View Full Version : IEEE Std 754 for Float32

Earl Foster
04-02-2009, 11:27 PM
I have a 2 part question:

I have a device that will be sending data that includes Float32 information, such as heading, and I am trying to figure out how I would go about parsing that information into it component parts. I know where in the data stream the information will start and stop but I am not sure I have the correct understanding of parsing it properly.

Part 1: Understanding the IEEE Std 754 format for single precision

Let’s say the decimal number -10.6 is coming in and is in big endian format. The binary/hex equivalent in IEEE Std 754 for the float32 would look like this (broken down into octets for readability)

Binary: 11000000 00000000 00000000 01101010 Hex: C0 00 00 6A

Bit 31: sign (1 = negative)
Bits 30-23: Exponent (0 < Exponent < 255)
Bits 22 – 0: Number without decimal point, ex 106

Part 2:

I am assuming I can assign C000006A to a variable then use the Float32 object to perform calculations, however, the Prop uses LITTLE ENDIAN so will I need to reverse my information and then assign to the variable?

Change C000006A to 6A0000C0

Would appreciate someone setting me straight if I am on the wrong path.

WWW.HAPB.NET (http://www.HAPB.NET)

"Don't ask yourself what the world needs - ask yourself what makes you come alive, and then go do it." - H.T.Whitman

Mike Green
04-02-2009, 11:39 PM
It all depends on how the data is actually sent. Assuming that the data is being sent 8 bits at a time with the most significant byte sent first, then you will have to assemble it into a 32 bit long word in the proper order. The most significant byte will be stored in the highest address byte of a long and the least significant byte will be stored in the lowest address byte of the long. If the bytes are stored in a buffer, then you will have to reverse their order to make a single 32 bit long word out of them.

In your example, the 32 bit value is written $C000006A. It's stored in memory as the following bytes: $6A, $00, $00, $C0 from lowest to highest address. If you're receiving them a byte at a time, you'll have to store them in reverse order.

Earl Foster
04-02-2009, 11:53 PM
Then Big Endian and Little Endian does not make any difference in this senerio as long as I know the correct incoming format. However, if the data stream was transmitted little endian (6A0000C0) then I would NOT need to reverse the bytes (C000006A) before performing any calculations. The device I am using allows me to change from big endian to little endian, thats why I am asking.

Thanks Mike

WWW.HAPB.NET (http://www.HAPB.NET)

"Don't ask yourself what the world needs - ask yourself what makes you come alive, and then go do it." - H.T.Whitman

Post Edited (Earl Foster) : 4/2/2009 6:00:53 PM GMT