redeboer · Sep 15, 2026 · Sep 15, 2026 · Sep 15, 2026 · Sep 15, 2026 · Sep 15, 2026
diff --git a/docs/_extend_docstrings.py b/docs/_extend_docstrings.py
@@ -670,9 +670,9 @@ def extend_relativistic_breit_wigner_with_ff() -> None:
     .. math:: {sp.latex(rel_bw_with_ff)}
         :label: relativistic_breit_wigner_with_ff
 
-    where :math:`\Gamma(s)` is defined by :eq:`EnergyDependentWidth`, :math:`B_L^2` is
-    defined by :eq:`BlattWeisskopfSquared`, and :math:`q^2` is defined by
-    :eq:`BreakupMomentumSquared`.
+    where :math:`\Gamma(s)` is defined by :eq:`EnergyDependentWidth`,
+    :math:`\hat{{B}}_L^2` is defined by :eq:`BlattWeisskopfSquared`, and :math:`q^2` is
+    defined by :eq:`BreakupMomentumSquared`.
     """,
     )
 

diff --git a/docs/analyticity/integration-algorithms.ipynb b/docs/analyticity/integration-algorithms.ipynb
@@ -147,7 +147,7 @@
     "  \\frac\n",
     "  {\\rho\\!\\left(s'\\right) n_\\ell^2\\!\\left(s'\\right) ds'}\n",
     "  {\\left(s' - s_\\mathrm{thr}\\right) \\left(s'- s\\right)} \\\\\n",
-    "n_\\ell^2(s') \\;&=\\; \\mathcal{F}_\\ell^2\\!\\left(s', m_1, m_2\\right) \\\\\n",
+    "n_\\ell^2(s') \\;&=\\; \\hat{\\mathcal{F}}_\\ell^2\\!\\left(s', m_1, m_2\\right) \\\\\n",
     "s_\\mathrm{thr} \\;&=\\; (m_1 + m_2)^2\n",
     "\\end{aligned}\n",
     "$$\n",

diff --git a/docs/conf.py b/docs/conf.py
@@ -192,7 +192,6 @@ def _get_dataclasses(module):
     "sphinx_comments",
     "sphinx_copybutton",
     "sphinx_design",
-    "sphinx_hep_pdgref",
     "sphinx_pybtex_etal_style",
     "sphinx_thebe",
     "sphinx_togglebutton",

diff --git a/docs/dynamics.ipynb b/docs/dynamics.ipynb
@@ -80,7 +80,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "AmpForm uses Blatt–Weisskopf functions $B_L$ as _barrier factors_ (also called _form factors_, see {class}`.BlattWeisskopfSquared` and **[TR-029](https://compwa.github.io/report/029)**):"
+    "AmpForm uses Blatt–Weisskopf functions $\\hat{B}_L$ as _barrier factors_ (also called _form factors_, see {class}`.BlattWeisskopfSquared` and **[TR-029](https://compwa.github.io/report/029)**). The hat indicates that they are normalized to $\\hat{B}_L(1)=1$:"
    ]
   },
   {
@@ -151,7 +151,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "The Blatt–Weisskopf form factor is used to 'dampen' the breakup-momentum at threshold and when going to infinity. A usual choice for $z$ is therefore $z=q^2d^2$ with $q^2$ the {class}`.BreakupMomentumSquared` and $d$ the impact parameter (also called meson radius). The {class}`.FormFactor` expression class can be used for this:"
+    "The Blatt–Weisskopf form factor is used to 'dampen' the breakup-momentum at threshold and when going to infinity. A usual choice for $z$ is therefore $z=q^2d^2$ with $q^2$ the {class}`.BreakupMomentumSquared` and $d$ the impact parameter (also called meson radius). The {class}`.FormFactor` expression class can be used for this. Set `normalize=False` to omit the normalization constant $\\left|h_L^{(1)}(1)\\right|$, which gives the non-normalized vertex factor $n_L$ of Equation (50.33) in the [PDG review on resonances](https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=12):"
    ]
   },
   {
@@ -692,7 +692,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.13.13"
+   "version": "3.13.15"
   }
  },
  "nbformat": 4,

