Hex to float like python struct.unpack()

Hi

I'm trying to convert a hex string to a float in the same manner python function struct.unpack(), https://docs.python.org/3/library/struct.html , which is using the IEEE754 binary32 0r 64 depending on foat or double.

As an example I have

hex = 'C956F53D'

When i run it in a struct.unpack() i get the value:

data = 0.1198

And this is correct. However, when i try the following matlab function i get something totally different.

>> hex2dec(hex)

ans =

   3.3779e+09
   
>> typecast(uint32(hex2dec(hex)),'single')

ans =

  single

 -8.8047e+05

Does anybody know the difference or what I'm doing wrong?

Thanks!

Answer 1

[Ameer Hamza]

This is the correct way

hex = 'C956F53D';

x = hex2dec(reshape(hex, 2, []).');
dec = typecast(uint8(x), 'single')

[Sebastian Thuné]

You my friend are a superstar! Thanks!

So i basically have to put the bytes in an array first. Why is that? Is it because typecast has to go byte by byte?

[Ameer Hamza]

Yes, you need to convert it into an uint8 array. Then MATLAB reads the bytes from the input array and converts those bytes into 'single'. It is pretty similar to pointer typecasting in C.

I am glad to be of help!

Answer 2

[James Tursa]

Could be a Big Endian vs Little Endian thing. E.g., inserting a swapbytes( ) step:

>> hex = 'C956F53D'
hex =
    'C956F53D'
>> typecast(swapbytes(uint32(hex2dec(hex))),'single')
ans =
  single
    0.1198