In the course of a discussion with Alexandre I ran into something really interesting: [delread~] and [vd~] have different delay limits! While [delread~] has always the buffersize minus the blocksize of the subpatch where it is located, the limit of [vd~] is 64 samples greater. Any explanations?

 

In my example patch, simply choose any blocksize, then set the delay time to maximum 100 (which is actually beyond the maximum), and then toggle between [vd~] and [delread~] to see the 64 samples difference...