You want to use a BinaryReader:
If you have a byte array that you want to deserialize, you wrap it in a memory stream and then use the BinaryReader. Like this:
byte[] inputArray; // somehow you've obtained this
using (var inputStream = new MemoryStream(inputArray))
{
using (var reader = new BinaryReader(inputStream))
{
float f1 = reader.ReadSingle();
short s1 = reader.ReadInt16();
int i1 = reader.ReadInt32();
}
}
You can also do it with the BitConverter class, but you have to maintain state. For example, you can read a float, a short, and an int like this:
byte[] inputArray;
int ix = 0;
float f1 = BitConverter.ToSingle(inputArray, ix);
ix += sizeof(float); // increment to the next value
short s1 = BitConverter.ToInt16(inputArray, ix);
ix += sizeof(short);
int i1 = BitConverter.ToInt32(inputArray, ix);
ix += sizeof(int);
Of the two, I'd suggest using the BinaryReader, because it's more flexible and easier to work with in most cases. BitConverter is handy if you only have a handful of items to deserialize. I suppose it has the potential of being faster, but that wouldn't matter unless your app is highly performance sensitive. And if deserializing data is that critical, you'd probably write a custom deserializer.