Fact-checked by Grok 2 weeks ago

Angular spectrum method

The angular spectrum method is a fundamental technique in scalar for propagating a monochromatic scalar field from an initial (typically at z=0) to a parallel observation plane (at z>0) by decomposing the field into a continuous spectrum of plane via a two-dimensional . This representation, also known as the angular spectrum representation, expresses the field as an over spatial frequencies k_x and k_y, where each component propagates with a z-directed k_z = \sqrt{k^2 - k_x^2 - k_y^2} (with k = 2\pi / \lambda the in the medium), enabling exact computation of effects under the . The method distinguishes between propagating (where k_z is real) and evanescent (where k_z is imaginary, leading to ), providing a unified framework for both near- and far-field . Introduced in the context of Fourier optics by Joseph W. Goodman in his 1968 textbook Introduction to Fourier Optics, the angular spectrum method builds on earlier diffraction theories like those of Kirchhoff and Rayleigh-Sommerfeld, offering a frequency-domain alternative that aligns with linear systems analysis. The propagation kernel derives from the Helmholtz equation for the transverse field component, yielding the transfer function H(k_x, k_y, z) = \exp(i k_z z) for propagating components, which is applied after Fourier transformation of the input field U(x, y, 0). In numerical implementations, fast Fourier transforms (FFTs) facilitate efficient computation on discrete grids, though care must be taken to handle evanescent waves and avoid aliasing via techniques like zero-padding or low-pass filtering. The method's versatility extends beyond to acoustics and other wave phenomena, with applications in simulating patterns, holographic , optical device design (such as photonic circuits and fiber delay lines), and modeling near-field effects like . Recent advances, including band-extended and scaled variants, enhance accuracy for long-distance propagation and off-axis calculations, addressing limitations in traditional Fresnel or Fraunhofer approximations. Its computational efficiency and physical insight into wave decomposition have made it a cornerstone for both theoretical analysis and practical simulations in modern .

Fundamentals

Definition and overview

The angular spectrum method is a technique for modeling the propagation of scalar wave fields, such as or , by decomposing the field into a continuous of plane waves with different propagation angles. This approach allows for the precise description of wave evolution in free space or homogeneous media without approximations beyond the scalar . Physically, the method represents the wave field at an initial plane as a superposition of obliquely propagating plane waves, where each component is characterized by spatial frequencies that correspond to specific angles relative to direction. These plane waves propagate independently according to their wave vectors, and their recombination at a subsequent plane yields the propagated field. The underlying mathematical tool for this decomposition is the , which maps the to an angular spectrum. The method originated in the mid-20th century amid developments in Fourier optics and diffraction theory, with foundational work by J. A. Ratcliffe in 1956 applying it to ionospheric propagation problems. Key contributions to its use in optical diffraction were made by Joseph W. Goodman in 1968, who integrated it into the framework of Fourier optics for analyzing imaging and coherence. It offers general advantages as an exact solution for both paraxial and non-paraxial propagation in homogeneous media, while enabling computationally efficient numerical simulations through fast Fourier transform algorithms.

Relation to Fourier optics

The angular spectrum method is fundamentally grounded in the principles of , which conceptualizes as a linear filtering process within the domain. In this framework, the optical field distribution across a plane is represented by its two-dimensional , where each component corresponds to a propagating at an angle determined by the wave vector components k_x and k_y. This decomposition allows propagation to be treated as a multiplicative operation on these frequency components, aligning directly with ' emphasis on frequency-domain analysis for understanding wave behavior. A key connection arises through the , which underpins much of . Free-space propagation of the field can be expressed as a spatial between the initial field and the function, or , that describes how a spreads over distance z. In the angular spectrum domain, this transforms into a straightforward by the \exp(i k_z z), where k_z = \sqrt{k^2 - k_x^2 - k_y^2} and k = 2\pi / \lambda is the wave number. This duality enables efficient computation and conceptual clarity, as the method leverages the theorem to model without direct spatial-domain integration. Spatial frequency, defined as cycles per unit length (with dimensions of inverse length), serves as a prerequisite concept linking the method to propagation angles via the transverse wave vector components: k_x = k \sin\theta_x and k_y = k \sin\theta_y, where \theta_x and \theta_y are the angles relative to the optical axis. This mapping highlights how higher spatial frequencies correspond to steeper propagation angles. Unlike conventional Fourier optics, which typically assumes the paraxial approximation for small angles (where \sin\theta \approx \theta and evanescent contributions are neglected), the angular spectrum method provides a more general formulation by including all spatial frequencies. For components where k_x^2 + k_y^2 > k^2, k_z becomes imaginary, resulting in evanescent waves that decay exponentially along the propagation direction rather than oscillating. This inclusion extends the applicability to non-paraxial scenarios, such as near-field effects or high-numerical-aperture systems, while retaining the core frequency-domain insights of .

