FM synthesis looks intimidating, but you can learn it with three ideas: a carrier, a modulator and an index. One oscillator bends the pitch of another fast enough to create new harmonics.
FM sits alongside sound design basics as one of the most distinctive ways to build a tone from scratch.
| Item | Value |
|---|---|
| Year Chowning noticed the effect | Autumn of 1967 |
| Yamaha DX7 release | 1983 |
| Modulator rate where new timbre appears | Above 30 Hz |
| Upper sideband positions | C+M, C+2M, C+3M and so on |
| Example ratio in the SFU tutorial | 1:2 gives a first upper sideband at 3 |
The figures come from Yamaha’s history of digital FM and the SFU FM tutorial. Sources can change or be updated, so check them before you rely on the numbers.
What FM synthesis actually does
In frequency modulation, one sine wave (the modulator) changes the pitch of another (the carrier). At slow speeds you hear ordinary vibrato.
Speed the modulator up and something changes. The SFU tutorial puts it this way: “when we raise M to the audio range (above 30 Hz) then we hear a new timbre composed of frequencies called sidebands.”
That is the whole trick. Two simple sine waves, which sound dull alone, produce a rich spectrum when one modulates the other.
Where FM came from
The effect was found by accident. Yamaha dates the discovery to autumn 1967, during Chowning’s computer experiments with vibrato at Stanford.
When the vibrato rate rose past the point where it could be heard as a cycle, the sound became a timbral change. Chowning was John Chowning, a Stanford composer and researcher whose main interest was sound spatialization.
The path to products went like this:
- In the early 1970s, Stanford sought commercial licensees for Chowning’s patents.
- Yamaha engineer Kazukiyo Ishimura visited Chowning following Stanford’s approach to the company.
- Making FM practical for a portable keyboard required substantial investment in custom chip development.
- The Yamaha DX7 arrived in 1983 and was described by Yamaha’s own history as a watershed moment.
Carrier, modulator and sidebands
The carrier is the oscillator you hear as the pitch. The modulator is the one that wobbles it. Each has a frequency, written C and M.
The new frequencies are called sidebands. According to the SFU tutorial, the upper ones sit at “C+M, C+2M, C+3M, C+4M, C+5M,…”, and lower ones mirror them below the carrier.
Take a 1:2 ratio, where the modulator is twice the carrier frequency. The first upper sideband is C+M = 1+2 = 3, and the next is (C+M) + M = 3+2 = 5.
Ratios decide whether it sounds tuned or metallic
The C:M ratio controls which sidebands appear. The SFU tutorial notes that sidebands can be harmonic or inharmonic, meaning not multiples of the fundamental.
A practical rule of thumb follows:
- Whole-number ratios such as 1:1 or 1:2 give harmonic, pitched tones: basses, organs, electric pianos.
- Awkward ratios give inharmonic spectra: bells, gongs, metallic hits.
- Reducible ratios are redundant, because 2:2 is the same as 1:1.
Start with whole-number ratios when you want a note that sits in a key. Change ratios to explore different spectra; slight detuning can also introduce beating.
The modulation index is your brightness knob
The SFU tutorial explains that the amplitude of each sideband varies with the strength of modulation, expressed by the modulation index. A low index is close to a plain sine. A higher index adds more strong sidebands and a brighter tone.
Real sounds change over time, so FM patches usually put an envelope on the index. A short, high index at the start gives a bright attack that mellows as it decays, which is why FM electric pianos and plucks sound so natural.
Compared with filtering, this is a different mindset. In subtractive synthesis you carve harmonics away from a rich wave, while FM builds them up from sines. If you prefer scanning through stored waveforms, wavetable synthesis is another route to evolving timbres.
Why FM patches feel unpredictable
In subtractive designs, turning a knob usually does one obvious thing. Changing the ratio changes sideband positions, while changing the index changes their amplitudes.
That is not a flaw. It is why FM can produce sounds that filters struggle to imitate, such as glassy electric pianos, hollow mallets and sharp digital basses.
A few habits make the behaviour easier to predict:
- Change one parameter at a time, and write down the ratio and index of sounds you like.
- Keep the modulator and carrier as sines while learning, so every change you hear comes from the modulation itself.
- Use the envelope on the index to control brightness, and the envelope on the carrier’s level to control loudness. They are separate jobs.
- Check the sound across the keyboard. A patch that is warm in the middle can turn harsh or thin at the extremes.
Operators and algorithms in hardware FM
Classic FM instruments call each sine-wave unit an operator. Every operator can act as a carrier, a modulator, or both, depending on how it is connected.
The wiring pattern is called an algorithm. Some algorithms stack operators in a chain, so one modulates another that modulates the carrier. Others run several carriers side by side, each with its own modulator, which lets you layer distinct tones in one patch.
Chains let you modulate a modulator, while parallel carriers let you combine separately shaped tones. When you open a hardware or software FM synth, look for the algorithm diagram first, since it tells you which operators you hear and which only shape the others.
A simple first patch
- Set two sine oscillators, one carrier and one modulator, with the modulator’s output routed to the carrier’s frequency.
- Choose a 1:1 ratio and raise the index slowly. Listen to the tone brighten.
- Try 1:2, then an unusual ratio, to hear harmonic turn into bell-like.
- Add a decaying envelope to the index for a pluck.
- Add a second modulator only after the two-operator version makes sense.
Small changes to the index and ratio matter more than adding operators. Learn those two controls first.
Once the basic patch works, save it and make variations. Try moving the ratio by small steps, then try changing the index envelope’s length. You will quickly build a library of ratios you recognise by ear, which is the real skill in FM programming.
It also helps to compare each result against a plain sine. Hearing what the modulator adds, and what it removes when you pull the index back to nothing, trains your ear faster than any chart.
Frequently asked questions
Is FM synthesis hard to program?
Show answer
The idea is simple, but the results are not intuitive, because small ratio changes cause big timbral jumps. Working with two operators and one ratio at a time makes it manageable.
What is a sideband?
Show answer
A sideband is an extra frequency created by modulation. The upper ones are at C+M, C+2M and so on, according to the SFU tutorial.
Why do FM sounds go metallic?
Show answer
Ratios whose frequencies do not share a common fundamental produce inharmonic sidebands. Your ear hears these as bell-like or metallic.
Who invented FM synthesis?
Show answer
John Chowning discovered the effect in the autumn of 1967, per Yamaha’s history. Yamaha later licensed it and released the DX7 in 1983.
