
$$ x(t)=m(t)cos(\omega _c t) $$
Double Sideband Supressed Carrier
can’t use coherent detection because the envelope is not proportional to the signal.
$$ BW=2W_m $$
Single Sideband Modulation
extract either sideband using ideal filters as it contains the information signal.
$$ BW= W_m $$
Vestigial Sideband Modulation
ideal filters are not causal, use practical filters to extract most of either sidebands but due to the non-zero filter transition a vestige of the other sideband is transmitted.
$$ BW= W_m + W_v $$

$$ x(t)=cos(\omega _ct + k_p m(t)) $$

the phase of the carrier wave is varied in proportion to the instantaneous value of the modulating (low-frequency) signal.
$$ \theta_i(t)=\omega_ct + \phi(t) $$
Bandwidth of single tone sinusoidal angle modulated signal
$$ x(t) = \sum_{n=-\infty} ^ {\infty} AJ_n(\beta) cos[(\omega_c - n\omega_n)t] $$
where $\beta =\frac{ \Delta \omega} {\omega_m}$ and $J_n(\beta)$ is the bessel function.
as $\beta$ increases, more power goes to the sidebands which carry the information and less power goes to carrier $J_0(\beta)$
$$ BW= 2n_{max} W_m $$
$$ x(t)= cos(\omega _c t + k_f \int _{-\infty} ^t m(\tau) d\tau ) $$

To change the frequency of the carrier using the information signal.
$$ \omega_i(t) =\frac{d\Theta_i(t)}{dt}= \omega_c + \frac{d\phi(t)}{dt} = w_c + \Delta \omega $$
Bandwidth & Carson’s Rule
$$ BW=f_{max}-f_{min} + B_1 + B_2 $$
and if the signal is symmetric with equal time intervals
$$ BW=2\Delta f + 2B $$
for phase modulation $\Delta f =0$
Full AM
to make the signal proportional to its envelope such that it is demodulated non-coherently, add a DC component to the information such that
$$ x(t) = [A+m(t)]cos(w_ct) $$
$$ \mu = \frac{-min\{ m(t) \}} A \le 1 $$

AM is power inefficient since only 33.3% of the power is spent on the information signal and 66.7% on the carrier.
$$ \eta= \frac {P_m}{P_m + A^2} $$
converting an analog signal into a digital signal such that we get better detection and security.

Using Nyquist-Shannon Sampling theorem, pick $f_s$ such that $f_s \ge 2B$ for proper detection.
Sampling produces a signal with discrete time values and infinite amplitudes.
given the sampled amplitudes, quantization produces finite amplitudes with finite time values.
$$ \Delta = \frac{max\{m(t)\}-min\{m(t)\}}{M} $$

Each sample will be assigned the center point $(\frac{x_1 + x_2} 2)$ of the partition its value exists in.
In quantization, we have a max error of $\frac \Delta 2$ and min error of $- \frac \Delta 2$ uniformly distributed ⇒ $e(t) = U (0, \frac {\Delta^2} {12})$

Encoding produces a bit stream with a previously agreed upon code based on how many bits to represent each sample.
$$ BW= nf_s [Hz] $$
DSBSC Amplitude
Vestiginal Amplitude
Angle Phase Locked Loop (PLL)
using a line coder, the bit stream is converted to a digital signal, either polar or unipolar NRZ encoding.