The exercise divides the speech level interval between -1 and +1 into 2B uniform intervals (for a uniform quantizer) and into 2B psuedo-logarithmically spaced intervals (for a mu-law quantizer). The exact locations of each quantization interval depends on whether the quantization uses mid-riser or mid-tread intervals, and it also depends on whether the quantization uses rounding and saturating intervals, or truncating and saturating intervals.