When configured as a DDS (TP_MIXER_MODE=0), the output of the DDS is intended to be the components of a unit vector. For TT_DATA = cfloat, this means that the outputs will be in the range -1.0 to +1.0. For TT_DATA = cint16 and TT_DATA = cint32 the output is scaled by two to the power 15 and 31 respectively such that the binary point follows the most significant bit of the output. Therefore, if the DDS output is used to multiply/mix, you must account for this bit shift.