Byte Array / Bit Conversions in a Cell

I'm not too familiar with Matlab, and I'm having a trouble with cell operations. I have 2 questions:

1- ) Think a byte array cell which consist of 1000x4 uint8 type. I want to convert it into a cell which is 1000x1 uint32 type. Byte order of byte array is big endian. I'm having a trouble with implementing the 'typecast' method to this cell without any loop for time efficiency.

2-) Again think the uint8 type cell which is 1000x1. I need to split it to bits. As a result I want to obtain 8 cells 1000x1 for every bit. But again I could not handle with the 'bitget' method for cells.

Thanks a lot!

Answer 1

[Steven Lord]

If all your numeric data is the same type, I recommend storing it in a numeric array rather than as a cell array. Let's make some sample int8 data:

rng default
x = randi([0 intmax('int8')], 3, 4, 'int8')

x = 3×4

104 116 35 123 115 80 70 20 16 12 122 124

Now to convert this to int32 we need to first make it a vector. I'm going to convert each row of x into an int32 so:

xvec = reshape(x.', 1, [])

xvec = 1×12

104 116 35 123 115 80 70 20 16 12 122 124

y = typecast(xvec, 'int32').'

y = 3×1

2065921128 340152435 2088373264

Let's look at the hex digit patterns.

format hex
disp(x)

   68   74   23   7b
   73   50   46   14
   10   0c   7a   7c

disp(y)

   7b237468
   14465073
   7c7a0c10

Or perhaps you want the first column of x to have the most significant bits in y, the second column the next most significant, etc.? If so take a look at the swapbytes function.

disp(swapbytes(y))

   6874237b
   73504614
   100c7a7c

[Eore D]

Thanks a lot! It is exactly what I want.