{
  "id": "hilbert-transform",
  "title": "Hilbert Transform",
  "lang": "en",
  "source_lang": "en",
  "summary": "The Hilbert transform is a singular integral operator connecting harmonic conjugates, frequency-dependent phase shifts, analytic signals, and causal response functions.",
  "category": "mathematics",
  "subcategory": "analysis",
  "type": "concept",
  "aliases": [
    "Hilbert transformation",
    "conjugate-function operator"
  ],
  "generated_at": "2026-11-04",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/hilbert-transform/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/hilbert-transform/",
      "json": "https://aiwiki.page/zh/hilbert-transform.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "The Hilbert transform is a linear operator that converts a function into a related function by integrating against the singular kernel \\(1/(\\pi x)\\). Unlike an ordinary averaging operation, it combines contributions with opposite signs from opposite sides of each point. In the frequency domain, it shifts positive-frequency components by \\(-90^\\circ\\) and negative-frequency components by \\(+90^\\circ\\), without changing their amplitudes. These properties make it a fundamental object in harmonic analysis and a standard tool in signal processing.\n\nNamed after David Hilbert, the transform connects several mathematical descriptions of the same phenomenon: singular integration on the real line, conjugate harmonic functions in a half-plane, and separation into positive and negative frequencies. Its theory explains both its usefulness and its limitations. It is well behaved on many function spaces, but not on all; its kernel requires a carefully defined limiting process; and numerical implementations necessarily approximate an ideal, nonlocal operation."
    },
    {
      "heading": "Definition and the principal-value integral",
      "text": "For a sufficiently regular, decaying function \\(f\\) on the real line, the Hilbert transform is defined by\n\n\\[(Hf)(x)=\\frac{1}{\\pi}\\operatorname{p.v.} \\int_{-\\infty}^{\\infty}\\frac{f(t)}{x-t}\\,dt.\\]\n\nHere \\(\\operatorname{p.v.}\\) denotes the Cauchy principal value. More explicitly,\n\n\\[(Hf)(x)= \\lim_{\\varepsilon\\downarrow0} \\frac{1}{\\pi} \\int_{|x-t|>\\varepsilon}\\frac{f(t)}{x-t}\\,dt,\\]\n\nprovided the integral at infinity is meaningful and the displayed limit exists. The symmetric exclusion of an interval around \\(t=x\\) is essential: the two sides usually diverge separately, while their combined contribution can converge. Some authors use the kernel \\(1/(t-x)\\), producing the negative of this convention.\n\nThe singularity is controlled by cancellation rather than by absolute integrability. In a symmetric neighborhood of \\(x\\), the constant part \\(f(x)\\) contributes zero because the kernel is odd. The remaining local contribution can therefore be written using\n\n\\[\\frac{f(t)-f(x)}{x-t}.\\]\n\nIf \\(f\\) satisfies a local Hölder condition\n\n\\[|f(t)-f(x)|\\leq C|t-x|^\\alpha,\\qquad \\alpha>0,\\]\n\nthis quotient is bounded in magnitude by \\(C|t-x|^{\\alpha-1}\\), which is locally integrable. Conditions controlling the behavior at infinity are a separate requirement. This distinction between local cancellation and distant contributions is basic to the theory of singular integral operators.\n\nFormally, the transform is a convolution:\n\n\\[Hf=K*f,\\qquad K(x)=\\operatorname{p.v.}\\frac{1}{\\pi x}.\\]\n\nThe kernel is not an ordinary integrable function. It can instead be interpreted through distribution theory, where its action on a smooth, rapidly decreasing test function \\(\\varphi\\) is defined by\n\n\\[\\langle K,\\varphi\\rangle= \\lim_{\\varepsilon\\downarrow0} \\frac{1}{\\pi}\\int_{|x|>\\varepsilon}\\frac{\\varphi(x)}{x}\\,dx.\\]\n\nThis formulation gives a precise meaning to convolution and Fourier identities that would otherwise involve divergent integrals.\n\nA simple example exposes the transform’s nonlocal character. For the indicator function of an interval,\n\n\\[H\\mathbf{1}_{(a,b)}(x) =\\frac{1}{\\pi}\\log\\left|\\frac{x-a}{x-b}\\right|, \\qquad x\\neq a,b.\\]\n\nAlthough the original function vanishes outside a bounded interval, its transform extends throughout the line and develops logarithmic singularities at the endpoints. Thus compact support and boundedness need not survive the transformation. The example also shows why the principal-value interpretation is more than a technical convention: inside the interval, cancellation across the singular point is what makes the formula possible."
