Is there a simple way to convert a 32bit number into 8 chunks of 4bit?

[Morthawt]

I am currently manually producing binary 1's and 0's and re-processing into 4bit integers. Is there some easier method without having to manually convert down to the 1's and 0's binary level first?

[BrewManNH]

Huh?

[UEZ]

What about BitShift?

$i32bit = 123456789 ;0x075BCD15 = 0111.0101.1011.1100.1101.0001.0101
                    ;               7    5   11   12   13    1    5
For $i = 0 To 7
    ConsoleWrite(BitAND(0xF, BitShift($i32bit, 4 * $i)) & @CRLF)
Next

Br,

UEZ

[JohnOne]

Huh?

I'd like to see a 4 bit integer too

[Morthawt]

Like, take this 32 bit number: 55028615

The binary representation of it is: 00000011010001111010101110000111

Then you break it down into groupings of 4:

0000

0011

0100

0111

1010

1011

1000

0111

 

Then break each 4 bit block into a 4 bit number:

0

3

4

7

10

11

8

7

 

Is there some easier method than manually converting down and processing individual bits?

[Morthawt]

What about BitShift?

$i32bit = 123456789 ;0x075BCD15 = 0111.0101.1011.1100.1101.0001.0101
                    ;               7    5   11   12   13    1    5
For $i = 0 To 7
    ConsoleWrite(BitAND(0xF, BitShift($i32bit, 4 * $i)) & @CRLF)
Next

Br,

UEZ

 

That is looking very close to what I want. I just need to process and understand what it is doing. I think that would work, in reverse ordering.

[UEZ]

For $i = 7 To 0 Step - 1

Br,

UEZ

[Morthawt]

For $i = 7 To 0 Step - 1

Br,

UEZ

yes I got this done already. I am just trying to decipher what the code you wrote is doing. I am not too familiar with bitwise functions. I know bitand searches to see if a particular bit is flipped or not in a given number I think. But I am quite confused right now until I understand it. If you can provide an English explanation of what the functions are doing that would help me a lot most likely. Thanks. 

[UEZ]

The 1. step is to shift the bits of the 32 bit number 4 * $i bits to the right. This is done by the BitShift($i32bit, 4 * $i) command.

To mask the result only to the last 4 bits (2^4-1 = 15) I used the BitAND() command.

BitAND(0xF, the result of the bitshift) = the result.

Ok?

Br,

UEZ

Edited January 17, 2015 by UEZ

[Morthawt]

Erm... not really lol. I do not visually understand what it is doing. Plus I think BitAnd is the thing I use to check if a number has a particular binary bit flipped like finding out what options have been set on a combined number like option 1, option 2, option 4, option 8 etc. Then I can run a check to see if option 8 is part of that number. But in this situation I have no clue what bitand is doing or how the bits get shifted by the bitshift. I rarely ever deal with anything other than bitand, and in this situation my usage does not seem to fit with your usage.

Edited January 17, 2015 by Morthawt

[Morthawt]

I think I understand the shift, it moves to the right so each time it cycles through the last 4 digits are discarded and the 4 prior are now at the end I think. But it seems like BitAnd may have two functions. I use it to determine if a paritcular bit was flipped to create the big number presented. If the number is not 0 then it was part of creating it. But it seems adding a hex 0xF is acting as a mask? Only giving a single first 4 bit grouping? Is this correct?

[UEZ]

Here a more detailed description:

1.  BitShift(123456789, 4 * 0) -> no shifting -> $result = 123456789 
    BitAND(0xF, 123456789) = 5
        0111.0101.1011.1100.1101.0001.0101
        0000.0000.0000.0000.0000.0000.1111
    =   0000.0000.0000.0000.0000.0000.0101 -> = 5

2.  BitShift(123456789, 4 * 1) -> shift the bits 4 times to the right -> 0000.0111.0101.1011.1100.1101.0001 = 7716049
    BitAND(0xF, 7716049) = 1
        0000.0111.0101.1011.1100.1101.0001
        0000.0000.0000.0000.0000.0000.1111
    =   0000.0000.0000.0000.0000.0000.0001 -> = 1
    
3.  BitShift(123456789, 4 * 2) -> shift the bits 8 times to the right -> 0000.0000.0111.0101.1011.1100.1101 = 482253
    BitAND(0xF, 482253) = 13
        0000.0000.0111.0101.1011.1100.1101
        0000.0000.0000.0000.0000.0000.1111
    =   0000.0000.0000.0000.0000.0000.1101 -> = 13
...


