Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Aether Catalog definitions — Algebra_SmoothWindows_GaborOperators

Definition
Algebra_SmoothWindows_GaborOperators

by raver1975 · Sep 17, 2026 · Mathlib c5ea003 (Lean v4.30.0)

aether-catalogalgebra

Definition bundle for the Aether Catalog module Algebra.SmoothWindows.GaborOperators, transplanted by skeleton subtraction; supplies the types and constants the catalog's theorems import.

Definition code
-- Def bundle generated from Algebra/SmoothWindows/GaborOperators.lean by skeleton subtraction
import Mathlib

/-!
# Smooth windows I: translation, modulation and the Weyl relation

The Reciprocal-Zero Harmonics programme (`Algebra.ReciprocalZeroHarmonics.Core`,
`Algebra.ReciprocalZeroHarmonics.WindowDichotomy`) analyses spectral data through the
*rectangular* window `|Im ρ| ≤ T`.  This file lays the algebraic foundation for replacing that
sharp cutoff by a smooth (Gaussian / Schwartz) window: the two families of operators that move a
window around phase space,

* **translation** `(T_a f)(t) = f (t - a)`,
* **modulation** `(M_b f)(t) = e^{2πi b t} f (t)`,

and the *modulation/translation identity* (the Weyl commutation relation) that they satisfy.

## Main results

* `modOp_transOp` — **Weyl commutation relation**: `M_b T_a = χ(b a) · T_a M_b`, where
  `χ(x) = e^{2πix}`.  Translation and modulation commute only up to the phase `e^{2πi ab}`; this
  failure of commutativity *is* the Heisenberg group.
* `Heis` — the (circle-valued) **Heisenberg group** with multiplication
  `(a,b,z)·(a',b',z') = (a+a', b+b', z z' χ(b a'))`, with a full `Group` instance proved from the
  cocycle identity.
* `heisRep` — the **Schrödinger/Gabor representation** `Heis →* Function.End (ℝ → ℂ)`,
  `(a,b,z) ↦ z · T_a M_b`: a genuine monoid homomorphism, i.e. the Weyl relation is exactly what
  makes the composite of two Gabor shifts a third one.
* `heisRep_injective` — the representation is **faithful**: the Gaussian is a test vector which
  separates all Heisenberg elements.  Consequently the phase factor in `modOp_transOp` cannot be
  removed by any renormalisation of the operators.
* `modOp_transOp_ne` — an adversarial check: the phase is really needed, the two orders of
  composition differ on any window that is nonzero at the origin.

-- !-- Lab Notes -- !--
* **Hypothesis (Hypothesizer).** Windowing should be an *action of a group*, not an ad hoc
  cutoff; the correct group is the Heisenberg group, and the rectangular window of the catalog is
  merely one (non-smooth) vector in a representation space.
* **Experiment (Experimenter).** All operator identities reduce to `Complex.exp_add` after
  clearing casts; the group axioms for `Heis` reduce to the cocycle identity
  `b·a' + (b+b')·a'' = b'·a'' + b·(a' + a'')`.  Faithfulness needed a genuinely analytic input:
  the Gaussian test vector is nowhere zero and has a unique maximum, which pins down `a`, then
  the phase `z`, then the modulation `b` via `χ(1/4) = i ≠ 1`.
* **Analysis (Analyst).** The phase cocycle `χ(ba')` is not a coboundary — faithfulness shows the
  central `Circle` factor is not redundant.  This is the structural reason a *smooth* window can
  be moved freely in phase space while keeping all of its algebraic bookkeeping exact.
* **Critique (Critic).** Nothing here is definitional: `modOp_transOp` is false without the phase
  (`modOp_transOp_ne`), and `heisRep_injective` shows the group is not a proper quotient.
-/

namespace SmoothWindows

open Complex Real

/-! ## The character `χ(x) = e^{2πix}` -/

/-- The basic additive character `χ(x) = e^{2πix}` of `ℝ`. -/
noncomputable def chi (x : ℝ) : ℂ := Complex.exp ((2 * π * x : ℝ) * Complex.I)

@[simp] theorem chi_zero : chi 0 = 1 := by simp [chi]

theorem chi_add (x y : ℝ) : chi (x + y) = chi x * chi y := by
  unfold chi
  rw [← Complex.exp_add]
  push_cast
  ring_nf

@[simp] theorem chi_ne_zero (x : ℝ) : chi x ≠ 0 := Complex.exp_ne_zero _


theorem chi_neg (x : ℝ) : chi (-x) = (chi x)⁻¹ := by
  refine eq_inv_of_mul_eq_one_left ?_
  rw [← chi_add]
  simp

/-- `χ` is the coercion of `Circle.exp (2πx)`. -/
theorem coe_circleExp (x : ℝ) : (Circle.exp (2 * π * x) : ℂ) = chi x := by
  rw [Circle.coe_exp, chi]



/-! ## Translation and modulation -/

/-- **Translation** `(T_a f)(t) = f (t - a)`. -/
def transOp (a : ℝ) (f : ℝ → ℂ) : ℝ → ℂ := fun t => f (t - a)

/-- **Modulation** `(M_b f)(t) = e^{2πi b t} f (t)`. -/
noncomputable def modOp (b : ℝ) (f : ℝ → ℂ) : ℝ → ℂ := fun t => chi (b * t) * f t










