Clever Quest Lab ← Labs

Modulation Lab

One sound, your voice or a test tone, through every common kind of modulation: compare them, and hear what a slightly mistuned receiver makes of each.

Tip: click or tap the spectrum to tune the receiver there.

Sound is off. Press Play to hear the received signal.

99% bandwidth–
Carrier–
Sidebands–
Receiver–

Sound source

Transmitter

Receiver

Channelclean

The transmitter's peak power against the noise in a 3 kHz band. At the right end there is no noise.

Runs on your device. Uses your microphone; the sound is processed on your device. Nothing you type or record is uploaded. Keeps working without a connection while the page is open. Works with touch. Saves the transmitter, channel, receiver and view settings on this device; recordings and opened files are not kept.

Added

What this lab shows

A transmitter puts a sound a few kilohertz wide onto a carrier, and the receiver takes it off again. The ways of doing it differ in bandwidth, in how much of the power carries the sound, and in how they cope with a receiver set a little wrong. Here one sound goes through all of them, so the differences can be heard and measured side by side.

Things to try

What to look for

How it works

Everything runs at the sample rate of your sound output, and the “radio frequency” is a quarter of that rate: 12 kHz at 48 kHz, 11.025 kHz at 44.1 kHz. The spectrum is labelled in kilohertz from that carrier. Real stations sit at megahertz, but the arithmetic of modulation is the same.

Transmitter

The sound x(t), scaled to a peak of 1, makes a complex baseband signal z(t), and the transmitted wave is

s(t) = Re{ z(t) · ej2πfct }

The microphone and files pass a 250 to 2900 Hz speech filter first, as in a real transmitter.

Channel

Noise is white Gaussian noise added to the wave. Two-path fading adds an echo, z(t) + a·ejθ(t)·z(t − τ), whose phase θ turns slowly: that cuts notches 1/τ apart which drift across the signal (with a 1 ms echo, one every kilohertz). The second station is an AM signal playing its own tune.

Receiver

The receiver multiplies the wave by its own oscillator at the carrier frequency plus the tuning offset and passes the result through a 2049-tap filter of the chosen width: above the oscillator for USB and CW, below it for LSB, on both sides otherwise. Then it detects:

An automatic gain control evens out the loudness, as in a real receiver.

Bandwidth

The “99% of power” readout is the usual definition of occupied bandwidth: 0.5% of the power lies below the band and 0.5% above it. For FM, Carson's rule gives a quick estimate,

B ≈ 2(Δf + fmax)

which keeps about 98% of the power. It is named after John Renshaw Carson, whose 1922 paper on the theory of modulation it comes from.

Limits of the model

Controls

Frequently asked questions

Why does a receiver need to know USB or LSB?

Because the two sidebands are mirror images. A receiver in LSB looks below its dial, where a USB signal has nothing. Tune the dial 3 kHz above the carrier and the signal falls into the filter upside down: with Two tones sent in USB, 700 Hz comes out at 2300 Hz and 1900 Hz at 1100 Hz. Data modes sent through SSB transmitters turn over the same way: the mark and space tones of radioteletype swap places, so the text is garbage until the sideband is right or the decoder is set to reverse.

Sources