    },
    {
      "heading": "Fourier representation and operator identities",
      "text": "Fix the Fourier transform convention\n\n\\[\\widehat f(\\omega)=\\int_{\\mathbb R}f(x)e^{-i\\omega x}\\,dx.\\]\n\nThe defining frequency-domain identity is\n\n\\[\\widehat{Hf}(\\omega) =-i\\,\\operatorname{sgn}(\\omega)\\widehat f(\\omega).\\]\n\nThe multiplier equals \\(-i\\) at positive frequencies and \\(+i\\) at negative frequencies. Its value at zero is irrelevant for functions in \\(L^2(\\mathbb R)\\), since changing a multiplier at one point does not affect an \\(L^2\\) operator. Constants and other nondecaying objects require an additional convention or an appropriate generalized formulation.\n\nThis identity gives a particularly direct interpretation for sinusoidal signals. For \\(\\omega_0>0\\),\n\n\\[H[\\cos(\\omega_0x)]=\\sin(\\omega_0x),\\qquad H[\\sin(\\omega_0x)]=-\\cos(\\omega_0x),\\]\n\nunderstood in a generalized or periodic setting because these sinusoids are not square-integrable on the entire real line. The transform is therefore a quadrature operator, not a universal time delay. A delay multiplies frequency components by \\(e^{-i\\omega\\tau}\\), whose phase depends linearly on frequency; the Hilbert multiplier has a constant phase on each frequency half-axis.\n\nOn the Hilbert space \\(L^2(\\mathbb R)\\), the multiplier immediately yields\n\n\\[H^2=-I,\\qquad H^{-1}=-H,\\qquad \\|Hf\\|_2=\\|f\\|_2.\\]\n\nThus \\(H\\) is an isometry and a unitary operator on complex \\(L^2\\), although it is skew-adjoint rather than self-adjoint:\n\n\\[H^*=-H.\\]\n\nFor real-valued \\(f\\),\n\n\\[\\int_{\\mathbb R}f(x)Hf(x)\\,dx=0.\\]\n\nThe original function and its transform are orthogonal in the real \\(L^2\\) inner product. This does not imply pointwise independence or any general statistical independence.\n\nThe same algebra separates the frequency axis into two parts. The operators\n\n\\[P_+=\\frac{I+iH}{2},\\qquad P_-=\\frac{I-iH}{2}\\]\n\nare orthogonal projections onto the subspaces supported at positive and negative frequencies, respectively. Their multipliers are the corresponding half-axis indicators, apart from the immaterial zero-frequency point. They satisfy\n\n\\[P_++P_-=I,\\qquad P_+P_-=0.\\]\n\nThese projections explain why the Hilbert transform appears whenever a real signal must be completed into a complex signal with one-sided spectrum.\n\nTranslation invariance follows from convolution, and positive dilations commute with the transform. Differentiation also commutes with it whenever both operations are defined:\n\n\\[H(f')=(Hf)'.\\]\n\nConsequently, the Hilbert transform preserves the order of regularity in many Sobolev spaces. These structural facts are often more useful than direct evaluation of the principal-value integral, particularly in differential equations and spectral computation."
