User:BudjarnLambeth/12edo as a 2.3.5.17.19 tuning: Difference between revisions
| Line 20: | Line 20: | ||
# If the edo has >40% relative error but <15-20c^ absolute error on any primes N where N is 11 or smaller, then use dual-N for all those primes; then also include every other prime as a single-prime up to the last prime before M, where M is the first prime above 11 with >40% relative error. | # If the edo has >40% relative error but <15-20c^ absolute error on any primes N where N is 11 or smaller, then use dual-N for all those primes; then also include every other prime as a single-prime up to the last prime before M, where M is the first prime above 11 with >40% relative error. | ||
# If the edo has >40% relative error and >15-20c^ absolute error on any primes N where N is 11 or smaller, then use no-N for all those primes; then include every other prime up to the last prime before M, where M is the first or second^ prime above 11 to have >15-20c^ absolute error. | # If the edo has >40% relative error and >15-20c^ absolute error on any primes N where N is 11 or smaller, then use no-N for all those primes; then include every other prime up to the last prime before M, where M is the first or second^ prime above 11 to have >15-20c^ absolute error. | ||
# An addition to the previous step: if the edo approximates any less-than- | # An addition to the previous step: if the edo approximates any less-than-35ish^ multiple of an excluded prime N, with <15-20c^ absolute error, then turn the edo's subgroup into a composite subgroup, and include that multiple as a [[basis element]]. | ||
# Another addition to the previous step: if the edo approximates any 30-integer-limit interval of an excluded prime N, with <~10c^ absolute error, then turn the edo's subgroup into a fractional subgroup, and include that multiple as a basis element, or include a fractional basis element which would make that interval accessible. | # Another addition to the previous step: if the edo approximates any 30-integer-limit interval of an excluded prime N, with <~10c^ absolute error, then turn the edo's subgroup into a fractional subgroup, and include that multiple as a basis element, or include a fractional basis element which would make that interval accessible. | ||
# If none of the above cases are true, but the edo <15-20c^ absolute error on any primes N where N is 11 or smaller on all primes in a prime limit 7 or higher, use that prime limit. If that is the case for all but a small number^ of primes P, then just use the no-P version of the prime limit. | # If none of the above cases are true, but the edo <15-20c^ absolute error on any primes N where N is 11 or smaller on all primes in a prime limit 7 or higher, use that prime limit. If that is the case for all but a small number^ of primes P, then just use the no-P version of the prime limit. | ||