Theoretical formulation

Angular spectrum representation

The angular spectrum representation provides a fundamental decomposition of the scalar wave field into a superposition of plane waves, each characterized by its direction of propagation. In this formulation, the complex scalar field u(x, y, 0) at a reference plane z = 0 is expressed as the two-dimensional inverse Fourier transform of the angular spectrum A(f_x, f_y), where f_x and f_y denote spatial frequencies in the x- and y-directions, respectively: u(x, y, 0) = \iint_{-\infty}^{\infty} A(f_x, f_y) \exp\left[i 2\pi (f_x x + f_y y)\right] \, df_x \, df_y. This representation originates from the plane wave expansion of the field and is a cornerstone of Fourier optics. The angular spectrum A(f_x, f_y) itself is defined as the two-dimensional Fourier transform of the field at z = 0: A(f_x, f_y) = \iint_{-\infty}^{\infty} u(x, y, 0) \exp\left[-i 2\pi (f_x x + f_y y)\right] \, dx \, dy. The spatial frequencies f_x and f_y correspond to the directional components of the wave vectors, related to the angles \theta_x and \theta_y by \sin \theta_x \approx \lambda f_x and \sin \theta_y \approx \lambda f_y under the for small , where \lambda is the . The components of the angular spectrum can be interpreted in terms of the longitudinal wave number k_z = \sqrt{(2\pi / \lambda)^2 - (2\pi f_x)^2 - (2\pi f_y)^2}. For propagating waves, where the transverse spatial frequency magnitude \sqrt{f_x^2 + f_y^2} < 1 / \lambda, k_z is real, corresponding to plane waves that carry energy away from the reference plane without decay. In contrast, evanescent waves arise when \sqrt{f_x^2 + f_y^2} > 1 / \lambda, making k_z imaginary; these components decay exponentially with distance from the plane and are associated with near-field effects, such as those in total internal reflection or subwavelength imaging scenarios. This representation assumes the field is defined over an infinite transverse plane at z = 0, ensuring the integrals converge and the decomposition is complete. For practical scenarios with finite apertures, such as limited illumination or observation areas, edge effects introduce artifacts like Gibbs ringing in the spectrum, requiring careful handling in applications. The angular spectrum at subsequent planes can then be obtained by multiplying A(f_x, f_y) by a propagation phase factor, enabling the modeling of free-space evolution.

