Question

I am writing some signal processing code in C that has a communications channel. At the output I get a bunch of bits as they arrive.

for (n=0; n<BUFFER_LENGTH; n++) {
    /* do some processing that calculates x */
    output[n] = x > 0;
}

Here are my questions:

1. Is there a good type to represent
   the output array? At first I thought
   uint1_t would be ideal but I hear
   that doesn’t necessarily represent
   one bit in memory.
2. Once I find a
   sync pattern in the data I know the
   format of the next bits, how can I
   convert a bunch of 1’s and 0’s in
   the array into integers, floats,
   doubles, characters, etc.? I’ve heard of using
   a union but I don’t think that will work with an
   array of bits.

Answer 1

Just store the data in a sufficiently large block of bytes and then iterate through the bits using shifting and masking to extract individual bits sequentially.

e.g. to print out the contents of a buffer as individual bits:

#include <stdio.h>
#include <stdint.h> // uint8_t et al
#include <limits.h> // CHAR_BIT

uint8_t buffer[256];
int b, bit;

for (b = 0; b < 256; ++b)
{
    for (bit = CHAR_BIT - 1; bit >= 0; --bit)
    {
        uint8_t mask = 1 << bit;
        printf("%2d", (buffer[b] & mask) != 0);
    }
    printf("\n");
}