    },
    {
      "heading": "Function spaces, boundedness, and endpoint behavior",
      "text": "The natural range of strong boundedness is\n\n\\[1<p<\\infty.\\]\n\nFor each such \\(p\\), the Hilbert transform extends to a bounded operator on the L^p space:\n\n\\[\\|Hf\\|_p\\leq C_p\\|f\\|_p.\\]\n\nFor functions in these spaces, the principal-value integral exists almost everywhere and agrees with the bounded operator obtained by extension from smooth functions. This agreement connects two distinct constructions: a pointwise limiting integral and a continuous operator defined through norm completion.\n\nThe exact operator norm is known:\n\n\\[\\|H\\|_{L^p\\to L^p} = \\begin{cases} \\tan\\!\\left(\\dfrac{\\pi}{2p}\\right),&1<p\\leq2,\\\\[6pt] \\cot\\!\\left(\\dfrac{\\pi}{2p}\\right),&2\\leq p<\\infty. \\end{cases}\\]\n\nThese sharp constants, established by Stylianos Pichorides in 1972, equal \\(1\\) at \\(p=2\\) and diverge as \\(p\\) approaches either endpoint. Their behavior quantifies the loss of control near \\(L^1\\) and \\(L^\\infty\\), rather than merely indicating that a proof based on square-integrability has failed.\n\nAt \\(p=1\\), the transform is not bounded from \\(L^1\\) to itself. It does satisfy a weak-type estimate:\n\n\\[m\\{x:|Hf(x)|>\\lambda\\} \\leq \\frac{C}{\\lambda}\\|f\\|_1, \\qquad \\lambda>0,\\]\n\nwhere \\(m\\) is Lebesgue measure. This controls the size of sets on which the transform is large without requiring the transform to be integrable. For a compactly supported function with nonzero integral, the distant behavior typically includes\n\n\\[Hf(x)\\sim\\frac{1}{\\pi x}\\int_{\\mathbb R}f(t)\\,dt.\\]\n\nThe \\(1/x\\) tail gives a concrete obstruction to integrability on the full line.\n\nAt the other endpoint, a bounded input need not have a bounded transform, as the interval-indicator example demonstrates. The appropriate replacement is bounded mean oscillation, abbreviated BMO. Its seminorm measures average deviation from a local mean:\n\n\\[\\|g\\|_{\\mathrm{BMO}} = \\sup_I\\frac{1}{|I|}\\int_I|g(x)-g_I|\\,dx.\\]\n\nThe Hilbert transform maps \\(L^\\infty\\) into BMO when its action is normalized appropriately and functions differing by constants are identified. Logarithmic singularities can belong to BMO even though they are unbounded.\n\nA complementary endpoint theory uses Hardy spaces. The real-variable space \\(H^1(\\mathbb R)\\) admits a characterization through the simultaneous integrability of \\(f\\) and \\(Hf\\), with an equivalent norm\n\n\\[\\|f\\|_{H^1}\\asymp\\|f\\|_1+\\|Hf\\|_1.\\]\n\nHere \\(H^1\\) denotes a Hardy space, not the first-order Sobolev space sometimes written with the same symbol. Its cancellation requirements remove precisely the kinds of low-frequency behavior responsible for troublesome tails. Hardy spaces and BMO consequently provide a more faithful endpoint framework than ordinary \\(L^1\\) and \\(L^\\infty\\)."
