-----
You wrote:
Good observations. Comparative historical linguistics also comes into play here. If ἀδελφοί was pronounced as ἀαδελφί during the Classical period, why would it be transliterated into Latin as adelphoe during that same period? One tiny example, but the evidence against Dr. Zachariou's thesis is overwhelming, and supported by the majority of scholars, including native Greek speakers. Perhaps we could email my old classmate Maria Pantelia, now at UC-Irvine?

-----
Here is my response, but first, some background—even some extra tips—from my book. (See www.JesusspokeGreek.com)
Erasmus figured that the ancient Greek pronunciation could be reconstructed using Latin as a basis. Ironically, by then (early 1500s) Latin had long been metamorphosed into several phonologically different major languages (French, Italian, Portuguese, Romanian, Spanish), and a number of now extant dialects. Here are some of the risks of “erasmianizing” Latin in explaining away Greek spelling and sounds:
1. Latin often renders both η and ε as e, but these two are not the same any more than the i in Virgilius and the e in Vergilius. Until about 450 B.C. or later Attic Greek often used E for H and E. So, Latin would naturally use E as well, whether the actual sound was ε /ε/ or η /i/. Also, Greek sometimes used η and ε alternatively as in εὐλόγησεν, ηὐλόγησεν “he blessed,” a difference in spelling and pronunciation. Latin, of course, would use e in either case.
2. The digraph αι in Καίσαρ [kεsar] Caesar is the same sound as ε, but Erasmians see αι as the diphthong in aisle, thus mispronouncing Kaiser [kaɪzər]. Interestingly, Spanish e in Cesar [sεsar] betrays the original [ε] sound.
3. And now regarding your comment. Latin gives οι as oe, but οι is neither the oe sound in poet nor the oi in oil, as Erasmian would have it. Even English(!) attests to Latin oe as οι (= ι):
Phoenix [finɪx], Phoebe [fibɪ], amoeba [əmibə] from Greek Φοίνιξ, Φοίβη, ἀμοιβάς.
Cheers,
Philemon Zachariou