1: simple oscillator */

        opcode Oscillator, a, kk

kamp, kcps      xin             ; read input parameters
a1      vco2 kamp, kcps         ; sawtooth oscillator
        xout a1                 ; write output


/* example opcode 2: lowpass filter with local ksmps */

        opcode Lowpass, a, akk

        setksmps 1              ; need sr=kr
ain, ka1, ka2   xin             ; read input parameters
aout    init 0                  ; initialize output
aout    =  ain*ka1 + aout*ka2   ; simple tone-like filter
        xout aout               ; write output


/* example opcode 3: recursive call */

        opcode RecursiveLowpass, a, akkpp

ain, ka1, ka2, idep, icnt       xin     ; read input parameters
        if (icnt >= idep) goto skip1    ; check if max depth reached
ain     RecursiveLowpass ain, ka1, ka2, idep, icnt + 1
aout    Lowpass ain, ka1, ka2           ; call filter
        xout aout                       ; write output


/* example opcode 4: de-click envelope */

        opcode DeClick, a, a

ain     xin
aenv    linseg 0, 0.02, 1, p3 - 0.05, 1, 0.02, 0, 0.01, 0
        xout ain * aenv         ; apply envelope and write output


/* instr 1 uses the example opcodes */

        instr 1

kamp    =  .7                ; amplitude
kcps    expon 50, p3, 500       ; pitch
a1      Oscillator kamp, kcps                   ; call oscillator
kflt    linseg 0.4, 1.5, 0.4, 1, 0.8, 1.5, 0.8  ; filter envelope
a1      RecursiveLowpass a1, kflt, 1 - kflt, 10 ; 10th order lowpass
a1      DeClick a1
        outs a1, a1



i 1 0 4
e5		;extra second before quitting


See Also

endop, setksmps, xin, xout


Author: Istvan Varga, 2002; based on code by Matt J. Ingalls

New in version 4.22