    },
    {
      "heading": "Harmonic conjugates and complex boundary values",
      "text": "The Hilbert transform has a geometric interpretation in complex analysis. Let \\(f\\) be a suitable real boundary function, and extend it into the upper half-plane using the Poisson kernel:\n\n\\[u(x,y)=P_y*f(x),\\qquad P_y(x)=\\frac{1}{\\pi}\\frac{y}{x^2+y^2}, \\quad y>0.\\]\n\nThe resulting function is harmonic:\n\n\\[u_{xx}+u_{yy}=0.\\]\n\nIts conjugate harmonic extension can be written\n\n\\[v(x,y)=Q_y*f(x),\\qquad Q_y(x)=\\frac{1}{\\pi}\\frac{x}{x^2+y^2}.\\]\n\nThe pair satisfies the Cauchy–Riemann equations, so \\(F=u+iv\\) is analytic in the upper half-plane.\n\nAs \\(y\\) decreases to zero, the boundary values are\n\n\\[u(\\cdot,y)\\longrightarrow f,\\qquad v(\\cdot,y)\\longrightarrow Hf,\\]\n\nwith the precise convergence depending on the function space. For \\(1<p<\\infty\\), convergence holds in \\(L^p\\), and suitable boundary limits also hold almost everywhere. Consequently, \\(f+iHf\\) is the boundary trace of an analytic function in the upper half-plane. The additive constant in a harmonic conjugate is fixed here through the chosen extension and normalization.\n\nThis construction can also be expressed by a Cauchy integral. For sufficiently decaying \\(f\\), define\n\n\\[Cf(z)=\\frac{1}{2\\pi i} \\int_{\\mathbb R}\\frac{f(t)}{t-z}\\,dt, \\qquad \\operatorname{Im}z\\neq0.\\]\n\nThe upper and lower boundary values obey the Sokhotski–Plemelj formulas:\n\n\\[C_+f=\\frac12f+\\frac{i}{2}Hf,\\qquad C_-f=-\\frac12f+\\frac{i}{2}Hf.\\]\n\nTheir difference is \\(f\\), while their sum is \\(iHf\\). The jump across the boundary and the singular integral along it are therefore two aspects of the same analytic construction.\n\nThese identities are central to a boundary-value problem in which the real or imaginary part of an analytic function is prescribed along a boundary. Specifying the real boundary values determines the conjugate values under appropriate growth and normalization conditions. The transform is the operator that performs this reconstruction on a straight boundary. For a circle or a more general curve, the kernel changes, but the connection between boundary traces and conjugate functions remains.\n\nThis viewpoint also clarifies why sign conventions matter. Choosing the lower half-plane instead of the upper half-plane reverses the relevant conjugacy relation. Reversing the orientation of a contour or changing the Fourier exponential convention can likewise reverse signs. Identities from complex analysis and signal processing are compatible only after these choices have been matched; the names “Hilbert transform” and “harmonic conjugate” alone do not specify every convention."
    },
    {
      "heading": "Periodic and finite-interval transforms",
      "text": "For a \\(2\\pi\\)-periodic function, the natural analogue is the circular Hilbert transform:\n\n\\[(H_{\\mathbb T}f)(\\theta)= \\frac{1}{2\\pi}\\operatorname{p.v.} \\int_{-\\pi}^{\\pi} f(\\phi)\\cot\\!\\left(\\frac{\\theta-\\phi}{2}\\right)\\,d\\phi.\\]\n\nThe cotangent kernel respects periodicity and has the same leading local singularity as the real-line kernel. Near \\(\\phi=\\theta\\),\n\n\\[\\cot\\!\\left(\\frac{\\theta-\\phi}{2}\\right) \\sim\\frac{2}{\\theta-\\phi}.\\]\n\nThe factor \\(1/(2\\pi)\\) therefore gives the same local normalization \\(1/[\\pi(\\theta-\\phi)]\\).\n\nIn terms of Fourier series, its action is\n\n\\[H_{\\mathbb T}(e^{in\\theta}) =-i\\,\\operatorname{sgn}(n)e^{in\\theta}, \\qquad n\\neq0,\\]\n\nand \\(H_{\\mathbb T}(1)=0\\). Consequently,\n\n\\[H_{\\mathbb T}^2f=-f+\\overline f,\\]\n\nwhere\n\n\\[\\overline f=\\frac{1}{2\\pi}\\int_{-\\pi}^{\\pi}f(\\theta)\\,d\\theta.\\]\n\nUnlike a nonzero constant on the whole real line, a constant periodic function belongs to every finite-\\(p\\) space on the circle. Its presence therefore cannot be ignored. The circular transform is invertible with inverse \\(-H_{\\mathbb T}\\) only after restricting to mean-zero functions.