There are 2 separate parts to determining the speed of processing: the scan engine and the document class processes.

Capture Engine will extract documents from Laserfiche and queue them up on disk for processing. The speed here can vary based on the activity in the server and the method used to find documents (finding them by direct path is always faster than search). Other factors potentially affecting the speed are the network connection and the disk speed. You can see what the speed is by running the same Capture Engine settings through a session with no processes and only one document class.

The speed of processing documents once they're queued up locally depends on the number of document classes and each individual process. Lookup can be on the slow side, depending on the type of database and network connection. Pattern Matching would depend on how generic your pattern is trying to be.

RAM is less relevant to Quick Fields, processing speed tends to be disk-bound.