Brian McFee

Ok, google analytics is back up - looks like we get about 4% traffic to the 0.7.2 site, the rest splits between `main` (dev) and `latest` (stable) over the past...

Following up, I have a basic (slow) implementation of the [crystal method](https://www.youtube.com/watch?v=t6twhXA1Gyw) ```python def crystal(N, primes): S = set() root = tuple([0]*len(primes)) S.add(root) seeds = list() for i in range(len(primes)):...

As another aside, summarizing some offline conversation with @lostanlen : - If we do something in this direction, it will apply to VQT but not CQT. CQT should adhere to...

@napulen > I see a common theme here. You could potentially have 12 bins, and not have Equal Temperament (ET). What makes something ET is that the frequency ratio between...

A couple more comments: - In light of the above, I think we might want to not have Pythagorean be an alias of ji3, but a separate implementation that only...

(Brain-dumping a bit more) If we assume that the above function (perhaps with some modification) is satisfactory for constructing intervals, there are now a host of other questions to resolve,...

A couple more comments after spending some more time playing around with tiebreaking: 1. I think I might be okay with the aforementioned delaying of 7 factors compared to 3s...

A fun observation: the above algorithm explores the frontier by selecting the next interval that's closest on average to all previously selected points. If you replace that "average" by a...

> There are two relatively mainstream JI notations (Helmholtz-Ellis, Ben Johnston's) but neither would be very nice to work with Unicode / etc. I think they're actually both doable with...

> I respectfully disagree. Maybe you wouldn't call that serious but i would very much like to have a diatonic transform: (do-re-mi-fa-soh-la-ti), in other words, 7 bins per octave. This...