diff --git a/docs/dynamics/k-matrix.ipynb b/docs/dynamics/k-matrix.ipynb
@@ -16,7 +16,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "While {mod}`ampform` does not yet provide a generic way to formulate an amplitude model with $\\boldsymbol{K}$-matrix dynamics, the (experimental) {mod}`.kmatrix` module makes it fairly simple to produce a symbolic expression for a parameterized $\\boldsymbol{K}$-matrix with an arbitrary number of poles and channels and play around with it interactively. For more info on the $\\boldsymbol{K}$-matrix, see the classic paper by Chung {cite}`Chung:1995dx`, {pdg-review}`2021; Resonances`, or this instructive presentation {cite}`Meyer:2008-MatrixTutorial`.\n",
+    "While {mod}`ampform` does not yet provide a generic way to formulate an amplitude model with $\\boldsymbol{K}$-matrix dynamics, the (experimental) {mod}`.kmatrix` module makes it fairly simple to produce a symbolic expression for a parameterized $\\boldsymbol{K}$-matrix with an arbitrary number of poles and channels and play around with it interactively. For more info on the $\\boldsymbol{K}$-matrix, see the classic paper by Chung {cite}`Chung:1995dx`, [PDG2026, §Resonances, p.14](https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=14), or this instructive presentation {cite}`Meyer:2008-MatrixTutorial`.\n",
     "\n",
     "Section {ref}`dynamics/k-matrix:Physics` summarizes {cite}`Chung:1995dx`, so that the {mod}`.kmatrix` module can reference to the equations. It also points out some subtleties and deviations.\n",
     "\n",
@@ -415,7 +415,11 @@
     "\n",
     "with $\\gamma_{R,i}$ some _real_ constants and $\\Gamma^0_{R,i}$ the **partial width** of each pole. In the Lorentz-invariant form, the fixed width $\\Gamma^0$ is replaced by an \"energy dependent\" {class}`.EnergyDependentWidth` $\\Gamma(s)$.[^phase-space-factor-normalization] The **width** for each pole can be computed as $\\Gamma^0_R = \\sum_i\\Gamma^0_{R,i}$.\n",
     "\n",
-    "[^phase-space-factor-normalization]: Unlike Eq. (77) in {cite}`Chung:1995dx`, AmpForm defines {class}`.EnergyDependentWidth` as in {pdg-review}`2021; Resonances; p.6`, Eq. (50.28). The difference is that the phase space factor denoted by $\\rho_i$ in Eq. (77) in {cite}`Chung:1995dx` is divided by the phase space factor at the pole position $m_R$. So in AmpForm, the choice is $\\rho_i \\to \\frac{\\rho_i(s)}{\\rho_i(m_R)}$."
+    ":::{warning}\n",
+    "Eq. (50.28) no longer appears in [PDG2026, §Resonances, p.12](https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=12). The width is now defined in terms of bare couplings, $\\Gamma_b(s) = g_b^2\\rho_b(s)n_b^2(s)/m_\\mathrm{BW}$ (Eq. (50.32)), and Eq. (50.35) trades $g_b$ for the partial width $\\Gamma_{\\mathrm{BW},b}$. Combining the two gives the old Eq. (50.28), but the PDG stresses that this substitution is only valid for narrow resonances with all channel thresholds below $m_\\mathrm{BW}$.\n",
+    ":::\n",
+    "\n",
+    "[^phase-space-factor-normalization]: Unlike Eq. (77) in {cite}`Chung:1995dx`, AmpForm defines {class}`.EnergyDependentWidth` as in [PDG2021, Eq. (50.28)](https://pdg.lbl.gov/2021/reviews/rpp2021-rev-resonances.pdf#page=9). The difference is that the phase space factor denoted by $\\rho_i$ in Eq. (77) in {cite}`Chung:1995dx` is divided by the phase space factor at the pole position $m_R$. So in AmpForm, the choice is $\\rho_i \\to \\frac{\\rho_i(s)}{\\rho_i(m_R)}$."
    ]
   },
   {
@@ -441,7 +445,7 @@
     "\n",
     "with $B_{R,i}(q(s))$ the **centrifugal damping factor** (see {class}`.FormFactor` and {class}`.BlattWeisskopfSquared`) for channel $i$ and $\\beta_R^0$ some (generally complex) constants that describe the production information of the decaying state $R$. Usually, these constants are rescaled just like the residue functions in {eq}`residue-function`:\n",
     "\n",
-    "[^damping-factor-P-parametrization]: Just as with [^phase-space-factor-normalization], we have smuggled a bit in the last equation in order to be able to reproduce Equation (50.23) in {pdg-review}`2021; Resonances; p.9` in the case $n=1,n_R=1$, on which {func}`.relativistic_breit_wigner_with_ff` is based.\n",
+    "[^damping-factor-P-parametrization]: Just as with [^phase-space-factor-normalization], we have smuggled a bit in the last equation in order to be able to reproduce [PDG2026, Eq. (50.37)](https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=13) in the case $n=1,n_R=1$, on which {func}`.relativistic_breit_wigner_with_ff` is based.\n",
     "\n",
     "```{margin}\n",
     "{cite}`Chung:1995dx` Eq. (121)\n",
@@ -1332,7 +1336,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "[^pole-vs-resonance]: See {pdg-review}`2021; Resonances`, Section 50.1, for a discussion about what poles and resonances are. See also the intro to Section 5 in {cite}`Chung:1995dx`."
+    "[^pole-vs-resonance]: See [PDG2026, §50.1](https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=1) for a discussion about what poles and resonances are. See also the intro to Section 5 in {cite}`Chung:1995dx`."
    ]
   },
   {

