Ah. I had to try that. So I made a little patch to test it. It has an osc/sine which outputs to the left channel. A copy is sent throuch bipolar2unipolar to the right channel.
My first Axoloti patch ever!
Then I hooked it up to my pocket scope:
Sorry for the bad photo, but you should be able to see that
- the red channel is centered about 0V and has a large peak-to-peak range. This is the left, unconverted channel.
- the converted channel shows only positive values and has about half the range.
- there is no phase change. There can't be, because the conversion code operates on single samples at a time. What you're probably seeing with 2 separately triggered software scopes is that they both trigger on the zero crossing, making it look like a phase change.
So apparently the bipolar ranges from -1 to +1 in terms of DAC output voltages (minimum/maximum voltage), and the unipolar ranges from 0 to +1. In terms of bits, a unipolar is an unsigned 7.25 fixed-point, and the bipolar is an 8.24 signed fixed point (kind of the other way than I originally thought). It is starting to make sense to me now.
Correct me if I still got it wrong. I'm tired.