Export CCS3.3 data and use MATLAB to deal with the method

The CCS3.3 is a DSP integrated development environment (IDE). When doing DSP development, it is the verification algorithm. It is often necessary to use MATLAB for algorithm validation, and the validation algorithm requires data. Thus, one way to interact is to:

1. Using the DSP Development Board to connect to CCS
2. Use "File->data->save." In the CCS3.3 menu. function to export DSP in-memory data into the PC's. dat file
3. Use MATLAB to read. dat files for data processing

Here's how to read a dat file exported by CCS using MATLAB:

% =========================================================================% Read *dat file of CCS3.3% xiahouzuoxin% 2014.04.21% =========================================================================% clc;clear all;close all;%    dialog box select *.dat file [Fname,pname]=uigetfile (... {' *.dat '; '  *. * '}, ' Input *.dat File '); fid = fopen (FullFile (Pname,fname)); Fseek (FID, 21,-1); % go to file header, 21 bytes. For example, 1651 1 80000006 0 100% read Resolution FM = 4;SWITCH (FM) Case 4 in 4Byte format, such as 0x 80000000 x = Textscan (FID, '%2s%8          S ');            Z (:, 1) = Hex2dec (x{2}); Case 2% read in 2Byte format.          such as 0x 8000 0000 x = Textscan (FID, '%2s%4s%4s ');        Z (:, 1) = Hex2dec (x{3});    Z (:, 2) = Hex2dec (x{2}); Case 1% read in 1Byte format.          0x x = Textscan (FID, '%2s%2s%2s%2s%2s ');        Z (:, 1) = Hex2dec (x{5});        Z (:, 2) = Hex2dec (x{4});        Z (:, 3) = Hex2dec (x{3});    Z (:, 4) = Hex2dec (x{2}); otherwise z = [];endif ~isempty (z)% place the data processing code here Endfclose (FID);

In the program, the default DSP memory data storage format is an integer type, and for floating-point DSP such as tms320c6713b, often in memory is a float or double stored in the format,

Then. We also need an operation that translates to floating point numbers. So I wrote a function that was converted to the float type,

function y = integer2float (x, ishex)% ======================================================== =================% Some data y is originally float type. In memory is stored in float format% and now reads it from memory as an integer to X percent This function is used to convert an integer read-out to y% ishex~=0 for input to hexadecimal format% Note: To complete the function, you must understand the format of the IEEE floating-point number xiahouzuoxin% 2014.04.21% =========================================================================if Nargin = = 2 if (ishex) x = Hex2dec (x); Endend[h W] = size (x), y = zeros (h,w), for i = 1:h for j = 1:w sign = bitget (x (i,j), 32); exponent = Bitget (x (i,j), 24:31) * 2.^ (0:7). '; Fraction = Bitget (x (i,j), 1:23) * 2.^ ( -23:-1). '; Y (i,j) = ( -1) ^sign * (1+fraction) * 2^ (exponent-127); EndEnd 

Depending on the input situation (can be 16 binary string representation or decimal number, but must be ≤8bytes), use to give a sample:

>> x=‘4565A012‘>> y=integer2float(x,1);>> x = 23974881923;>> y = integer2float(x)

Later found that Matlab has already had the corresponding method,

y = typecast(uint32(z),‘single‘);

The ability to convert z directly into a single-precision floating-point number. For more information, please see

>> help typecast

Export CCS3.3 data and use MATLAB to deal with the method