A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
Quadrature modulation uses two carriers in phase quadrature (90° apart) to transmit two information streams: s(t) = I(t)·cos(ωct) – Q(t)·sin(ωct). Complex notation s_c(t) = (I + jQ)·ejωct simplifies analysis where s_c is the complex envelope. Demodulation by correlating with local carriers recovers I and Q components, enabling bandwidth-efficient communication.
From your study of AM, FM, and shift-keying modulation, you know the basic idea: modulation is the process of imprinting information onto a high-frequency carrier so it can be transmitted over a channel. AM encodes information in the carrier's amplitude; FM in its frequency. Each of these uses a single real carrier and transmits one stream of information per channel. But a fundamental insight — partially anticipated by the analytic signal concept you encountered with the Hilbert transform — is that a sinusoidal carrier actually has two degrees of freedom available simultaneously: its amplitude and its phase. Quadrature modulation exploits both degrees of freedom at once, doubling the information density without using any additional bandwidth.
The core idea is that two sinusoids at the same frequency but 90° apart in phase are orthogonal: their product integrated over a complete cycle is exactly zero. This means they can coexist on the same carrier without interfering with each other. The transmitted signal is s(t) = I(t)·cos(ωct) − Q(t)·sin(ωct), where I(t) is the in-phase component (the baseband signal multiplied onto the cosine carrier) and Q(t) is the quadrature component (multiplied onto the sine carrier, which lags the cosine by 90°). To recover I(t) at the receiver, you multiply s(t) by 2cos(ωct) and lowpass filter — the cosine×cosine term gives I(t) while the cosine×sine term (the Q contribution) integrates to zero. Symmetrically, multiplying by −2sin(ωct) recovers Q(t). The orthogonality of sine and cosine is what makes this separation perfect.
The complex envelope notation unifies this. Define the complex baseband signal as ẑ(t) = I(t) + jQ(t). The transmitted passband signal is s(t) = Re[ẑ(t)·ejωct]. This is exactly the analytic signal framework from your Hilbert transform study: ẑ(t) is the complex envelope, and the passband signal is its real part modulated onto the carrier. The magnitude |ẑ(t)| = √(I² + Q²) is the instantaneous amplitude; the angle ∠ẑ(t) = arctan(Q/I) is the instantaneous phase. AM is ẑ(t) = A(t) with constant phase; PM is ẑ(t) = e^(jφ(t)) with constant magnitude; QAM and PSK are ẑ(t) choosing from a constellation — a discrete set of complex points in the I-Q plane. Plotting I vs Q gives the constellation diagram, where different modulation orders (QPSK, 16-QAM, 64-QAM) appear as grids of points at different distances and angles from the origin.
The practical power of I/Q representation is that all signal processing — pulse shaping, equalization, matched filtering, carrier recovery — can be done entirely in the complex baseband domain at low frequency, without ever working at the RF carrier frequency. The hardware architecture separates naturally: a quadrature modulator chip takes digital I and Q samples, converts them to analog, and multiplies by cosine and sine carriers (with a 90° phase shift between them), summing the result for transmission. The receiver uses a quadrature demodulator to split the incoming RF signal into I and Q paths and downconvert each to baseband. All the intelligence — OFDM, MIMO, adaptive equalization — operates on the I/Q streams. This is why every modern software-defined radio, WiFi chip, LTE modem, and satellite receiver is built around I/Q architecture: it cleanly separates the carrier physics from the signal processing.