Free-space propagation

The angular spectrum method provides an exact analytical for describing the of scalar electromagnetic fields in homogeneous, free-space , governed by the . The field distribution u(x, y, [z](/page/Z)) at a longitudinal [z](/page/Z) from the initial plane (z = 0) evolves through the of its angular spectrum components, representing a superposition of plane waves with varying transverse wavevectors. This approach treats as a linear filtering operation in the domain, where each spectral component advances with its corresponding longitudinal phase factor derived from the of waves in free space. The propagation transfer function H(f_x, f_y, z) multiplies the initial angular spectrum A(f_x, f_y, 0) to yield the spectrum at distance z: A(f_x, f_y, z) = A(f_x, f_y, 0) \exp\left[i k z \sqrt{1 - \lambda^2 (f_x^2 + f_y^2)}\right], where k = 2\pi / \lambda is the , \lambda is the , and f_x, f_y are the spatial frequencies in the x and y directions, respectively. This expression arises from the plane-wave decomposition, where the longitudinal component of the wavevector k_z = k \sqrt{1 - \lambda^2 (f_x^2 + f_y^2)} ensures satisfaction of the wave equation (\nabla^2 + k^2) u = 0. For propagating waves, where \lambda^2 (f_x^2 + f_y^2) < 1, k_z is real, leading to oscillatory phase advancement. The field at the propagated plane is reconstructed via the inverse two-dimensional Fourier transform of the evolved spectrum: u(x, y, z) = \iint_{-\infty}^{\infty} A(f_x, f_y, z) \exp\left[i 2\pi (f_x x + f_y y)\right] \, df_x \, df_y. This integral, often termed the angular spectrum propagator, sums the contributions of all plane waves, each tilted at angles \theta_x \approx \lambda f_x and \theta_y \approx \lambda f_y for small angles, to form the diffracted field. Equivalently, the propagation can be expressed directly in the spatial domain as a convolution with the propagator kernel, though the frequency-domain formulation is computationally preferable for its separability. Evanescent waves arise when \lambda^2 (f_x^2 + f_y^2) > 1, making the argument of the negative; in this regime, k_z = -i \kappa with \kappa = k \sqrt{\lambda^2 (f_x^2 + f_y^2) - 1} > 0, transforming the to \exp[-\kappa z] for forward (z > 0). These components, corresponding to high spatial frequencies or supercritical angles, decay exponentially away from the initial plane and do not contribute to far-field , effectively filtering near-field details. This behavior is crucial for understanding limits in systems. The method is rigorously valid for any propagation distance z in non-absorbing, homogeneous media, providing an exact solution to the scalar without approximations on the field or distance. In the paraxial of small angles (\lambda^2 (f_x^2 + f_y^2) \ll 1), the approximates to \sqrt{1 - \xi} \approx 1 - \xi/2 with \xi = \lambda^2 (f_x^2 + f_y^2), reducing the to the quadratic of the integral and recovering the paraxial approximation. This exactness distinguishes the angular spectrum approach from approximate methods like or , enabling accurate modeling across near-, intermediate-, and far-field regimes.

Numerical implementation

Discrete Fourier transform approach

The discrete Fourier transform (DFT) approach provides a straightforward for implementing the angular spectrum propagation in free space, leveraging (FFT) algorithms for efficiency. This technique discretizes the continuous angular spectrum representation, where the input complex field u(x, y, 0) is sampled on a uniform 2D grid of size N \times N with spatial sampling intervals \Delta x and \Delta y (often equal for isotropic cases). The propagation to a distance z follows a three-step . First, the 2D DFT of the discretized input field yields the discrete angular spectrum A(m, n), where m, n = 0, 1, \dots, N-1. In the second step, the discrete angular is multiplied element-wise by the H(m, n) = \exp\left[ i k z \sqrt{1 - \lambda^2 (f_x(m)^2 + f_y(n)^2)} \right], with wave number k = 2\pi / \lambda, spatial frequencies f_x(m) = m / (N \Delta x) and f_y(n) = n / (N \Delta y), and \lambda the ; this discrete form approximates the continuous free-space . The resulting is then transformed back via a 2D inverse DFT to obtain the output field u(x, y, z) on the same grid. Appropriate sampling is critical to prevent and ensure accurate representation of diffracted fields. The spatial sampling must satisfy the \Delta x \leq \lambda / 2 to capture the highest spatial frequencies in the input field. Additionally, to avoid in the propagated field due to the nonlinear phase in the , the sampling interval should fulfill \Delta x \leq \lambda z / (N \Delta x), where N \Delta x is the total grid extent; the corresponding frequency sampling is \Delta f_x = 1 / (N \Delta x). Zero-padding techniques enhance the implementation by extending the grid size beyond the input , typically by appending zeros to reach a larger N_{\text{pad}} > N. This increases the effective , improves output resolution by effectively reducing \Delta x in the reconstructed plane, and mitigates finite-grid artifacts such as wrap-around errors from periodic DFT assumptions. Padding factors of 2–4 are common, balancing accuracy and computational cost. In practice, this DFT-based method is routinely implemented in software environments like , utilizing built-in functions such as fft2 for the forward transform and ifft2 for the inverse, or in via NumPy's np.fft.fft2 and np.fft.ifft2. Both forward (z > 0) and backward (z < 0) propagation are handled by adjusting the sign in the exponent; for backward cases, conjugating the phase term equivalently reverses the direction while preserving .

