How to translate 32-bit number to float number

The acquired data is expressed as '6D61803F', this may be a little-endian. I translated it to '3F80616D', then I got the float number 1.002973, this was the real data!

Now I used

q = quantizer(q);
hex2num(q, '6D61803F') or hex2num(q, '3F80616D')

Both of those did not get the real one!

What's the problem?

Thanks

[Walter Roberson]

Please show the struct q that you are passing to quantize() to create the quantizer object.

[wenlong]

 q =

        DataMode = fixed
       RoundMode = floor
    OverflowMode = saturate
          Format = [16  15]

Sorry about q = quantizer(q); it should be q = quantizer

[wenlong]

Thanks Walter!

Answer 1

[Walter Roberson]

>> q = quantizer; q.quantize(1.002973)
Warning: 1 overflow(s) occurred in the fi quantize operation. 
> In embedded.quantizer/quantize (line 77) 
ans =
         0.999969482421875

You are exceeding the range of the quantizer.

>> q.numeric2hex(1.002973)
ans =
7fff

You are providing more hex digits than fit into the quantizer.

A [16 15] quantizer is a total of 16 bits long.

I would suggest that you need

q = quantizer('mode', 'single');
hex2num(q, '3F80616D')

 ans =
          1.00297319889069