diff --git a/pyproject.toml b/pyproject.toml
@@ -71,7 +71,6 @@ doc = [
     "sphinx-comments",
     "sphinx-copybutton",
     "sphinx-design",
-    "sphinx-hep-pdgref",
     "sphinx-pybtex-etal-style",
     "sphinx-thebe",
     "sphinx-togglebutton",

diff --git a/src/ampform/dynamics/__init__.py b/src/ampform/dynamics/__init__.py
@@ -5,7 +5,7 @@
 
 from __future__ import annotations
 
-from typing import TYPE_CHECKING, Any
+from typing import TYPE_CHECKING, Any, Literal
 from warnings import warn
 
 import sympy as sp
@@ -35,24 +35,47 @@
 
 @unevaluated
 class SimpleBreitWigner(sp.Expr):
-    r"""Simple Breit–Wigner with :math:`m_0 \Gamma_0` in the numerator."""
+    r"""Simple Breit–Wigner with a configurable numerator.
+
+    With the default ``numerator="mass-width"``, the propagator is multiplied by
+    :math:`m_0 \Gamma_0`, so that
+    :math:`\left|\hat{\mathcal{R}}^\mathrm{BW}(m_0^2)\right| = 1`. Set
+    ``numerator="unity"`` for the dressed propagator of `PDG2026, Eq. (50.31)
+    <https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=12>`__, which is
+    also the convention of `MultichannelBreitWigner`.
+    """
 
     s: Any
     mass: Any
     width: Any
-    _latex_repr_ = R"\mathcal{{R}}^\mathrm{{BW}}\left({s}; {mass}, {width}\right)"
+    numerator: Literal["mass-width", "unity"] = argument(
+        default="mass-width", kw_only=True, sympify=False
+    )
 
     def evaluate(self):
         s, m0, w0 = self.args
-        return m0 * w0 * _formulate_breit_wigner(s, m0, w0)
+        numerator = _formulate_numerator(self.numerator, m0, w0)
+        return numerator * _formulate_breit_wigner(s, m0, w0)
+
+    def _latex_repr_(self, printer: LatexPrinter, *args) -> str:
+        s, mass, width = map(printer._print, self.args)
+        function_symbol = _get_breit_wigner_symbol(self.numerator)
+        return Rf"{function_symbol}\left({s}; {mass}, {width}\right)"
 