Eg: 1101 = 1*2^3 + 1*2^2 + 0*2^1 + 1*2^0 = 8 + 4 + 0 + 1 = 13

I hope it helps you to understand now.

 

 

And operation on bit level:

 

0 and 0 = 0

0 and 1 = 0

1 and 0 = 0

1 and 1 = 1

 

Br,

UEZ

Edited January 17, 2015 by UEZ

[Morthawt]

Here a more detailed description:

1.  BitShift(123456789, 4 * 0) -> no shifting -> $result = 123456789 
    BitAND(0xF, 123456789) = 5
        0111.0101.1011.1100.1101.0001.0101
        0000.0000.0000.0000.0000.0000.FFFF
    =   0000.0000.0000.0000.0000.0000.0101 -> = 5

2.  BitShift(123456789, 4 * 1) -> shift the bits 4 times to the right -> 0000.0111.0101.1011.1100.1101.0001 = 7716049
    BitAND(0xF, 7716049) = 1
        0000.0111.0101.1011.1100.1101.0001
        0000.0000.0000.0000.0000.0000.FFFF
    =   0000.0000.0000.0000.0000.0000.0001 -> = 1
    
3.  BitShift(123456789, 4 * 2) -> shift the bits 8 times to the right -> 0000.0000.0111.0101.1011.1100.1101 = 482253
    BitAND(0xF, 482253) = 1
        0000.0000.0111.0101.1011.1100.1101
        0000.0000.0000.0000.0000.0000.FFFF
    =   0000.0000.0000.0000.0000.0000.1101 -> = 13
...


Eg: 1101 = 1*2^3 + 1*2^2 + 0*2^1 + 1*2^0

I hope it helps you to understand now.

 

And operation on bit level:

0 and 0 = 0

0 and 1 = 0

1 and 0 = 0

1 and 1 = 1

Br,

UEZ

I think so. So I was right, shift is lopping bits off the end and shifting the remaining over to the right (or left if you use negative number) and I think after some experimentation and looking at your depiction that I have a better understanding of BitAnd. I think BitAnd uses the first parameter to return all the values that are in the second parameter by checking only the bits used that are represented in the first parameter. So if on the first parameter I put in 3 and a number in the right it will only return the value based on the 1 and 2 bits, so it could produce a 0, 1, 2 or 3 since those bits can only produce those numbers. So if you use 0xF for the first or even 15, it returns the value of the second number but only up to the value produced by the status of the bits that can make up that first number, in this instance being 15 maximum.

Does that sound about right?

Edited January 17, 2015 by Morthawt

[JohnOne]

Excuse me for butting in UEZ, that seems just like a logic And gate in electronics, is that correct?

Edited January 17, 2015 by JohnOne

[JohnOne]

In an Or gate it would be.

0 or 0 = 0

0 or 1 = 1

1 or 0 = 1

1 or 1 = 1

[UEZ]

Sorry, I made a mistake by adding the FFFF. That makes no sense here! It must be 1111 (=15 or 0xF) of course!

But your understanding seems to be correct.

BitAnd masks the bits only. In this case you just cut off the bits from 32 to 4 because the 1st 4 bits are what you are looking for.

@J1: correct it is the same as the logic AND gate in electronics

Br,

UEZ

Edited January 17, 2015 by UEZ

[JohnOne]

Cheers, my virtually non existent understanding of their use just increased tenfold.

[Morthawt]

Thank you UEZ. I love being able to improve my code with just these little elegant pieces of functionality. Once you truly understand the what and why, it is so much more useful. The help file does not really explain what the bitwise functions do very well. It assumes you know what they already are and just shows examples of them being used. I think better documentation would help. I might not have needed to make a post on the forum if I understood them from the beginning. I immediately went to the bitwise section but it was all greek to me. So I did a whole thing converting to bit binary and grouping things up etc :S very CPU time consuming compared to your method. So, thanks again for help me to understand it.

[UEZ]

You are welcome guys!

PS: Bitshift to the right is also a division. In this case 4 bits to the rights means divison with 2^4

Br,

UEZ

Edited January 17, 2015 by UEZ

[JohnOne]

You're not alone, I've read up about the bitwise operations many times, and just never got them at all.