Programmer Guide/Command Reference/NUM

command	return value	value of RC
NUM expression	value of expression or empty string if the evaluation fails	0 error code
NUMCHECK expression	value of expression or empty string if the evaluation fails	0 warning code

The expression will be evaluated numerically, and the result (textual representation) will be returned. The expression may consist of the following parts:

decimal numbers

123.456, 17.5e3, 100, -312.123, 1e-3

hexa-decimal numbers

0x1234, 0xabc, 0XabC

special constants

pi (=3.1415...), e (=2.71828...),

rand (a linear distributed pseudo random number r; -1≤r<1)

lran (a linear distributed pseudo random number r; 0≤r<1)

numerical operators

-a (negate), a+b (add), a-b (subtract), a*b (multiply), a/b (divide), a%b (modulo), a^b (power)

logical operators (bitwise)

!a (not), a&b (and), a|b (or), a^b (exclusive or)

brackets

(expression)

functions

setlran(seed)	linear distributed pseudo random number r (0≤r<1) seed is the start value (0≤seed&lt<1)
sin(x), cos(x), tan(x)	sine, cosine or tangent of x
asin(x), acos(x), atan(x)	inverse sine, cosine or tangent of x
exp(x)	computes ex
log(x)	common logarithm of x (base 10)
ln(x)	natural logarithm of x (base e)
sqrt(x)	square root of x (base 10)
abs(x)	absolute value of x (base 10)
int(x)	integer part of x, the fractional part is truncated (range of x: 32bit signed integer)
round(x)	round to the nearest integer of x (range of x: 32bit signed integer)
db(x)	convert level to factor (10x/20)
sinc(x), sinx(x)	sinc function: sin(x) / x
sign(x)	sign of x; returns -1 if x<0, otherwise 1
hz2bark(x)	convert x from Hertz to Bark
bark2hz(x)	convert x from Bark to Hertz
bit(n)	returns the integer with only bit n (0≤n<32) is set and all other are cleared; can be used to generate bit masks
floor(x)	returns a floating-point value representing the largest integer that is less than or equal to x (range of x: 64bit float)
iserr(rc), iserror(rc)	returns 1 if rc is a STx error code, otherwise 0 this function can be used to test the completion code RC of a command
iswarn(rc), iswarning(rc)	returns 1 if rc is a STx warning code, otherwise 0 this function can be used to test the completion code RC of a command
npow2(n)	next power of 2; returns the smallest possible value 2m ≥ n

Notes

• In case of the expression being syntactically ill-formed, an error (NUM) or warning (NUMCHECK) is reported.
• The special numerical objects (vectors, matrices) available in EVAL-expressions, can not be used in the NUM expressions. Only the use of numerical constants and variables is possible.
• The same expression syntax is used for the INT and INTCHECK, for the numerical expressions in conditions (e.g. IF or WHILE) and for the evaluation of numerical arguments of commands.
• Numerical expressions are evaluated in 64bit floating point precission.
• For logical (operators !, &, |, ^) and integer expressions (functions round, int, bit) 32bit signed integers are used.

See also

INT, INTCHECK, EVAL, EVALCHECK, SEGMENT

Examples

// #wave = wave item
#t := num $#wave[!length] / $#wave[!srate]          // duration in seconds
// #t = frame length in seconds, #o = overlap in percent
#lfrm := int $#t * $#wave[!srate]                   // frame length in samples
#lhop := int $#t * (1 - $#o / 100) * $#wave[!srate] // hopsize in samples
#lfft := int npow2($#lfrm)                          // fft length
#df   := num $#wave[!srate] / $#lfft                // frequency resolution

// using expressions in a IF statement
if 'db($#level)' > 1 then
    conlog 'level must be lower or equal 0dB'
    #level := 0
end