What Is A Complex Number

The Bare Bones of It

A complex number is just a number that has two parts: a real part and an imaginary part. You write it as a + bi, where a and b are regular real numbers and i is the square root of negative one. That's it. Nothing mystical about it. When you add two complex numbers, you add their real parts together and their imaginary parts together. When you multiply them, you use distributive properties like regular algebra, then replace i² with -1. You'll end up with another complex number. People in engineering and signal processing use them constantly because they make certain calculations vastly simpler than working with real numbers alone. Electrical engineers dealing with alternating current circuits use complex impedance all day. Control theorists use complex roots to determine system stability. Quantum mechanics is built on them. The practical reason to care is that some problems in physics and math are nearly impossible to solve cleanly without them, even if the final answer you extract from them is purely real. I ran into a specific problem last year while working on a filter design project. I was computing the frequency response of a fourth-order Butterworth filter and kept getting garbage results near the cutoff frequency. The transfer function had complex conjugate poles sitting very close to the imaginary axis. When I computed the response numerically using only real arithmetic, roundoff errors were dominating the result. The workaround was straightforward: I restructured the transfer function into second-order sections, each section represented in complex form, and I evaluated each section separately before combining them. This kept the pole locations well-defined and eliminated the cancellation error that was poisoning the overall result. It took me about three hours to refactor the code and the results were correct on the first run after that.

How It Actually Works in Practice

Complex numbers live in a two-dimensional plane called the complex plane. The horizontal axis is the real part and the vertical axis is the imaginary part. When you multiply two complex numbers, their magnitudes multiply and their angles add. This is why the polar or exponential form is so useful. You write a complex number as r times e to the power of i theta, where r is the magnitude and theta is the angle. Multiplication in this form is just: multiply the magnitudes and add the angles. It is dramatically easier than using rectangular form for that operation. Addition is different though. Addition works best in rectangular form. You literally add the real parts and the imaginary parts. Mixing the two forms in your head during a calculation is how most mistakes happen. I still see people try to add complex numbers in polar form and then wonder why the answer is wrong. It is a basic rule but it gets forgotten under time pressure. Here is a concrete example that covers the typical workflow. Take the complex number z equals 3 plus 4i. Its magnitude is 5, calculated as the square root of 9 plus 16. Its angle is approximately 53.13 degrees or about 0.927 radians. If you need to multiply this by w equals 1 minus i, you can convert both to polar form first. The magnitude of w is the square root of 2, and its angle is negative 45 degrees. Multiply the magnitudes to get 5 times the square root of 2, which is roughly 7.07. Add the angles to get about 8.13 degrees. Convert back to rectangular form and you get approximately 6.99 plus 0.91i. If you do this same multiplication in rectangular form directly, you get 3 plus 3i minus 4i minus 4i squared, which simplifies to 3 minus i plus 4, giving 7 minus i. The polar method and rectangular method match, but the rectangular method was faster here because one of the numbers was already simple.

Things Nobody Warns You About

The first pitfall is assuming that you can take the square root of a negative number and move on without tracking branches. The square root of negative one is i, but so is minus i. In most casual calculations this does not matter. In control theory and signal processing, the choice of branch can flip a stability conclusion. Always be explicit about which branch you are using when the problem involves roots. The second thing that catches people off guard is that complex conjugates are not just a mathematical curiosity. When you have a polynomial with real coefficients, complex roots always come in conjugate pairs. This means if a plus bi is a root, then a minus bi is also a root. This has a direct impact on how you implement filters and. You group conjugate poles into second-order sections rather than handling each pole individually, which reduces computational cost and numerical sensitivity. There is also a limitation that is worth stating plainly. Complex numbers do not solve every problem. If you are doing heavy numerical simulation involving millions of operations on complex data, memory bandwidth becomes a bottleneck. A complex double-precision number takes 16 bytes. A real double takes 8 bytes. Operations on complex arrays are roughly twice as expensive in memory, and many processors do not have native complex arithmetic units, so the compiler generates extra instructions. If performance is critical and your problem can be reformulated in real arithmetic, it often will run significantly faster. I have seen simulation runtimes cut from around 40 minutes down to about 18 minutes simply by switching a complex-only implementation to a real-valued formulation using trigonometric identities instead of complex exponentials.

When Complex Numbers Break Down

They are not a universal replacement for real numbers. If you are building a simple calculator app or doing basic financial computations, complex numbers will only slow you down and confuse users. If your data is inherently one-dimensional and real, forcing it into complex form adds unnecessary complexity to both the code and the mental model. Quaternion notation exists for three-dimensional rotation problems and sometimes serves better than treating each dimension as a separate complex plane. Octonions exist too, but they lose associativity and are rarely useful outside pure mathematics. For most practical engineering work, the standard approach is sufficient. Use rectangular form for addition and subtraction. Use polar or exponential form for multiplication and division. Keep conjugate pairs grouped when factoring polynomials or designing filters. And do not be afraid to convert between forms whenever one form makes the current operation obviously simpler.

A Quick Reference for Common Operations

Conjugate of a plus bi is a minus bi. Magnitude is the square root of a squared plus b squared. The product of a complex number and its conjugate equals the magnitude squared, which is a useful identity when rationalizing denominators. Division of two complex numbers is handled by multiplying numerator and denominator by the conjugate of the denominator, which turns the denominator into a real number. Euler's formula states that e to the i theta equals cosine theta plus i sine theta, and this is the bridge between exponential and trigonometric representations that makes polar arithmetic work. Argand diagrams are the standard way to visualize these numbers geometrically, though they are mostly pedagogical tools rather than practical ones in professional work.