\n\nThe finite Hilbert transform is a different operator:\n\n\\[(H_{(a,b)}f)(x)= \\frac{1}{\\pi}\\operatorname{p.v.} \\int_a^b\\frac{f(t)}{x-t}\\,dt, \\qquad a<x<b.\\]\n\nIt is not the periodic transform, nor does it retain the full real-line transform’s inversion identity. It can be viewed as extending \\(f\\) by zero, applying the real-line transform, and then discarding every output value outside \\((a,b)\\). That final restriction loses information, so applying the same finite operator twice generally does not recover \\(-f\\).\n\nEndpoint behavior is particularly important. On \\((-1,1)\\),\n\n\\[H_{(-1,1)} \\left[\\frac{1}{\\sqrt{1-t^2}}\\right](x)=0, \\qquad -1<x<1.\\]\n\nThe input is nonzero and belongs to \\(L^p(-1,1)\\) for \\(1\\leq p<2\\). Thus, on these spaces, the finite transform has a nontrivial null space. The inverse problem depends on the admissible endpoint singularities, the chosen function space, and any additional normalization. Results for the whole-line transform cannot simply be transferred by changing integration limits.\n\nFinite transforms arise in integral equations associated with aerodynamics, elasticity, and reconstruction problems involving restricted data. Circular transforms instead describe conjugacy on the unit disk and periodic signals. Their common local kernel explains shared singular-integral techniques, but their global geometry determines different null spaces, inversion formulas, and boundary conditions."
    },
    {
      "heading": "Analytic signals, envelopes, and instantaneous frequency",
      "text": "For a real signal \\(x(t)\\), its analytic signal is\n\n\\[z(t)=x(t)+iHx(t).\\]\n\nUnder the Fourier convention already specified,\n\n\\[\\widehat z(\\omega)= \\begin{cases} 2\\widehat x(\\omega),&\\omega>0,\\\\ 0,&\\omega<0. \\end{cases}\\]\n\nThe term “analytic” refers to its connection with analytic boundary values and its one-sided spectrum, not to the requirement that \\(x(t)\\) possess a convergent Taylor series. A constant component, when present in a periodic or generalized setting, is retained rather than doubled.\n\nWriting\n\n\\[z(t)=A(t)e^{i\\phi(t)}\\]\n\ndefines the magnitude\n\n\\[A(t)=\\sqrt{x(t)^2+[Hx(t)]^2}\\]\n\nand a phase \\(\\phi(t)\\), wherever \\(z(t)\\neq0\\). If the phase is locally differentiable, the instantaneous angular frequency is\n\n\\[\\omega_{\\mathrm{inst}}(t)=\\phi'(t) =\\operatorname{Im}\\frac{z'(t)}{z(t)}.\\]\n\nFrequency in cycles per unit time is \\(\\phi'(t)/(2\\pi)\\). Phase unwrapping selects a continuous branch across ordinary \\(2\\pi\\) jumps, but it cannot remove genuine indeterminacy at zeros of the analytic signal.\n\nFor a pure tone \\(x(t)=A_0\\cos(\\omega_0t+\\theta)\\), with \\(A_0>0\\) and \\(\\omega_0>0\\),\n\n\\[z(t)=A_0e^{i(\\omega_0t+\\theta)}.\\]\n\nThe analytic magnitude is exactly \\(A_0\\), and the instantaneous frequency is constant. Amplitude-modulated signals require stronger qualifications. A useful sufficient condition comes from Bedrosian’s theorem, published by Edward Bedrosian in 1963. In a common formulation, if the spectrum of \\(a\\) lies below a frequency threshold and that of \\(b\\) lies entirely above it, then, under appropriate integrability assumptions,\n\n\\[H(ab)=a\\,Hb.\\]\n\nFor a low-frequency envelope multiplying a sufficiently high-frequency carrier, this identity justifies extracting the modulation envelope from the analytic magnitude.\n\nWithout spectral separation, the analytic magnitude need not equal an independently intended physical amplitude. For example, the analytic signal of two positive-frequency tones is\n\n\\[z(t)=e^{i\\omega_1t}+e^{i\\omega_2t},\\]\n\nwhose magnitude is\n\n\\[|z(t)|=2\\left| \\cos\\!\\left(\\frac{\\omega_1-\\omega_2}{2}t\\right) \\right|.