hi eureka and swiftnicholas
i only put together that model so that i'd have an idea of how greek might have sounded on the basis of the evidence i read... all it does is show how homer sounds if you apply the statistically significant patterns found in extant greek music (found by devine and stephens). i can't say for sure that there aren't glaring errors in it because i've read references about e.g. a papyrus fragment containing a pitch transcription of a menander iambic line, and the "new guide to accenting greek" suggests that the 2nd beat of a long acute is on a higher pitch than the 1st beat. i'm also not sure if non-accented proclitics cause anathesis, devine and stephens equivocated on that so i'd have to read all the greek fragments again... another thing is that there's this opinion that non-strophic music (like homer) followed the pitch accents strictly but strophic music (like choruses of drama and pindar) didn't, i think it's on the basis of dionysus of hal's pitch description of a line of euripides' lyrics where he describes non-accented pitches sung lower than accented pitches and things like that... on first glance to me it looked like the pitch of the strophe in euripides might be following in responsion the pitch of the antistrophe, which would be interesting, but i haven't had a chance to look at things like this more closely
the 7-tone scale came out naturally as i tried to model the things in devine and stephens, given that i found they described 4 different pitch peaks and 3 different types of pitch drop; as i applied this to texts they all seemed to stay consistently within a 7-tone range, only then i read up on the standard 7-tone scale of ancient greek music. it might be coincidence.
i think the singers of homer would have kept strictly to whatever scale they used, which is what aristoxenus says expressly in his book on music theory. there are lots of unknowns though, in addition to the uncertainty of the whole model, e.g. whether it was sung to an enharmonic scale or chromatic or diatonic (or to different scales at different times), my assumption is that they used the enharmonic scale, which would explain why accented syllables in names of people and in words following grave-accented words are on average 2 whole tones higher (i think that's right from memory) than accented syllables in other words.
but because of all these uncertainties there's no point me recording anything; it's just my general guess about how it might have sounded; i only put it together because i'm an auditory learner and i didn't want the terrible stress pronunciation of academics i've heard to get stuck in my head as a beginner. there was nothing out there which described a technique for doing this so i had to make up one; even devine and stephens just present data about pronunciation and leave it to the reader to come up with a technique about how to apply it.
i've seen that stefan hagel has written some sort of software to do what my model does but automatically, he sent me the first few lines of the iliad which have a pitch trace line over the top of each line. it just shows relatively how the syllables probably sat next to each other, which is all my model did as well. i hope this ramble answers what you asked, but if you have any other questions let me know, thanks