The ONLY difference between JPLAY and any other control point (which JPLAY attribute to sound improvement) is reduced polling to the renderer (to get track progress).
If you use the WHA to start a queue then closed the app, there will be no communication between the two at all. JPLAY on the other hand will still need to serve the URL of the next track when requested.
If users are unable to discern the difference between an active or inactive WHA, they will be incapable of perceiving any distinction with JPLAY.
BubbleUPnP can be configured to poll every 10 seconds; if that affects sound quality then there's a fundamental problem with the streamer.