v1.3.0
This release changes the strategy to gradually increase the permitted adjustment range as confidence builds, instead of doing it in steps.
Gain limits start narrow and widen linearly until full adjustment range is available at 30s, allowing the algorithm to make small initial corrections before committing to its final decision.
Full Changelog: v1.2.2...v1.3.0