Bitwise OR and Bitwise shift

I am trying to write a matlab code pretty much like the following C code :

uint8_t data[6];

uint32_t tdata = data[3] & 0x0F;

tdata <<= 8;

tdata |= data[4];

tdata <<= 8;

tdata |= data[5];

temperature = ((float)tdata * 200 / 0x100000) - 50;

I recently started using MATLAB and how I translated the above code is something like this :

      uint32(tdata) = data(3) & 0x0F; 
      uint32(tdata) = bitshift(tdata,8);
      uint32(tdata) = bitor(tdata,data(4));
      uint32(tdata) = bitshift(tdata,8);
      uint32(tdata) = bitor(tdata,data(5));
      temperature = (tdata*200/1048576)-50;

It's throwing a lot of errors and I am quite stuck. Please help me out here. Thank you in advance.

[Matt J]

We cannot run your code without being given "data" and "tdata".

[Das Siddharth]

Actually, these are the datas I am getting from a sensor. It's in real-time. That's why I didn't post it here. I just wanted to know if we translate that code what's the correct way.

Answer 1

[Walter Roberson]

temperature = double(bitand(data(4), 0x0f) * 65536 + data(5) * 256 + data(6)) * 200 / double(0x100000) - 50;

[Das Siddharth]

Wow ! Can we do this ? I mean without any bitshifts ? These are datas from an AHT20 temperature sensor btw.

[Walter Roberson]

Use. You are using non-overlapping fields. Shift by multiplying, and add the components.

There are other ways of doing the calculation, but you run into hassles about the fact that MATLAB is little-endian on all current platforms, so doing bit operations is not the same as doing mathematical operations.

For example,

typecast( [data(6), data(5), bitand(data(4), 0x0f), uint8(0)], 'uint32')  %slam together little-endian style