AFTD Auto-Formalizing Theoretical Domains

energy_filter_delta_renyi_constant

theoremverified

For the Kronecker delta filter delta : ℤ → ℝ defined by delta 0 = 1 and delta j = 0 for j ≠ 0, and any real q > 0, energy_filter_renyi_constant delta q equals (4 / Real.sqrt (2 * Real.pi)) ^ q * Real.sqrt (Real.pi / (8 * q)).

Statement

theorem energy_filter_delta_renyi_constant (q : ℝ) (hq : 0 < q) :
    let delta : ℤ → ℝ := fun j => if j = 0 then 1 else 0
    energy_filter_renyi_constant delta q =
      (4 / Real.sqrt (2 * Real.pi)) ^ q * Real.sqrt (Real.pi / (8 * q))
Source
knowledgebase, OP-1115
Verified
05 Nov 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 6
  • energy_filter_renyi_constantFor a filter sequence a : ℤ → ℝ and real exponent q, energy_filter_renyi_constant a q is (4 / Real.sqrt (2 * Real.pi)) ^ q * Real.sqrt…
  • energy_filter_fourier_symbolFor a sequence a : ℤ → ℝ and real θ, energy_filter_fourier_symbol a θ is the Fourier symbol A(θ) = ∑' j : ℤ, (a j : ℂ) * Complex.exp…
  • energy_filter_sequence_lattice_s2For a real sequence a on Z, energy_filter_sequence_lattice_s2 a is the sum of (a j)^2 over all j in Z.
  • energy_filter_delta_lattice_sumsFor the Kronecker delta filter delta : ℤ → ℝ defined by delta 0 = 1 and delta j = 0 for j ≠ 0, the second-moment lattice sum…
  • energy_filter_sequence_lattice_s4For a real sequence a on Z, energy_filter_sequence_lattice_s4 a is the sum over m in Z of the squared discrete convolution (sum_{j in Z}…
  • energy_filter_sequence_purity_constantFor a real sequence a on Z, energy_filter_sequence_purity_constant a is (2 / sqrt(pi)) times the ratio of the fourth-moment convolution…

Read back from the Lean

For any real number $q$ such that $q > 0$, letting $\delta : \mathbb{Z} \to \mathbb{R}$ be the sequence defined by $\delta(j) = 1$ if $j = 0$ and $0$ otherwise, the quantity energy_filter_renyi_constant delta q equals $\left(\frac{4}{\sqrt{2\pi}}\right)^q \sqrt{\frac{\pi}{8q}}$.

Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.

The proof, explained

The proof evaluates the Rényi constant of the Kronecker delta filter delta directly from its definition. By definition, the Rényi constant equals C_q * ((1 / (2π)) * ∫₀^{2π} ‖Fourier(delta)(θ)‖^{2q} dθ) / S₂^q, where C_q is the target factor (4 / √(2π))^q √(π / (8q)). Because delta is supported solely at index 0, its Fourier symbol is identically 1, making the integral equal to 2π. Since the lattice sum S₂ is also 1, the integral and normalization factors cancel out completely, leaving only C_q.

  1. The Fourier symbol of the delta filter is identically 1: for all θ ∈ ℝ, energy_filter_fourier_symbol delta θ = 1.

    We introduce the filter delta, expand the definition of the Fourier symbol as the infinite sum ∑ j, delta(j) * exp(i * j * θ), and apply tsum_eq_single at j = 0. Since delta(j) = 0 for all j ≠ 0, all other terms vanish, leaving delta(0) * exp(0) = 1.

    lines 6–13

    uses tsum_eq_single

      intro delta
      have hfour : ∀ θ : ℝ, energy_filter_fourier_symbol delta θ = 1 := by
        intro θ
        simp only [energy_filter_fourier_symbol]
        rw [tsum_eq_single (0:ℤ)]
        · simp [delta]
        · intro b hb
          simp [delta, hb]
  2. For all θ ∈ ℝ, ‖energy_filter_fourier_symbol delta θ‖ ^ (2 * q) = 1.

