When I am experimenting in converting matlab code into vhdl with hdl coder, I am getting the error such as Function 'log' is not defined for values of class 'embedded.fi' or Function 'exp' is not defined for values of class 'embedded.fi'.
The specific block of code comes from the block:
f_0 = fi( 0, 0, 1, 0, fm ); % 133.33333; f_sp = fi( 200*fi( 1/3, 0, 14, 15 ), 0, 14, 7, fm ); % 66.66667; brkfrq = fi( 1000, 0, 11, 0, fm ); brkpt = fi( fi_div( (brkfrq - f_0), f_sp ), 0, 14, 10, fm ); logstep = fi( 1.0711703, 0, 14, 13, fm ); linpts = (f<brkfrq); z = fi( fi( 0, 0, 1, 0 )*f, 0, 6, 0, fm ); z( linpts ) = fi( fi_div( (f( linpts ) - f_0), f_sp ), 0, 6, 0, fm ); % error with line below. z( ~(linpts~=0) ) = fi( brkpt + (log( fi_div( f( ~(linpts~=0) ), brkfrq ) )) ./ log( logstep ), 0, 6, 0, fm );
The original matlab code is from audspec.m of rastamat. I have tried converting fixed point to double before calling log(), but then another error pops up.
As the error message states, there is no fixed-point version of the 'log' or 'exp' commands in MATLAB HDL Coder. The complete list of functions that have fixed point versions and are supported for HDL code generation is at http://www.mathworks.com/help/hdlcoder/ug/fixed-point-run-time-library-support.html.