≤ c(µ, Y )ψ(X) f BMO(dµ) for all cubes Q in R n and every function f ∈ BMO(dµ), where µ is a doubling measure in R n , Y is some positive functional defined on cubes,is a sufficiently good quasi-norm and c(µ, Y ) and ψ(X) are positive constants depending on µ and Y , and X, respectively. That abstract scheme allows us to recover the sharp estimatefor every cube Q and every f ∈ BMO(dµ), which is known to be equivalent to the John-Nirenberg inequality, and also enables us to obtain quantitative counterparts when L p is replaced by suitable strong and weak Orlicz spaces and L p(•) spaces.Besides the aforementioned results we also generalize [OPRRR20, Theorem 1.2] to the setting of doubling measures and obtain a new characterization of Muckenhoupt's A∞ weights.
Contents1. Introduction 1 2. The A ∞ condition of a functional with respect to a quasi-norm 7 3. A new quantitative self-improving theorem for BMO functions 13 3.1. Lemmata 13 3.2. Proof of the self-improving theorem 14 4. Applications of the self-improving theorem 17 4.1. BMO-type improvement at the Orlicz spaces scale 18 4.2. BMO-type improvement at the variable Lebesgue spaces scale 24 Appendix A. Proof of Theorem 2.1 26 Acknowledgements 30 References 31