A forward-looking sonar and a multibeam echosounder are both "acoustic cameras" in the sales literature, and that is where most procurement mistakes begin. One is built to answer is something there, and where; the other is built to answer how deep is it, and where exactly. The hardware differs accordingly, and no specification sheet will tell you that you bought the wrong one.
This article separates the two by geometry, by resolution arithmetic, and by the data product each supports — then shows where a single survey legitimately needs both.
Detection and measurement are different businesses
A forward-looking sonar (FLS) projects a sector ahead of the platform and renders it as an image. Its job is to reveal structure: a wall, a boulder, a pipeline, a diver, a shoal of fish, a berth fender. The output is a picture with ranges and bearings, updated fast enough to be useful while the platform is moving.
A multibeam echosounder (MBES) projects a fan of narrow beams downward and across track, measures the two-way travel time to the seabed on each beam, and converts those ranges into soundings. Its job is to produce a georeferenced depth surface with a stated uncertainty. The output is a point cloud with vertical and horizontal accuracy, and it is only as good as the motion, heading and sound-speed data that accompany it.
The distinction matters commercially because the two are specified with different words. An FLS datasheet leads with field of view, angular resolution and update rate. An MBES datasheet leads with beam count, across-track beam width, swath width as a multiple of depth, and depth accuracy. Compare an FLS and an MBES on "resolution" alone and you will conclude that the FLS is a better measuring instrument, which is false; it simply resolves differently.
Geometry: a sector ahead, a swath beneath
An FLS sees an angular sector: our FLS100 covers 105° horizontally and 26° vertically in imaging mode, narrowing to 7° in collision-avoidance mode to concentrate energy and range. The frame is the platform's, and it moves with the platform.
An MBES sees a swath: a fan across the track, with a footprint that grows with depth. For a 130° total opening the swath is roughly 4.3 times the water depth — 129 m wide in 30 m of water — while along-track resolution depends on beam width, speed and ping rate. The frame of reference is the seabed, in absolute coordinates, and that is the whole point.
A practical consequence follows immediately: an FLS covers a region the vehicle has not reached yet, an MBES covers a region the vehicle has just passed over. For obstacle avoidance that difference is everything; for charting it is irrelevant.
The second consequence is about gaps. An FLS sector has a vertical aperture of 20–26°, so it images a slice, not a volume, and its coverage falls off with range as the sector widens. At 120 m, a 150° horizontal field spans roughly 896 m across the far edge while the 20° vertical slice is only about 42 m tall — a wide, shallow ribbon rather than a picture of the seabed. An MBES intersects the seabed, so its coverage is continuous along track provided the line spacing is right. Judging either instrument by "how much area does it cover per ping" mixes two incompatible geometries.
Resolution arithmetic: angular resolution sets target size
Range resolution and angular resolution are quoted side by side in every FLS datasheet, and only one of them usually matters. Range resolution — 4 mm at 750 kHz, 2.5 mm at 1.2 MHz on our dual-frequency unit, 3.75 cm on the 100 kHz FLS100 — tells you how finely a target's range is sliced. Angular resolution tells you how large a target must be, at a given range, to occupy one beam and therefore be distinguishable at all.
The conversion is the small-angle relation: cross-range footprint equals range times angular resolution in radians. At 0.96° (0.01676 rad), that is 0.67 m at 40 m, 1.68 m at 100 m, 3.35 m at 200 m and 8.4 m at 500 m. A 100 kHz FLS that promises 500 m of detection is promising, at that range, to resolve objects about 8 m across — which is honest for a dam wall and useless for a mooring shackle.
Running the relation backwards gives the specification you should actually tender. For a 0.5 m target at 50 m, the required angular resolution is 0.5 / 50 rad, or 0.57°; at 100 m it is 0.29°; at 200 m it is 0.14°. That is why long-range obstacle avoidance and short-range inspection are different products, and why "0.5 m resolution" without a range next to it is a meaningless claim.
Our dual-frequency unit shows the trade explicitly: 750 kHz gives 120 m at 1° (2.1 m footprint at full range), while 1.2 MHz gives 40 m at 0.6° (0.42 m at full range). The higher frequency is not "better"; it is narrower in range and finer in angle, which is a different mission.
The outputs, and the tasks they support
| Task | What the task actually needs | Forward-looking sonar | Multibeam echosounder |
|---|---|---|---|
| Obstacle avoidance ahead of a vehicle | detection at range, low latency, platform frame | fits | not suited: looks down, not ahead |
| Harbor wall, pile and fender inspection | fine angular resolution at close range | fits | partial: gives depth and backscatter, not a face-on image |
| Charting and volume computation | georeferenced soundings with stated uncertainty | not suited: no vertical reference | fits |
| Search and recovery | wide sector, fast update, target reacquisition | fits | partial: needs overlapping lines |
| Dredge and berth monitoring | repeatable depth surface | not suited | fits |
| AUV forward navigation | small mass, low power, 20–40 Hz | fits | not suited |
| Fish biomass and behaviour | imaging of a moving target in the water column | fits | partial: echo integration, not imaging |
The table is deliberately blunt. Most failed procurements come from asking one instrument to do the row it cannot do, then compensating with software. The short version of this table is the forum discussion on imaging sonar versus multibeam echosounder, which starts from the same question buyers ask.