Computational efficiency and scalability

The numerical implementation of the angular spectrum method () relies on the (FFT) for efficient computation, achieving a of O(N^2 \log N) per propagation step for an N \times N grid, primarily due to the two-dimensional FFT and inverse FFT operations. Memory requirements scale as O(N^2), accommodating the field data at both source and destination planes. These scalings enable rapid simulations for moderate grid sizes but pose challenges for high-resolution or large-scale problems, where increasing N quadratically amplifies both time and storage demands. Scalability limitations arise from the fixed sampling grid in traditional , which constrains zoom factors and adjustments between planes, often necessitating zero-padding to avoid and thereby inflating computational overhead. Near the source plane (z = 0), accurate representation of evanescent waves—inhomogeneous plane waves with imaginary longitudinal wavenumbers that convey subwavelength details—demands fine spatial sampling to resolve high spatial frequencies, as coarser grids lead to errors and loss of near-field fidelity. For distant (large z), the fixed pitch results in oversized grids to maintain the , exacerbating inefficiency without adaptive resizing. A notable advance is the scalable angular spectrum (SAS) method, introduced in 2023, which enables variable pixel pitches between source and destination planes through pre-compensation in the angular spectrum domain, allowing zoom capabilities while preserving accuracy in homogeneous media. By avoiding extensive padding and supporting scalable grids (e.g., destination pitch \Delta_d = \lambda z^2 N / L^2, where L is the source extent), SAS reduces computational demands by factors up to 50 for large z distances compared to padded traditional ASM, with complexity still tied to three FFTs but at reduced effective grid sizes. Further optimizations enhance ASM's practicality. GPU acceleration, particularly for variants like the scaled ASM (Sc-ASM), leverages to simulate large-scale propagation, yielding significant speedups for high-numerical-aperture scenarios and enabling or near- computations in applications such as coherent combining. For non-uniform media, spectrum slicing divides the propagation volume into homogeneous slabs, applying sequentially with interface corrections, which maintains efficiency while handling variations without full-volume resampling. Hybrid approaches combining with -based methods, such as the Rayleigh-Sommerfeld , improve overall efficiency by using frequency-domain for bulk propagation and spatial-domain for boundary layers, achieving faster and more accurate results than pure in layered media. As of 2025, further developments include a universal least-sampling spectrum method for modeling between arbitrary non-parallel planes, enabling efficient off-axis calculations, and an improved based on for enhanced reconstruction accuracy in cellular holograms.

Applications

Diffraction and beam propagation

The angular spectrum method enables accurate simulation of diffraction patterns from apertures by decomposing the input field into its angular spectrum of plane waves, which includes both propagating components (for far-field effects) and evanescent components (for near-field details). For a single slit or circular aperture, the method propagates this spectrum to compute the diffracted field at any distance, capturing phenomena like Fresnel zones and the Poisson spot in the near field, where traditional approximations may fail. This approach satisfies the Helmholtz equation exactly in homogeneous media, allowing the inclusion of subwavelength features through evanescent waves that decay exponentially beyond the critical angle. In simulations, the method models the evolution of Gaussian beams by applying the kernel to their , revealing effects such as spreading due to and the conservation of in vortex beams during free-space travel. For Gaussian beams, the method computes the waist expansion and curvature radius variation along the axis, providing a non-paraxial description that extends beyond the standard paraxial formula. In vortex beams, like Laguerre-Gaussian modes, the topological charge is largely preserved, with minor distortions only under high-intensity conditions, as the helical structure is maintained in the decomposition. Self-focusing in can be simulated iteratively by alternating steps with nonlinear updates, enabling studies of collapse or filamentation. Representative examples illustrate the method's utility: in Fraunhofer diffraction through a circular , it yields the Airy disk pattern, where the central spot's radius is approximately $0.61 \lambda / \mathrm{NA}, with approximately 84% of the energy contained within the (central spot up to the first dark ring), directly from the inverse of the aperture function. For near-field evanescent effects, the method simulates subwavelength imaging by retaining high-spatial-frequency evanescent terms, allowing resolution of features smaller than the , such as in aperture-limited setups where propagating waves alone would blur details. Numerical implementation typically employs the for efficient computation of these spectra. Compared to the Rayleigh-Sommerfeld diffraction integral, the angular spectrum method offers superior computational efficiency, reducing calculation times by factors of 20 or more through fast Fourier transform-based propagation, while maintaining high accuracy for both near- and far-field regimes without singularities in the . It excels in handling non-paraxial beams, where the Rayleigh-Sommerfeld approach becomes more intensive due to direct spatial-domain integration.

