You have analog QAM and digital QAM confused. CQUAM limited negative peaks at 96% modulation because at maximum modulation, the carrier is phase shifted relative to the linear sidebands. This difference results in an effective carrier that instantaneously appears to be 96% of what it should be to the sidebands. Negative modulation peaks are limited to 96% for THAT reason. There are no such constraints with an IBOC analog OR digital signal.
No, there are no such constraints, but running the modulation past the point where the carrier is so weak or neutralized that there is
no reference for the constellation, the "bits" in the constellation do not "exist".
Sufficent data redundancy, buffering, and high/low data modes permit some loss of data.
If you have a suggestion for how the constellation can exist without reference to a "somewhat" continuous rate function such as a sine wave or even a square wave, please elaborate.
Even though the power available in the "carrier" varies as modulation subdivides the power into sdiebands, enough remains
(if not cut off completely) to provide reference for the 500 watts of data modulation at the frequencies above 5 khz.
I've never had occasion to find that my mind has generated "false facts", so I can generally consider that when my mind tells
me I've read something factworthy AND it fits in with reality as I have experienced and been taught in radio engineering school,
then it's a fact, amd I retain it for reference at times such as this.
Please cite ibiquity documentation that would suggest uncontrolled, over 100% negative modualtion peaks are "just fine".
Even the digital-only ibiquity AM method REQUIRES a small "vestigal" carrier for the bits to be referenced to.