Demo — Loudest Band#
Give one aircraft four signature bands — radar, IR, visual, and acoustic — each detected by a matched threat with its own range law. Detection is a union: if any band crosses its threshold you are found, so the range that matters is the largest one. This demo lets you pour signature reduction into each band and watch the fused detection range refuse to move until you finally hit the limiter.
The fusion rule#
Each band has its own detection range at its baseline signature, and each shrinks under investment along a different law:
The visual band is the one with a floor, and the floor is the point. Paint and finish genuinely buy contrast, so the band has a slider — but a contrail is a cloud the airframe manufactures, and a glint is the sun on a flat panel. Neither is a coating problem, so contrast reduction bottoms out: this demo stops the visual band at 6 km however far you drive it. That floor sits above the 4 km acoustic range on purpose — for a high-altitude penetrator the eye outranges the ear, so acoustic is never the band that finds you here.
Because the defender needs only one band to cross its threshold, the fused detection range is the maximum, not the average:
At baseline the ranges are radar 45 km, IR 30 km, visual 12 km, acoustic 4 km — so \(R_\text{det} = 45\) km and the loudest band is radar. Everything interesting follows from what happens when you attack one band at a time.
Interactive demo#
Walkthrough#
The three sliders apply signature reduction in decibels to the radar, IR, and visual bands; acoustic has no knob and sits fixed at 4 km, and the visual band stops at its 6 km contrail/glint floor. The three preset chips drive the sliders to their 40 dB corner for you, but you can dial any value in between.
Read the baseline. Four bands, four ranges — radar 45, IR 30, visual 12, acoustic 4 km. The bar chart darkens the loudest band, the dashed red line marks the fused detection range, and the readouts agree: \(R_\text{det} = 45\) km, loudest band radar. That single number is what a fused defender achieves.
Spend on the wrong band. Pour 10 dB of reduction into IR alone. IR’s bar drops to about 9 km, and \(R_\text{det}\) does not move — it stays pinned at 45 km, because radar was always the limiter. A non-dominant band buys nothing.
Spend on the loudest band. Now pour reduction into radar. \(R_\text{det}\) falls — then flattens at 30 km when IR inherits the limit. The lower plot sweeps 0–40 dB into whichever band is currently loudest and draws that whole story as one curve: a fall, then a flat, with the handoff point marked. More radar reduction past the handoff is wasted; the defender simply leans on IR.
Spend on both. Only cutting radar and IR together drives \(R_\text{det}\) down to 12 km, where the visual band takes over. Nothing in a radar or IR package buys below that, because neither treatment touches what the eye sees.
Push the visual band into its floor. The Visual (C) slider models contrast reduction — paint, finish, glint control. Drive it from the balanced state and \(R_\text{det}\) falls from 12 km to 6 km and then stops dead, about 6 dB in, with the dotted ochre line on both plots marking where. Visual is still the loudest band at its floor: radar has been driven to 4.5 km and IR to 0.3 km, and neither is above 6. This is the lesson’s sharpest point made mechanically — you can buy contrast, but you cannot buy your way under a contrail, so past that dB the only levers left are route, altitude and time of day.
Key observations#
Your worst band decides your range. \(R_\text{det} = \max_b R_b\) means the loudest band alone sets how close a fused defender can get. The average signature is irrelevant.
Reduction on a quiet band is invisible. Until a band becomes the limiter, cutting it changes nothing — this is the single most common way to waste weight, money, and maintenance.
The limiter moves. Each cut hands the lead to the next band, so the range curve falls in steps and flattens between them. Attack the limiter, expect it to jump, re-plan, repeat.
Some floors resist coatings. The eyeball (contrail, glint) and, tail-on, the plume set floors no material buys below — floors you manage with route, altitude, and finish rather than with treatment. The demo makes this one enforceable rather than advisory: the visual band stops at 6 km no matter what the slider says. This is the loudest-first budget of L26, now spanning the whole spectrum — and it is the analytical core of Project 3.
Source#
MATLAB bundle · L28_LoudestBand.m↓
The companion script builds one aircraft with four bands, computes each band’s detection range under its own law, and takes the maximum for the fused \(R_\text{det}\). It then sweeps signature reduction into the wrong band (nothing moves), into the loudest band (the range falls, then flattens as IR takes over at 30 km), and into both (reaching 12 km, where the visual band takes over). The script carries no visual knob — it treats 12 km as the floor a radar-plus-IR package can reach — so it reproduces steps 1 through 4 above, not step 5.