/-! ## The Heisenberg group and the Gabor representation -/

/-- The (circle-valued) **Heisenberg group**: triples `(a, b, z)` of a translation, a modulation
and a unimodular phase, multiplied through the Weyl cocycle. -/
structure Heis where
  /-- the translation parameter -/
  a : ℝ
  /-- the modulation parameter -/
  b : ℝ
  /-- the central phase -/
  z : Circle

namespace Heis

@[ext] theorem ext {g h : Heis} (ha : g.a = h.a) (hb : g.b = h.b) (hz : g.z = h.z) : g = h := by
  cases g; cases h; simp_all

noncomputable instance : One Heis := ⟨⟨0, 0, 1⟩⟩

noncomputable instance : Mul Heis :=
  ⟨fun g h => ⟨g.a + h.a, g.b + h.b, g.z * h.z * Circle.exp (2 * π * (g.b * h.a))⟩⟩

noncomputable instance : Inv Heis :=
  ⟨fun g => ⟨-g.a, -g.b, g.z⁻¹ * Circle.exp (2 * π * (g.b * g.a))⟩⟩


@[simp] theorem one_a : (1 : Heis).a = 0 := rfl
@[simp] theorem one_b : (1 : Heis).b = 0 := rfl
@[simp] theorem one_z : (1 : Heis).z = 1 := rfl
@[simp] theorem mul_a (g h : Heis) : (g * h).a = g.a + h.a := rfl
@[simp] theorem mul_b (g h : Heis) : (g * h).b = g.b + h.b := rfl
@[simp] theorem mul_z (g h : Heis) :
    (g * h).z = g.z * h.z * Circle.exp (2 * π * (g.b * h.a)) := rfl
@[simp] theorem inv_a (g : Heis) : g⁻¹.a = -g.a := rfl
@[simp] theorem inv_b (g : Heis) : g⁻¹.b = -g.b := rfl
@[simp] theorem inv_z (g : Heis) : g⁻¹.z = g.z⁻¹ * Circle.exp (2 * π * (g.b * g.a)) := rfl

/-- The Heisenberg group law: associativity is precisely the 2-cocycle identity for the Weyl
phase `2π b a'`. -/
noncomputable instance : Group Heis where
  mul_assoc g h k := by
    refine Heis.ext (by simp [add_assoc]) (by simp [add_assoc]) ?_
    have e1 : Circle.exp (2 * π * ((g.b + h.b) * k.a))
        = Circle.exp (2 * π * (h.b * k.a)) * Circle.exp (2 * π * (g.b * k.a)) := by
      rw [← Circle.exp_add]; ring_nf
    have e2 : Circle.exp (2 * π * (g.b * (h.a + k.a)))
        = Circle.exp (2 * π * (g.b * h.a)) * Circle.exp (2 * π * (g.b * k.a)) := by
      rw [← Circle.exp_add]; ring_nf
    simp only [mul_z, mul_a, mul_b, e1, e2]
    simp [mul_comm, mul_left_comm, mul_assoc]
  one_mul g := Heis.ext (by simp) (by simp) (by simp)
  mul_one g := Heis.ext (by simp) (by simp) (by simp)
  inv_mul_cancel g := by
    refine Heis.ext (by simp) (by simp) ?_
    have : Circle.exp (2 * π * (g.b * g.a)) * Circle.exp (2 * π * (-g.b * g.a)) = 1 := by
      rw [← Circle.exp_add, show 2 * π * (g.b * g.a) + 2 * π * (-g.b * g.a) = 0 by ring]
      simp
    simp only [mul_z, inv_b, inv_z, one_z]
    rw [mul_assoc, mul_mul_mul_comm, inv_mul_cancel, this, mul_one]

end Heis

/-- The **Schrödinger (Gabor) representation** of the Heisenberg group on functions `ℝ → ℂ`:
`(a, b, z) ↦ z · T_a M_b`. -/
noncomputable def gaborAct (g : Heis) (f : ℝ → ℂ) : ℝ → ℂ :=
  fun t => (g.z : ℂ) * transOp g.a (modOp g.b f) t

theorem gaborAct_apply (g : Heis) (f : ℝ → ℂ) (t : ℝ) :
    gaborAct g f t = (g.z : ℂ) * chi (g.b * (t - g.a)) * f (t - g.a) := by
  simp [gaborAct, transOp, modOp, mul_assoc]

@[simp] theorem gaborAct_one (f : ℝ → ℂ) : gaborAct 1 f = f := by
  funext t; simp [gaborAct_apply]

/-- **The representation property.**  The composite of two Gabor shifts is the Gabor shift of the
Heisenberg product — this is exactly the modulation/translation identity in group form. -/
theorem gaborAct_mul (g h : Heis) (f : ℝ → ℂ) :
    gaborAct (g * h) f = gaborAct g (gaborAct h f) := by
  funext t
  simp only [gaborAct_apply, Heis.mul_a, Heis.mul_b, Heis.mul_z, Circle.coe_mul, coe_circleExp]
  rw [show t - (g.a + h.a) = (t - g.a) - h.a by ring,
    show chi ((g.b + h.b) * (t - g.a - h.a))
      = chi (g.b * (t - g.a)) * chi (h.b * ((t - g.a) - h.a)) * (chi (g.b * h.a))⁻¹ by
      rw [← chi_add, ← chi_neg, ← chi_add]; congr 1; ring]
  field_simp

/-- The Gabor representation packaged as a monoid homomorphism into the endomorphism monoid of
the space of windows. -/
noncomputable def heisRep : Heis →* Function.End (ℝ → ℂ) where
  toFun g := gaborAct g
  map_one' := by funext f; exact gaborAct_one f
  map_mul' g h := by funext f; exact gaborAct_mul g h f


/-- The Gaussian test window used to prove faithfulness. -/
noncomputable def gaussTest : ℝ → ℂ := fun t => ((Real.exp (-t ^ 2) : ℝ) : ℂ)




end SmoothWindows

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me