 
 @unevaluated
 class BreitWigner(sp.Expr):
-    r"""Relativistic Breit–Wigner with :math:`m_0 \Gamma_0` in the numerator.
-
-    Uses an `EnergyDependentWidth` in the denominator (see Equations
-    :eq:`BreitWigner` and :eq:`EnergyDependentWidth`).
+    r"""Relativistic Breit–Wigner with a configurable numerator.
+
+    Uses an `EnergyDependentWidth` in the denominator (see Equations :eq:`BreitWigner`
+    and :eq:`EnergyDependentWidth`). With the default ``numerator="mass-width"``, the
+    propagator is multiplied by :math:`m_0 \Gamma_0`, so that
+    :math:`\left|\hat{\mathcal{R}}^\mathrm{BW}(m_0^2)\right| = 1`, because
+    :math:`\Gamma(m_0^2) = \Gamma_0`. Set ``numerator="unity"`` for the dressed
+    propagator of `PDG2026, Eq. (50.31)
+    <https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=12>`__, which is
+    also the convention of `MultichannelBreitWigner`. The numerator does not affect the
+    `.FormFactor` inside the `EnergyDependentWidth`, where its normalization cancels.
     """
 
     s: Any
@@ -65,12 +88,14 @@ class BreitWigner(sp.Expr):
     phsp_factor: PhaseSpaceFactorProtocol = argument(
         default=PhaseSpaceFactor, sympify=False
     )  # ty: ignore[invalid-assignment]
+    numerator: Literal["mass-width", "unity"] = argument(
+        default="mass-width", kw_only=True, sympify=False
+    )
 
     def evaluate(self):
         width = self.energy_dependent_width()
-        return (
-            self.mass * self.width * _formulate_breit_wigner(self.s, self.mass, width)
-        )
+        numerator = _formulate_numerator(self.numerator, self.mass, self.width)
+        return numerator * _formulate_breit_wigner(self.s, self.mass, width)
 
     def energy_dependent_width(self) -> EnergyDependentWidth | sp.Basic:
         s, m0, w0, m1, m2, ang_mom, d = self.args
@@ -80,7 +105,7 @@ def energy_dependent_width(self) -> EnergyDependentWidth | sp.Basic:
 
     def _latex_repr_(self, printer: LatexPrinter, *args) -> str:
         s = printer._print(self.s)
-        function_symbol = R"\mathcal{R}^\mathrm{BW}"
+        function_symbol = _get_breit_wigner_symbol(self.numerator)
         mass = printer._print(self.mass)
         width = printer._print(self.width)
         arg = Rf"\left({s}; {mass}, {width}\right)"
@@ -94,10 +119,20 @@ def _latex_repr_(self, printer: LatexPrinter, *args) -> str:
 class EnergyDependentWidth(sp.Expr):
     r"""Mass-dependent width, coupled to the pole position of the resonance.
 
-    See Equation (50.28) in :pdg-review:`2021; Resonances; p.9` and
+    See `PDG2021, Eq. (50.28)
+    <https://pdg.lbl.gov/2021/reviews/rpp2021-rev-resonances.pdf#page=9>`__ and
     :cite:`ParticleDataGroup:2012pjm`, equation (6). Default value for
     :code:`phsp_factor` is `.PhaseSpaceFactor`.
 