\\]\n\nThis beat envelope vanishes at particular times, making the phase undefined there. More generally, a positive-frequency spectrum does not guarantee nonnegative instantaneous frequency at every point. Local phase behavior can be strongly affected by interference.\n\nAnalytic signals are widely used for quadrature demodulation, envelope measurement, vibration analysis, and the study of oscillatory time series. In single-sideband modulation, combining a signal and its Hilbert transform with quadrature carriers can suppress one of the two translated sidebands. Such applications rely on spectral structure, not merely on the formal existence of \\(A(t)\\) and \\(\\phi(t)\\); a complicated multicomponent signal may have a mathematically valid analytic representation without a unique interpretation as one oscillation."
    },
    {
      "heading": "Discrete computation and filter implementation",
      "text": "A common numerical implementation uses the discrete Fourier transform of a finite sample sequence. For an even record length \\(N\\), the discrete Hilbert-transform multiplier is assigned as\n\n\\[M[k]= \\begin{cases} 0,&k=0,\\\\ -i,&1\\leq k<N/2,\\\\ 0,&k=N/2,\\\\ +i,&N/2<k<N. \\end{cases}\\]\n\nAfter multiplying the transformed samples by \\(M[k]\\), an inverse transform gives the discrete quadrature sequence. For odd \\(N\\), there is no Nyquist-frequency bin, and all nonzero bins split into positive and negative groups. The fast Fourier transform makes this computation possible in \\(O(N\\log N)\\) arithmetic operations.\n\nTo construct an analytic sequence directly, the spectral weights are instead \\(2\\) on positive-frequency bins, \\(0\\) on negative-frequency bins, and \\(1\\) at zero and, for even \\(N\\), at the Nyquist bin. The exceptional treatment of the Nyquist component follows from sampled-frequency geometry: it has no distinct negative-frequency partner. A real alternating sequence at that frequency cannot be assigned an independent quadrature component by this periodic construction.\n\nThe ideal infinite discrete-time Hilbert transformer has frequency response\n\n\\[H_d(e^{i\\omega})=-i\\,\\operatorname{sgn}(\\omega), \\qquad -\\pi<\\omega<\\pi,\\]\n\nand impulse response\n\n\\[h[n]= \\begin{cases} \\dfrac{2}{\\pi n},&n\\text{ odd},\\\\[5pt] 0,&n\\text{ even}, \\end{cases}\\]\n\nwith \\(h[0]=0\\). This response extends indefinitely in both time directions. Practical finite-impulse-response approximations truncate or optimize an antisymmetric coefficient sequence and add delay to make the filter causal. They achieve an approximate quadrature relationship over a specified frequency band rather than an exact ideal response everywhere.\n\nAn FFT calculation implicitly treats the record as periodic. If its endpoints do not join smoothly, the artificial discontinuity affects the transformed signal throughout the record. Zero padding changes the assumed continuation and can reduce wraparound interaction, but it does not reconstruct unknown data outside the observation window. Tapering reduces boundary discontinuities while also modifying amplitudes and spectral resolution. These effects arise from the transform’s nonlocality, not simply from floating-point arithmetic.\n\nDirect principal-value quadrature is useful when samples are irregularly spaced or when a finite-interval integral is the intended operator. A standard technique subtracts the singular value:\n\n\\[\\operatorname{p.v.}\\int_a^b\\frac{f(t)}{x-t}\\,dt = \\int_a^b\\frac{f(t)-f(x)}{x-t}\\,dt + f(x)\\log\\left|\\frac{x-a}{x-b}\\right|.\\]\n\nFor differentiable \\(f\\), the first integrand has a removable singularity. This formula permits ordinary quadrature after treating the removable value correctly and evaluating the logarithmic term explicitly. Numerical design must therefore begin by distinguishing three different tasks: a periodic transform, a whole-line transform inferred from a finite record, and a genuinely finite-interval transform."
