A sine machine driven by a trap signal will have the same torque ripple / noise issue (and for the same reason).
I think the issue is more inductance-related rather than sine vs trap. I haven't seen a stock vs rewound bEMF trace. I can't think of why they'd be different, as most of the time we replace one CFP winding scheme with another CFP wind... I have been wrong in the past, and will be wrong in the future!
Anyone got a scope?