Ok, since you acknowledge this is a homework question and are asking for a method not a solution, first, you need the signal values
x = Recorder.getaudiodata('int16');
Since you are going to be quantizing you might as well start with values in the right range, as opposed to doubles which would be normalized [-1, 1]
So, next you need to quantize from 16 bits to 15. This is actually pretty easy in matlab, if you're writing a ton of code, it's probably not working. There are several ways you could do it.
Essentially, you're throwing away half the information when you remove 1 bit.
Take the numbers 0, 1, 2, 3 which have 2 bit precision. In binary, 00, 01, 10, 11. If I quantize to one bit, I throw away half the information. Imagine the LSB is gone, I'd have 0, 0, 1, 1. But, my original numbers span 0 to 3... Typically you want your numbers to have the same dynamic range. So, I threw half the numbers away, what if I scale the new numbers by 2? I get 0, 0, 2, 2. This is a correct quantization from 2 to 1 bit.
In your problem, what would happen if all the odd numbers were even?