According to my experience and turntables that I measured, direct drive's wow&flutter performance isn't better than belt drive. Basic and very cheap direct drives can be very good up to a certain point but after that even high end direct drives can not improve w&f figures much further. Another strange thing that some modern direct drives offer worse performance than old, vintage ones. The best w&f results can be achieved with top belt drives according to my experience. I haven’t measure modern idlers but old ones are not far behind of direct drives.
Belt drive’s problem is belt, direct drive and idler’s problem is not being able to separate motor from the plinth (there are some exceptions like Esoteric T1). Direct drive motors have another disadvantage like turning at platter speed (33.3rpm). Idler's disadvantage is noise because of idler mechanism.
IMHO/IME there are more important factors than drive type which makes a turntable sound good or bad. One of those factors is high mass platter. A high mass platter (preferably over 20kg) is the key factor for a good sounding turntable and helps solving problems introduced by drive type (BD, DD or idler). I believe implementation and how good a turntable is built is more important than drive type.
Keeping steady speed at 33.3rpm and w&f performance are two different things. Another very important problem for all drive types is aggressively trying to keep 33.3rpm speed by control electronics. That aggressive control may achieve 33.3rpm average but for the cost of increased w&f. Piano and violin sounds are very sensitive to w&f. It's a good idea to check how piano and violin sound when evaluating a turntable.
Wonderful post Mtemur. I remember being surprised reading the W&F measurements in I think HiFi+ of the Air Force one and the SME 30. They were very low and I think the best turntables in this regard that the magazine had measured.
I agree that a high mass platter and build quality are the key factors but also how are the motor controls the platter and the speed control electronics themselves. If the motor pulley is able to maintain constant speed without constant correction, aided by a fly wheel and heavy mass platter with very low bearing friction, speed stability and consistency should be very accurate.
I have found through experimentation that the less influence the motor has on the platter the better. Separation and isolation helps, but so does a very light touch between driving pulley and platter. My thin non-stretch reinforced thread is quite loose. My platter weighs 65kg, and it floats on air. Speed might change from 33.333 to 33.334 on the RoadRunner Tachometer. And as you say, violin and piano, the complete note from attack to decay makes speed consistency quite easy to hear. Clarity, especially in bass and overall nuance also improve.
33.3 is pretty easy to achieve. 33.33 or 33.333 is more difficult. I would like to see studies about what is actually audible.