Floating point to 16 bit hexadecimal

Hello all, how can i convert floating point decimal to 16 bit hexadecimal value?

Thank you

[Guillaume]

Specifying the encoding used would be helpful. Is it IEEE 754 half precision you're after?

[Tousif Ahmed]

yes IEEE 754 half precision

Answer 1

[Ameer Hamza]

You can use this file exchange submission. It will convert your number to 16-bit half precision format. But instead of returning as a string, it will return the equivalent uint16 MATLAB object. To get binary string, you can do this,

halfPrecisionNumber = halfprecision(1.2);  % halfprecision() is provided in the file exchange package.
halfPrecisionString = dec2bin(halfPrecisionNumber, 16);

In case you face the following error on running the package

"Floating point bit pattern is not IEEE 754"

you can try following

• Open halfprecision.c file.
• Add following to include section

#include <stdint.h>

• Replace 4 #define with the following macros

#define INT16_TYPE int16_t 
#define UINT16_TYPE uint16_t 
#define INT32_TYPE int32_t 
#define UINT32_TYPE uint32_t

• Run mex halfprecision.c.

[Guillaume]

halfPrecisionString = dec2bin(halfPrecisionNumber, 16)

will force dec2bin to output 16 bits regardless of the magnitude of the number and thus avoids having to pad afterwards.

[Ameer Hamza]

Thanks. I have modified the code in my Answer.

[Tousif Ahmed]

it is showing the floating point value in binary, when i am using num2hex for the answer to convert it into hexadecimal, its showing

halfPrecisionString =

0011110000000000

>> num2hex(halfPrecisionString); Error using num2hex (line 32) Inputs must be floating point, and may not be of class char.

[Guillaume]

Use dec2hex instead of dec2bin in Ameer's answer. Do not use num2hex which is not all suited for what you want.

halfPrecisionNumber = halfprecision(1.2);  % halfprecision() is provided in the file exchange package.
halfPrecisionHex = dec2hex(halfPrecisionNumber, 4);

[Tousif Ahmed]

cool that works perfectly. Thanks to Mr. Guillaume and Mr. Ameer Hamza