Getting Through Goodman Without Losing Your Mind
Fourier optics is one of those subjects that looks elegant on paper and falls apart the moment you try to use it in a real lab. Goodman's book is the standard reference. It's also not easy to work through without some background. The 3rd edition tightened up the notation and added a few more chapters on computational methods, but the core difficulty remains the same. The Introduction To Fourier Optics Goodman 3rd Edition starts with scalar diffraction theory, moves into Fourier transforms and their properties, then applies those tools to lenses, imaging systems, and coherence. That's the surface description. The book assumes you can already do complex analysis at an intermediate level and that you're comfortable with the idea of a function being represented in both spatial and frequency domains. If you aren't, you'll spend more time wrestling with the math than learning the optics. Chapter 2 is where most people hit a wall. The Fourier transform properties section is dense. Goodman lists them quickly because he assumes you've seen them before, but in optics they're used repeatedly in every derivation that follows. I've had students skip ahead and then get lost in Chapter 4 because they never fully internalized convolution vs. correlation in the spatial frequency domain. The fix is simple: work through the derivations yourself. Don't just read them. Write them out on paper with actual functions. The act of deriving the Fresnel diffraction integral from the angular spectrum method once, slowly, makes everything after it click in a way that reading never will.
A note on the exercises. They're not trivial. Some are straightforward plug-and-chug, others require you to set up approximations that the book doesn't spell out. I'd say roughly a third of the problems are useful for building intuition, another third are grinding, and the final third are the ones that actually teach you something. The grinding problems are still worth doing if you're preparing for qualifying exams. The useful ones are worth doing if you want to actually understand the material.
How I Used This Book in Practice
I worked with Fourier optics for computational imaging and holography. The book was my reference for everything from designing a 4f correlator to understanding the transfer function limits of a microscope. It's not a textbook you read cover to cover for practical work. It's a book you pull off the shelf when you need to derive a new system or debug why your simulation doesn't match the measurement. One specific case that comes to mind: I was modeling a partially coherent imaging system and needed the mutual coherence function propagating through a lens. The book covers coherence in Chapter 7, but the derivations assume paraxial propagation and perfect lenses. My setup had a finite aperture and the source wasn't perfectly incoherent. I spent about two days trying to make the textbook formula work and it kept giving unphysical results near the edges of the field. The workaround was to go back to the angular spectrum representation, apply the finite aperture as a multiplicative mask in the frequency domain, and then propagate numerically instead of using the closed-form impulse response. Goodman's treatment of the coherent transfer function gave me the starting point. The numerical approach was something I had to figure out outside the book. That's the pattern with this material. The theory is clean. The application is messy. Another thing that isn't obvious from the book: the scalar diffraction approximations break down when your features approach the wavelength. Goodman mentions this briefly, but he doesn't dwell on it because the whole framework is built on scalar theory. If you're working with subwavelength structures, metasurfaces, or near-field scanning, you need to move to vector diffraction or full electromagnetic simulation. The book won't help you there. I learned that the hard way when a design based on scalar Fourier optics kept failing at the nanoscale and I had to switch to FDTD.
Get the Full Details
Common Mistakes People Make
Mixing up the Fraunhofer and Fresnel regimes. The boundary between them isn't as clean as the book makes it seem. The rule of thumb about the Fresnel number being less than one is useful, but it depends on what you're measuring. If you're looking at intensity, the transition happens at a different distance than if you're looking at phase. I've seen this cause significant errors in alignment procedures for interferometers. Ignoring the bandlimit imposed by the numerical aperture. Every lens has a cutoff frequency. Goodman derives it in Chapter 6. Students often forget it when they simulate ideal systems. A simulation that doesn't include the NA limit will predict resolution that doesn't exist in practice. This is especially relevant for digital holography where the sampling grid can create aliases that look physically reasonable until you compare them to real data. Assuming the Fourier transform of a product is always a convolution. It is, but only under the right conditions. If your sampling isn't fine enough in the spatial domain, the convolution in the frequency domain wraps around and corrupts your result. This is a fundamental issue with discrete Fourier transforms, not a flaw in Goodman's theory. I've seen this ruin entire simulations. The fix is zero-padding. Pad your spatial domain data by at least a factor of two before taking the transform. It costs more compute but it prevents the aliasing.
Who Should Use This Book and How
If you're a graduate student in optics or electrical engineering, this is required reading. Not because it's enjoyable, but because it's the foundation. Most advanced topics in imaging, holography, and optical signal processing trace back to the equations in this book. The 3rd edition is worth getting over the 2nd because the newer chapters on computational Fourier optics are relevant to modern work. The older edition is fine for the classical theory, but it misses the numerical methods section. If you're an undergrad, you probably aren't ready for this unless you've already taken a course in complex variables and signals and systems. The math moves fast. The physical intuition builds gradually. There are better introductory books if you're seeing Fourier optics for the first time. Goodman is the book you go to after you've had that introduction. The book is widely available as a physical copy and through university libraries. There are scanned copies floating around online, but those are often lower quality and harder to reference. If you can afford the hardcover, get it. The paper quality and print are decent. The paperback has some binding issues that make it hard to keep open while you're writing derivations on the same page.
One final practical note. Don't try to memorize the formulas. Goodman himself says the book is meant to be read with a pen in hand. Work through the derivations. Derive the Abbe theory of imaging yourself. Work out the optical transfer function for an incoherent system from first principles. That's how this material sticks. Reading it passively will get you through a homework assignment and nothing else.
