Simulation¶

eSDIva splits simulation into two families, both driven by the SDI spatial impulse response engine:

Emission

Forward pressure field radiated by the transducer — monochromatic CW, transient pulsed, or with power-law attenuation. Returns p(x, y, z) (CW) or p(t, x, y, z) (transient).
Reception (RF)

Pulse-echo RF from point scatterers: PSF, phantoms, FMC, and PW/DW sequences. Returns per-element channel data (Erx, Nt). Field II-accurate,

20× faster on large apertures.

Field input format¶

Emission field points and reception scatterer lattices share the same grid dict:

field_points = {
    "x_extent": [-5, 5],    # mm — lateral
    "y_extent": [-0.5, 0.5],  # mm — elevation
    "z_extent": [5, 55],    # mm — axial (depth)
    "dx": 0.1,
    "dy": 1.0,
    "dz": 0.2,
}

Set one axis to a single value with its d=0 to compute a 2-D plane. Emission also accepts a raw (N, 3) mm array; you then reshape the flat output yourself.

Method selection (SDI)¶

Method	Meaning	When
`"auto"`	Picks FST or SDI from grid size	Always recommended
`"FST"`	Fully-Sampled Trapezoid — direct evaluation	Small grids, reference
`"sdi"`	Sparse Delta Integration — sparse deltas + integration	Large / dense grids

Both give the same result; SDI is the fast path for large apertures (>100× on emission, >20× on reception) with negligible difference from Field II.

Medium properties¶

Override defaults in the constructor (shared by Emission and Reception):

sim = Emission(tx, c=1540, rho=1.0, fs=200e6, alpha0=None)

Parameter	Default	Description
`c`	1540 m/s	Speed of sound
`rho`	1.0 kg/m³	Medium density
`fs`	100 MHz	Sampling frequency
`alpha0`	`None`	Power-law attenuation dB/(MHz·cm); `None` disables it