Question

I’m working on a project that includes WebSockets, and data between the server (Node.js) and the client (Chrome) is sent using a custom (very simple) format for data exchange I set up.

I’m sending data in pieces of 3 bits because I’m sending items which all have 8 possibilities. The data format looks like this:

            0          1
bit index   01234567 8901...
item        aaabbbcc cddd...

Currently, I’m parsing the items out of the bytes like this:

var itemA = bytes[0] >> 5;
var itemB = (bytes[0] >> 2) & 7;
var itemC = (bytes[0] & 3) << 1 | bytes[1] >> 7;
var itemD = (bytes[1] >> 4) & 7;

Personally, this feels as being too sophisticated. The problem is that it’s only complex because I’m getting data in bytes, which are a multiple of 8. To parse out items of 3 bits I have to bit-shift, doing AND operations, and because 8 is not divisible by 3 I sometimes even have to combine parts of two bytes like for itemC.

It would be much more effective to read this data as groups of 3 bits instead of groups of 8 bits.

What I’ve come up with is converting all bytes into bits to a string using .toString(2), then using .substring to get a substring with length 3, and converting back to a number with parseInt(bitString, 2), but I guess that’s not the way to do it, since string manipulation is slow and I’m actually not doing anything string-related.

Is it possible to read bits in groups of e.g. 3 instead of parsing them from bytes? Or is there a more efficient way to read bits out of bytes?

Answer 1

The binary AND and bit shifting operations are the fastest way of doing this. They translate well into machine code instructions. The only way to further speed this up is by sacrificing bandwidth for speed by for example simply not using more than 3 bits per byte, but judging from your question you’ve probably already considered and rejected that tradeoff.