    Substituting the previous result, the Fourier symbol is 1 for every θ. Its complex norm is ‖1‖ = 1, and 1 raised to any real power 2q is 1.

    lines 14–17

      have hnorm : ∀ θ : ℝ, ‖energy_filter_fourier_symbol delta θ‖ ^ (2*q) = 1 := by
        intro θ
        rw [hfour θ]
        simp
  3. The integral ∫ θ in 0..(2 * Real.pi), ‖energy_filter_fourier_symbol delta θ‖ ^ (2 * q) equals 2 * Real.pi.

    Replacing the integrand with the constant 1 using the previous identity, intervalIntegral.integral_const computes the integral over [0, 2π] as (2π - 0) * 1 = 2π.

    lines 18–21

    uses intervalIntegral.integral_const

      have hint : ∫ θ in (0:ℝ)..(2*Real.pi), ‖energy_filter_fourier_symbol delta θ‖ ^ (2*q) = 2*Real.pi := by
        simp only [hnorm]
        rw [intervalIntegral.integral_const]
        simp
  4. The second-order lattice sum S₂ for delta equals 1.

    We extract the first conjunct of the lemma energy_filter_delta_lattice_sums, which establishes that the lattice sum S₂ = ∑ j, (delta j)^2 equals 1 for the delta filter.

    lines 22–25

    uses energy_filter_delta_lattice_sums

      have hs2 : energy_filter_sequence_lattice_s2 delta = 1 := by
        have h := energy_filter_delta_lattice_sums
        simp only at h
        exact h.1
  5. The Rényi constant of delta equals (4 / Real.sqrt (2 * Real.pi)) ^ q * Real.sqrt (Real.pi / (8 * q)).

    We unfold the definition of energy_filter_renyi_constant and substitute the computed integral value 2π and lattice sum S₂ = 1. Simplifying 1^q = 1 via Real.one_rpow and dividing by 1 via div_one leaves the prefactor multiplied by (1 / (2π)) * (2π); field_simp finishes the calculation using Real.pi_ne_zero.

    lines 26–31

    uses Real.one_rpow, div_one, Real.pi_ne_zero

      simp only [energy_filter_renyi_constant]
      rw [hint, hs2]
      rw [Real.one_rpow]
      rw [div_one]
      have hpi : Real.pi ≠ 0 := Real.pi_ne_zero
      field_simp
tsum_eq_single
If all terms of a sum other than the one at index b are zero, the infinite sum equals the term at b.
intervalIntegral.integral_const
The integral of a constant c on the interval [a, b] is (b - a) • c.
energy_filter_delta_lattice_sums
For the Kronecker delta filter, the lattice sums satisfy S₂ = 1 and S₄ = 1, and its purity constant is 2 / √π.
Real.one_rpow
For any real x, 1^x = 1.
div_one
For any element a in a division monoid, a / 1 = a.
Real.pi_ne_zero
The real constant π is non-zero.

Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

/-- Exact evaluation of the Renyi constant for the Kronecker delta energy filter. -/theorem energy_filter_delta_renyi_constant (q : ℝ) (hq : 0 < q) :    let delta : ℤ → ℝ := fun j => if j = 0 then 1 else 0    energy_filter_renyi_constant delta q =      (4 / Real.sqrt (2 * Real.pi)) ^ q * Real.sqrt (Real.pi / (8 * q)) := by  intro delta  have hfour : ∀ θ : ℝ, energy_filter_fourier_symbol delta θ = 1 := by    intro θ    simp only [energy_filter_fourier_symbol]    rw [tsum_eq_single (0:ℤ)]    · simp [delta]    · intro b hb      simp [delta, hb]  have hnorm : ∀ θ : ℝ, ‖energy_filter_fourier_symbol delta θ‖ ^ (2*q) = 1 := by    intro θ    rw [hfour θ]    simp  have hint : ∫ θ in (0:ℝ)..(2*Real.pi), ‖energy_filter_fourier_symbol delta θ‖ ^ (2*q) = 2*Real.pi := by    simp only [hnorm]    rw [intervalIntegral.integral_const]    simp  have hs2 : energy_filter_sequence_lattice_s2 delta = 1 := by    have h := energy_filter_delta_lattice_sums    simp only at h    exact h.1  simp only [energy_filter_renyi_constant]  rw [hint, hs2]  rw [Real.one_rpow]  rw [div_one]  have hpi : Real.pi ≠ 0 := Real.pi_ne_zero  field_simp
See it in the graphDiscuss this resultChallenge it