But when I got Graham motor I tried rubber, original thread and some special fishing string as a belt again before I changed it to Delrin pulley and mylar tape. Difference was huge and striking. The highs were so dirty and compressed! I could not believe it. I measured speed (3150hz) and W&F ( 0.07% ) using AnalogMagik software. The only conclusion would be that these tiny thread and string transmitted huge noise to platter! Hard to believe but it is my experience.
From my experience of owning the Platine Verdier, the motor is the issue.
It is tiny, and not really designed for side loads on the spindle. Ultimately it will generate noise, which is transmitted via the belt. No doubt the delrin pulley helped you in this case. If you ran a double pulley, with a flywheel mounted on the opposite side of the pulley from the platter, you could reduce motor noise by "balancing the load" on the motor spindle.
By the way, I remounted my Verdier motor into a cradle clamp arrangement inside the main housing and eliminated the rubber mounts. This enabled me to reduce the motor noise significantly by reducing noise from the motor case itself by carefully tuned clamping. I replaced the rubber mounts with teflon washers. This arrangement also improved speed stability.
As a test to get a handle on the performance of the motor, I ran my Final Audio Parthenon AC motor and controller, driven by a 60wpc power amp on the Verdier in a trial ( both with thread drive ) - the improvement in timing and drop in noise floor were significant. The Final AC motor is dead silent, even though its about 4/5 times larger, it has a proper machined spindle, oilite bushes, thrust pad arrangement.