This is fine if you have estimates ahead of time that factor nicely, but I was looking for something that is relatively accurate and reliable where you don't have much to work with ahead of time, or the factors you do have do not nicely fit. I can do a linear approximation, but I have found that that method is not quite as accurate as I would like in general. An example would be log₂30 ≈ 4.90689 I can easily see that log₂16 =4 and log₂32 = 5, so it is between 4 and 5, but closer to 5, and I can approximate this by seeing where 30 falls between 2⁴=16 and 2⁵=32, ie, log₂30 ≈ 4 + (30-16)/(32-16) = 4⅞ = 4.875, but that estimate is off by over 0.031