    },
    {
      "heading": "Causality and dispersion relations",
      "text": "The Hilbert transform relates the real and imaginary parts of many physical response functions because causality imposes analyticity. Consider a linear time-invariant response\n\n\\[y(t)=\\int_{\\mathbb R}g(\\tau)x(t-\\tau)\\,d\\tau,\\]\n\nwith \\(g(\\tau)=0\\) for \\(\\tau<0\\). Its frequency response, using the convention \\(e^{-i\\omega t}\\), is\n\n\\[\\chi(\\omega)=\\int_0^\\infty g(t)e^{-i\\omega t}\\,dt.\\]\n\nUnder suitable integrability assumptions, this formula extends analytically into the lower half of the complex frequency plane. Exponential damping there suppresses contributions from large positive times.\n\nIf the response decays sufficiently at large frequency, analyticity yields the Kramers–Kronig relations. With the present Fourier and Hilbert-transform conventions, they take the form\n\n\\[\\operatorname{Re}\\chi =H(\\operatorname{Im}\\chi),\\qquad \\operatorname{Im}\\chi =-H(\\operatorname{Re}\\chi).\\]\n\nFor example,\n\n\\[\\operatorname{Re}\\chi(\\omega) = \\frac{1}{\\pi}\\operatorname{p.v.} \\int_{\\mathbb R} \\frac{\\operatorname{Im}\\chi(\\omega')} {\\omega-\\omega'}\\,d\\omega'.\\]\n\nUsing the opposite sign in the Fourier exponential changes the analytic half-plane and reverses the corresponding signs. The underlying causal constraint is unchanged.\n\nA simple check is the causal exponential response\n\n\\[g(t)=e^{-at}\\mathbf 1_{[0,\\infty)}(t),\\qquad a>0.\\]\n\nIts transform is\n\n\\[\\chi(\\omega)=\\frac{1}{a+i\\omega} =\\frac{a}{a^2+\\omega^2} -i\\frac{\\omega}{a^2+\\omega^2}.\\]\n\nSince\n\n\\[H\\left[\\frac{a}{a^2+\\omega^2}\\right] =\\frac{\\omega}{a^2+\\omega^2},\\]\n\nthe imaginary part is indeed minus the Hilbert transform of the real part. This example displays the relation without requiring any resonance approximation.\n\nDispersion relations are important in optics, dielectric response, electrical impedance, and spectroscopy. They connect two components of a complex response that are often associated with dispersion and absorption. Causality alone does not imply every physical positivity condition: passivity is an additional assumption, and the exact identification of absorptive behavior depends on the response quantity and sign convention. If a response tends to a nonzero constant or grows at infinity, an unsubtracted integral may fail; constants must be separated or subtracted dispersion relations used.\n\nA related connection joins magnitude and phase for minimum-phase systems. If a transfer function has the appropriate zero-free analytic continuation and growth behavior, an analytic logarithm can be formed. Its real part is the logarithm of the magnitude, and its imaginary part is the phase, so another Hilbert-transform relation follows. Magnitude alone does not determine phase for arbitrary systems: pure delays and all-pass factors can alter phase without changing magnitude. Moreover, dispersion integrals extend over the entire frequency axis, so finite-band measurements require additional assumptions about unmeasured frequencies."