Digital holography

The angular spectrum method (ASM) plays a central role in computing computer-generated holograms (CGHs) for digital holography by enabling accurate propagation of object fields from a virtual 3D scene to the hologram plane. In this process, the complex amplitude of the object wave is decomposed into its angular spectrum via a Fourier transform, propagated through free space using a transfer function that accounts for evanescent and propagating waves, and then inverse-transformed to obtain the field at the hologram plane. A reference wave, typically a plane wave tilted for off-axis configuration, is added to this propagated object field to form the interference pattern, which is quantized and encoded as a phase-only or amplitude hologram for display on spatial light modulators (SLMs). This approach ensures high-fidelity wavefront reconstruction, particularly for multilayered 3D objects, by avoiding paraxial approximations inherent in simpler methods. Reconstruction of digital holograms using ASM simulates optical playback by applying inverse propagation to the recorded interference pattern, yielding the original object wavefront while suppressing artifacts like the twin image. In off-axis setups, the reference wave's tilt spatially separates the zero-order, real, and virtual (twin) images in the Fourier domain, allowing selective filtering to isolate the desired virtual image before inverse ASM propagation restores the 3D field distribution. This numerical twin-image suppression is essential for quantitative phase imaging, as it mitigates conjugate noise that would otherwise overlap with the reconstructed object, enabling clear visualization of amplitude and phase details in applications such as microscopy. The method's accuracy stems from its rigorous handling of diffraction over various distances, making it suitable for both in-line and off-axis holograms recorded with CCD sensors. Specific techniques leveraging in include hybrids with for optimized performance across propagation distances, particularly in short-range scenarios common to SLM-based systems. For short distances where the Fresnel number is high, pure provides superior accuracy over Fresnel methods by correctly modeling evanescent waves, but hybrids switch between for near-field precision and Fresnel for far-field efficiency to reduce computational load without significant aliasing. These hybrids are widely used in dynamic on SLMs, where updates of CGHs enable interactive displays by rapidly recomputing and modulating patterns for applications like holographic video. Recent advances in ASM-optimized algorithms have enabled real-time simulations of 4K-resolution holograms, achieving frame rates exceeding 30 Hz for high-definition imaging through techniques like pre-computed angular spectra and neural operator enhancements that accelerate propagation calculations. These optimizations reduce computation time from seconds to milliseconds per frame, facilitating seamless integration in (AR) and (VR) near-eye displays where wide viewing angles and focus cues are critical for immersion. In microscopy, ASM-based supports quantitative phase-contrast of biological samples, providing sub-wavelength for non-invasive tracking of cellular dynamics without .

Extensions and variants

Vectorial angular spectrum method

The vectorial angular spectrum method generalizes the scalar angular spectrum representation to fully account for the vector nature of electromagnetic fields, enabling accurate modeling of effects during free-space . Building on the scalar free-space framework, it decomposes the into transverse electric () and transverse magnetic (TM) components in the angular spectrum domain, using an intrinsic defined by unit vectors perpendicular and parallel to the . This decomposition is expressed as \mathbf{E}(\mathbf{R}) = \mathbf{E}^{\mathrm{TM}}(\mathbf{R}) + \mathbf{E}^{\mathrm{TE}}(\mathbf{R}), where the TM component includes both transverse and longitudinal contributions, while the TE component is purely transverse. Propagation of the spectral amplitudes follows the \exp(i k_z z), with k_z = k \sqrt{1 - \lambda^2 (f_x^2 + f_y^2)}, where k = 2\pi / \lambda is the wave number and f_x, f_y are spatial frequencies. For the transverse components, the evolution is given by A_x(f_x, f_y, z) = A_x(f_x, f_y, 0) \exp\left[i k z \sqrt{1 - \lambda^2 (f_x^2 + f_y^2)}\right] and similarly for A_y, often handled via a 2×2 that relates the input and output transverse field spectra while preserving the vectorial structure. is managed through basis transformations from Cartesian to intrinsic coordinates (s, p), which introduce cross-coupling terms for non-normal incidence angles; specifically, the TM mode carries a longitudinal component proportional to k_t / k (with k_t = \sqrt{k_x^2 + k_y^2}), and evanescent waves (where k_z is imaginary) sustain nonzero longitudinal fields essential for near-field effects. Unlike the scalar method, which neglects and assumes isotropic fields suitable only for paraxial approximations, the vectorial approach captures and field ellipticity, particularly critical for tight focusing with high (NA) objectives where longitudinal components can become significant and even dominate the transverse intensity for certain polarizations, such as radial polarization. It also properly treats evanescent components in surface waves, avoiding artifacts in scalar models. These features make it indispensable for applications such as modeling focused vector beams, like radially or azimuthally polarized vortex beams used in optical trapping, where vectorial effects enhance axial resolution. Additionally, it facilitates simulations of plasmonic structures, where evanescent fields excite surface plasmons, enabling precise prediction of near-field enhancements in nanostructures.