+    .. warning:: Equation (50.28) no longer appears in
+        `PDG2026, §Resonances, p.12 <https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=12>`__.
+        The width is now defined in terms of bare couplings, :math:`\Gamma_b(s) =
+        g_b^2 \rho_b(s) n_b^2(s) / m_\mathrm{BW}` (Equation (50.32)), and Equation (50.35)
+        trades :math:`g_b` for the partial width :math:`\Gamma_{\mathrm{BW},b}`.
+        Combining the two gives the old Equation (50.28), but the PDG stresses that this
+        substitution is only valid for narrow resonances with all channel thresholds
+        below :math:`m_\mathrm{BW}`.
+
     Note that the `.FormFactor` of AmpForm is normalized in the sense that equal powers
     of :math:`z` appear in the nominator and the denominator, while the definition in
     the PDG (as well as some other sources), always have :math:`1` in the nominator of
@@ -144,7 +179,9 @@ class MultichannelBreitWigner(sp.Expr):
 
     where :math:`g_i^2` is the coupling squared, :math:`\rho_i` is a
     `.PhaseSpaceFactor`, and :math:`F_{L_i}` is a `.FormFactor`. Unlike an
-    `EnergyDependentWidth`, a channel term is not normalized at the pole position.
+    `EnergyDependentWidth`, a channel term is not normalized at the pole position. See
+    `PDG2026, Eqs. (50.31) and (50.32)
+    <https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=12>`__.
     """
 
     s: Any
@@ -174,6 +211,9 @@ class ChannelArguments(sp.Expr):
     .. math::
 
         \Gamma_i^\text{ch}(s) = \frac{g_i^2}{m_0} \rho_i(s) F_{L_i}^2(s)
+
+    See `PDG2026, Eq. (50.32)
+    <https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=12>`__.
     """
 
     s: Any
@@ -196,6 +236,27 @@ def _formulate_breit_wigner(s: Any, mass: Any, width: Any) -> sp.Expr:
     return 1 / (mass**2 - s - sp.I * mass * width)
 
 
+def _formulate_numerator(numerator: str, mass: Any, width: Any) -> Any:
+    _check_numerator(numerator)
+    if numerator == "mass-width":
+        return mass * width
+    return sp.S.One
+
+
+def _get_breit_wigner_symbol(numerator: str) -> str:
+    _check_numerator(numerator)
+    if numerator == "mass-width":
+        return R"\hat{\mathcal{R}}^\mathrm{BW}"
+    return R"\mathcal{R}^\mathrm{BW}"
+
+
+def _check_numerator(numerator: str) -> None:
+    allowed = ("mass-width", "unity")
+    if numerator not in allowed:
+        msg = f"Unknown numerator {numerator!r}, expected one of {', '.join(map(repr, allowed))}"
+        raise ValueError(msg)
+
+
 def relativistic_breit_wigner(s, mass0, gamma0) -> sp.Expr:
     """Relativistic Breit–Wigner lineshape.
 
@@ -220,7 +281,8 @@ def relativistic_breit_wigner_with_ff(  # ruff: ignore[too-many-positional-argum
 ) -> sp.Expr:
     """Relativistic Breit–Wigner with `.FormFactor`.
 
-    See :ref:`dynamics:_With_ form factor` and :pdg-review:`2021; Resonances; p.9`.
+    See :ref:`dynamics:_With_ form factor` and `PDG2026, §Resonances, p.12
+    <https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=12>`__.
     """
     ff = FormFactor(s, m_a, m_b, angular_momentum, meson_radius)
     bw = BreitWigner(

diff --git a/src/ampform/dynamics/builder.py b/src/ampform/dynamics/builder.py
@@ -101,8 +101,9 @@ class RelativisticBreitWignerBuilder:
 
     Args:
         form_factor: Formulate a relativistic Breit–Wigner function multiplied
-            by a Blatt–Weisskopf form factor (`.FormFactor`), like in Equation (50.26)
-            on :pdg-review:`2021; Resonances; p.9`.
+            by a Blatt–Weisskopf form factor (`.FormFactor`), like in `PDG2026, Eqs.
+            (50.33) and (50.37)
+            <https://pdg.lbl.gov/2026/reviews/rpp2026-rev-resonances.pdf#page=12>`__.
         energy_dependent_width: Use an `.EnergyDependentWidth` in the
             denominator of the Breit–Wigner.
         phsp_factor: A class that complies with the