    },
    {
      "heading": "Generalizations and the development of harmonic analysis",
      "text": "The Hilbert transform emerged from the study of conjugate functions, Fourier series, and analytic boundary values around the beginning of the twentieth century. Marcel Riesz’s work in the 1920s established the \\(L^p\\) boundedness of the conjugate-function operator for \\(1<p<\\infty\\). This result showed that the operator belongs naturally to a full scale of integrability spaces, rather than only to the square-integrable setting where its Fourier multiplier is easiest to understand. Later singular-integral theory generalized the role of its odd kernel and cancellation.\n\nIn \\(\\mathbb R^n\\), the principal analogues are the Riesz transforms:\n\n\\[\\widehat{R_jf}(\\xi) =-i\\frac{\\xi_j}{|\\xi|}\\widehat f(\\xi), \\qquad j=1,\\ldots,n.\\]\n\nTheir kernels have the form\n\n\\[c_n\\,\\operatorname{p.v.}\\frac{x_j}{|x|^{n+1}},\\]\n\nwhere \\(c_n\\) is chosen to produce the stated multiplier. For \\(n=1\\), the multiplier reduces to \\(-i\\operatorname{sgn}\\xi\\), recovering the Hilbert transform. In higher dimensions,\n\n\\[\\sum_{j=1}^nR_j^2=-I\\]\n\non \\(L^2\\). Riesz transforms connect harmonic extensions, gradients, and the Laplacian, making them central to estimates for partial differential equations.\n\nThe Hilbert transform is also the simplest model in Calderón–Zygmund theory. Its kernel has the characteristic size bound \\(|K(x)|\\lesssim |x|^{-1}\\), regularity away from the origin, and odd cancellation. In higher dimensions and on more general spaces, analogous conditions support boundedness theorems even when there is no simple Fourier multiplier. This extension is crucial for variable-coefficient equations and singular integrals along curves, where translation invariance is absent.\n\nWeighted inequalities reveal a finer aspect of the operator. For \\(1<p<\\infty\\), the Hilbert transform is bounded on \\(L^p(w\\,dx)\\) precisely when the positive weight \\(w\\) belongs to the Muckenhoupt class \\(A_p\\):\n\n\\[[w]_{A_p} = \\sup_I \\left(\\frac{1}{|I|}\\int_Iw(x)\\,dx\\right) \\left(\\frac{1}{|I|}\\int_Iw(x)^{-1/(p-1)}\\,dx\\right)^{p-1} <\\infty.\\]\n\nFor example, \\(w(x)=|x|^\\alpha\\) satisfies this condition exactly when\n\n\\[-1<\\alpha<p-1.\\]\n\nThe condition balances concentrations of the weight against concentrations of its reciprocal power.\n\nAnother major extension studies the commutator\n\n\\[[b,H]f=b\\,Hf-H(bf).\\]\n\nThis measures the failure of multiplication by \\(b\\) to commute with singular integration. Under the standard locally integrable formulation, boundedness of this commutator on \\(L^p\\), for \\(1<p<\\infty\\), characterizes BMO functions. The result turns the Hilbert transform into a test of oscillation: its interaction with a coefficient detects a regularity property that ordinary pointwise bounds do not capture. Weighted estimates, commutators, and transforms on irregular boundaries remain active parts of modern harmonic analysis."
    }
  ],
  "links": [
    "harmonic-analysis",
    "david-hilbert",
    "convolution",
    "indicator-function",
    "fourier-transform",
    "hilbert-space",
    "isometry",
    "orthogonal-projection",
    "sobolev-space",
    "lp-space",
    "almost-everywhere",
    "lebesgue-measure",
    "complex-analysis",
    "harmonic-function",
    "cauchy-riemann-equations",
    "analytic-function",
    "boundary-value-problem",
    "fourier-series",
    "time-series",
    "discrete-fourier-transform",
    "fast-fourier-transform",
    "floating-point-arithmetic",
    "spectroscopy",
    "laplace-operator",
    "partial-differential-equation"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}