Propagation in inhomogeneous media

The angular spectrum method (ASM) can be extended to inhomogeneous media by dividing the propagation path into homogeneous slabs of constant , where standard free-space ASM is applied within each slab, and interface conditions are enforced at boundaries to account for and . At each interface, governs the angular components of the plane waves in the spectrum, ensuring continuity of the tangential field components, while Fresnel coefficients determine the amplitude transmission and reflection for each spectral component. Phase accumulation within a slab of thickness \Delta z and refractive index n is incorporated via a propagation kernel \exp\left(i k n \Delta z \sqrt{1 - \left( \frac{\lambda f_x}{n} \right)^2 - \left( \frac{\lambda f_y}{n} \right)^2 }\right), where k = 2\pi / \lambda is the in vacuum, \lambda is the , and f_x, f_y are spatial frequencies. This layered approach accurately models effects such as and , particularly when the slab thicknesses are on the order of the or larger. Integration with the (TMM) enhances the for abrupt changes in multilayer structures by treating the angular spectrum as a superposition of plane waves and applying TMM to each individual component. The TMM propagates the field amplitudes across layers using a that incorporates shifts \exp(i k_z d) (with k_z the z-component of the wavevector and d the layer thickness) and interface matching via boundary conditions derived from . For gradual variations in n(z), the medium is approximated as a series of thin slices, where propagation in each slice uses \exp(i k n(z) \Delta z) to capture local evolution, effectively combining ASM's with TMM's handling of discontinuities. This hybrid method is particularly effective for structures like dielectric stacks or metamaterials, enabling computation of both transmitted and reflected spectra. The approach relies on approximations valid for slowly varying refractive index profiles, where the index modulation \delta n \ll \bar{n} (the average index) and the holds, allowing separation of rapid phase oscillations from slow amplitude changes. For rapid variations, such as in photonic crystals, these methods introduce cumulative errors from slicing, necessitating full-wave solutions like finite-difference time-domain methods for accuracy. Polarized light in such layers can incorporate vectorial effects briefly via TE/TM decompositions in the TMM. Examples include propagation through thin-film coatings, where the split-step ASM simulates wavefront distortion and energy loss across air-film-substrate interfaces, enabling characterization of optical properties like thickness and index via inverse fitting of measured intensities. In graded-index lenses, the medium is sliced into layers with linearly varying n(z), capturing beam focusing and refraction without ray-tracing approximations, as demonstrated in simulations of parabolic index profiles for optical beam delivery. These applications highlight the method's utility in modeling total internal reflection at high-angle components, where evanescent waves decay in lower-index layers.

