It will cost additional time for sure, but material cost can be limited greatly with some clever engineering upfront. One does not need to build many enclosure incarnations. One flexible enclosure can be used to test them all.While the D&D approach would seemingly be lower cost, if we have to iterate multiple prototypes to get the tuning right, am less inclined to take on the likely added cost of time/materials to do so.
Looking forward to the requirements list! Probably good to also write down a list of open questions that need to be answered to complete the list. That would help steer discussion.