Converting 1s and 0s to float

Question

So I have a separate class and function that take a float and convert it to a string of 1's and 0's. That then gets passed to another object where I would like to convert the string of 1's and 0's to back to the original float. Here is the code I been playing with, let me know if you have any suggestions.

 void TestFunc(string s, int startpos, int stoppos)
     {
         byte[] b = new byte[4];
         int bytepos = 0;
         for (int i = 0; i < 32; i += 8)
         {
             print(i);
             string a = "";
             for (int j = 0; j < 8; j++)
             {
                 a += s[i + startpos];
             }
             print(a);
             b[bytepos] = Convert.ToByte(a);
             bytepos += 1;
         }
     }

Answer 1

Uhm first of all your code doesn't look like it's using a "float" anywhere. You pass in a string and only work on single characters. Your "j" loop does simply add the same character 8 times. It also assumes the incoming string has at least 32 characters. Your "a" string would always contain 8 characters. Convert.ToByte expects a string that contains a decimal representation of a byte, i.e. 0 up to 255. You might wanted to use Convert.ToByte ovreload that takes an additional base where you would pass 2 for binary. However which order your "bits" are in depends on how you actually converted your value to a binary string. Your method just has way too many issues and mistakes.

Why do you actually want to convert it to a string containing binary digits? it's an incredible waste of space ^^.

Anyways I would recommend to first get the actual binary representation of the float value. You can use a helper struct like this one:

 [System.Runtime.InteropServices.StructLayout(System.Runtime.InteropServices.LayoutKind.Explicit)]
 public struct FloatHelper
 {
     [System.Runtime.InteropServices.FieldOffset(0)]
     public float floatVal;
     [System.Runtime.InteropServices.FieldOffset(0)]
     public int intVal;
 }

This struct uses a trick to "overlap" the "floatVal" and the "intVal" field in memory so they both occupy the same memory location. That way we can write a float value to floatVal and simply read out the binary representation using the intVal or vice versa. You can simply create two methods like this:

 public static string FloatToBinary(float aValue)
 {
     FloatHelper h = new FloatHelper();
     h.floatVal = aValue;
     var v = h.intVal;
     var sb = new System.Text.StringBuilder(32);
     for (int i = 0; i < 32; i++)
         sb.Append(((v&(1<<(31-i)))!=0)?'1':'0');
     return sb.ToString();
 }

 public static float BinaryToFloat(string aBinaryStr)
 {
     if (string.IsNullOrEmpty(aBinaryStr) || aBinaryStr.Length < 32)
         throw new System.FormatException("BinaryToFloat expects a binary string of 32 characters");
     int v = 0;
     for (int i = 0; i < 32; i++)
         v |= ((aBinaryStr[i] == '1') ? 1 : 0) << (31-i);
     FloatHelper h = new FloatHelper();
     h.intVal = v;
     return h.floatVal;
 }

This is untested but BinaryToFloat should recreate the exact same float value from the binary string that is produced by FloatToBinary.

I actually used a similar struct in my floating point number format editor window which might help to understand how floating point numbers work behind the scenes. There's also this online tool which allows you to convert 32bit floats to binary and inspect / manipulate the individual bits.

Converting 1s and 0s to float

1 Reply

Your answer

Follow this Question

Related Questions