References

  1. [1]
    6.3: Angular Spectrum Method - Physics LibreTexts
    Sep 16, 2022 · One way to see how light propagates from one plane to another is by using the angular spectrum method. We decompose the field in plane waves ...
  2. [2]
    The Angular Spectrum Method - Wiley Online Library
    Aug 7, 2015 · The approach presented is called the angular spectrum method for description of the diffraction problem. The chapter starts with the Helmholtz ...
  3. [3]
    Angular Spectrum Approach | Background Information | FOCUS
    The Angular Spectrum Approach (ASA) was first used in optics to propagate fields from a source plane to a destination plane.
  4. [4]
    Angular Spectrum Methods in Wave Propagation and Diffraction
    Angular spectrum methods constitute a powerful framework for simulating wave propagation and diffraction phenomena in optical systems.<|control11|><|separator|>
  5. [5]
    Introduction to Fourier Optics - Joseph W. Goodman - Google Books
    The second edition of this respected text considerably expands the original and reflects the tremendous advances made in the discipline since 1968.Missing: method | Show results with:method
  6. [6]
    [PDF] Chapter 7 Angular Spectrum Representation
    In this section we will derive the important result that Fourier Optics and Geomet- rical Optics naturally emerge from the angular spectrum representation.
  7. [7]
  8. [8]
    [PDF] the concept of an angular spectrum of plane waves, and its relation ...
    An angular spectrum of plane waves replaces the polar diagram, especially when the aperture is not limited in width, and is always applicable.
  9. [9]
    Numerical calculation of near field scalar diffraction using angular ...
    Feb 9, 2010 · In this work, a numerical implementation based on angular spectrum theory for near field diffraction is presented. The method uses Fast Fourier ...Missing: discrete | Show results with:discrete
  10. [10]
  11. [11]
  12. [12]
    [PDF] Numerical calculation of near field scalar diffraction using angular ...
    In this work, a numerical implementation based on angular spectrum theory for near field diffraction is presented. The method uses Fast Fourier Transforms (FFT) ...Missing: DFT | Show results with:DFT
  13. [13]
    Angular spectrum representation for the propagation of arbitrary ...
    An angular spectrum representation is applied for a description of statistical properties of arbitrary beamlike fields propagating through atmospheric ...
  14. [14]
    Evolution of orbital angular momentum spectrum of broadband ...
    Jan 2, 2023 · In this study, we numerically simulate the evolution of the orbital angular momentum (OAM) spectrum of a vortex laser beam in the optical parametric chirped ...
  15. [15]
    Extension of the angular spectrum method to calculate pressure ...
    Nov 16, 2011 · The angular spectrum method is an accurate and computationally efficient method for modeling acoustic wave propagation. The use of the ...
  16. [16]
    6.4: Rayleigh-Sommerfeld Diffraction Integral - Physics LibreTexts
    Sep 16, 2022 · The angular spectrum method amounts to a multiplication by exp ⁡ ( i ⁢ z ⁢ k z ) in Fourier space, while the Rayleigh-Sommerfeld integral is a ...
  17. [17]
    Review of computer-generated hologram algorithms for color ...
    Jul 26, 2022 · The global angular spectrum of an arbitrary polygon in the hologram plane is calculated by using the analytical spectrum of the primitive ...
  18. [18]
  19. [19]
  20. [20]
    Spatial filtering for zero-order and twin-image elimination in digital ...
    Off-axis holograms recorded with a CCD camera are numerically reconstructed with a calculation of scalar diffraction in the Fresnel approximation.<|control11|><|separator|>
  21. [21]
    Revisit to comparison of numerical reconstruction of digital ...
    Dec 4, 2019 · In digital holography, two most frequently used numerical reconstruction methods are angular spectrum method (ASM) and Fresnel diffraction ...
  22. [22]
  23. [23]
    Propagation-adaptive 4K computer-generated holography using ...
    Aug 20, 2025 · Computer-generated holography (CGH) enables photorealistic projection via numerical diffraction calculation, and has become a prominent method ...
  24. [24]
    DSCCNet for high-quality 4K computer-generated holograms
    These methods enable rapid hologram generation for real-time applications. Since the deep-learning-generated holography was proposed in 2018 [19], it has been ...
  25. [25]
    Quantitative phase-contrast microscopy by angular spectrum digital ...
    Techniques of digital holography are improved in order to obtain high-resolution, high-fidelity images of quantitative phase-contrast microscopy.
  26. [26]
    Propagation of vectorial laser beams
    **Authors, Title, Year, Journal:**
  27. [27]
    Modified vectorial angular spectrum formula for propagation of non ...
    The plane wave angular spectrum approach presents an effective integral technique to calculate propagation of electromagnetic waves or laser beams. For ...
  28. [28]
    Validation of vectorial theories for the focusing of high numerical ...
    Dec 15, 2018 · At present, there are two main vectorial diffraction methods for calculating the focusing properties of a high-NA FZP, namely vectorial angular ...
  29. [29]
    Propagation of focused scalar and vector vortex beams in ...
    Presented in this work is a versatile model designed for simulating the propagation of focused scalar and vector beams through three-layered systems, ...
  30. [30]
    Vector Vortex Beam Generation with a Single Plasmonic Metasurface
    Jul 20, 2016 · This method features the spin–orbit coupling and the superposition of the converted part with an additional phase pickup and the residual part ...
  31. [31]
  32. [32]
  33. [33]
  34. [34]