Our forward-looking lineup, and where each model fits
| Model | Frequency | Max range | Angular resolution | Field of view | Update rate | Mass / depth |
|---|---|---|---|---|---|---|
| UAC-FLS100 (Tingseek) | 100 kHz | ≥500 m (-15 dB target) | 0.96° | 105° × 26° (7° avoidance) | — | 30 kg / 600 m |
| MiniUAC-7512D (Tingseek) | 750 kHz / 1.2 MHz | 120 m / 40 m | 1° / 0.6° | 150° × 20° | ≥40 Hz | ≤1 kg / 600 m Al, 2,000 m Ti |
| FLS 750/1200 (Tingseek) | 750 kHz / 1.2 MHz | 120 m / 40 m | 1° / 0.6° | 130°/80° × 20°/12° | up to 40 Hz | 1.1 kg / 500 m |
| MiniUAC-3D-300 (Tingseek) | 300 kHz | 150 m | 50°×50° or 55°×55° FoV, ≤3 cm range | 3D | 2–20 Hz | ≤2.5 kg / 300 m |

UAC-FLS100: 100 kHz, 0.96° angular resolution, 105° horizontal field of view, and a 7° collision-avoidance mode for longer range on a narrower sector.
The three small units sit in different mission boxes rather than on a quality ladder. The 7512D and the 750/1200 are AUV-scale: about a kilogram, tens of watts, 40 Hz, with 512 beams and a 0.1 m minimum range on the 750/1200 variant, which is what confined-space inspection needs. The FLS100 pays for its 500 m detection range with mass, power and a 0.96° beam that is only useful at that range for large structures.

A field image from the FLS100 page: dam imaging at 1,300 m range. Large planar structures are exactly the target class where a 100 kHz long-range sector earns its place.
When a survey genuinely needs both
The two instruments coexist on survey vessels, ROVs and large AUVs more often than the procurement categories suggest. The pattern is consistent: an FLS forward for collision avoidance and target inspection, an MBES downward for the actual deliverable, and both running on the same vessel.
Coexistence has a cost that datasheets ignore. Acoustic interference between an FLS at 100–750 kHz, an MBES at 200–400 kHz and a side-scan sonar at 300–900 kHz is real, and it shows up as stripes, dropouts, or a false bottom. The standard mitigations are time synchronisation, a trigger hierarchy, and physical separation; our forum thread on acoustic interference covers the practical order to try them in.
There is also an image-quality trap on the multibeam side: an MBES produces soundings, and whether those soundings become a chart depends on motion compensation, sound-speed profile and tide. An FLS produces an image that looks convincing immediately, which is why it is easy to over-trust in a report. Neither output is self-validating.
Two surveys, two answers
Case one: inspect a harbor wall at 40 m stand-off, resolve 0.5 m features. Required angular resolution is 0.5 / 40 rad, about 0.72°. The 1.2 MHz mode of the dual-frequency unit gives 0.6° and 40 m of range, so the footprint at full range is 0.42 m — inside the requirement, with margin. The FLS100 at the same range would give 0.67 m, which is marginal for a 0.5 m feature and comfortable for a 1 m one. An MBES here would produce depth and backscatter of the wall base but not a face-on image, so it cannot answer the question at all.
Case two: chart 3 km² of a berth at 25 m depth. An MBES with a 130° swath covers about 4.3 × 25 m, or 108 m per line. At 2 m/s that is roughly 216 m² per second, so the area needs about 3.9 hours of line time before turning and overlap are counted. A forward-looking sonar cannot contribute soundings, so adding one does not shorten the survey; it adds obstacle detection along the same lines, which may be exactly what the vessel needs to run them safely.
Same water, same vessel, opposite instruments. The task sentence decides, not the specification sheet.
FAQ
Can a forward-looking sonar replace a multibeam echosounder for charting?
No. An FLS has no vertical reference to the seabed and reports ranges in the platform frame rather than soundings in a geodetic frame. It can show that a feature exists and roughly where; it cannot give you a depth with an uncertainty statement. Charting also needs motion, heading, sound speed and tide, none of which an FLS supplies.
Can a multibeam echosounder detect obstacles ahead of the vessel?
Only inside its swath, which points down and across, and only where the fan intersects an object. A bridge pier, a floating line or a boulder above the swath is invisible until the vessel is on top of it. That is precisely the gap a forward-looking sector fills, and why both exist on the same vehicle.
What angular resolution do I need for a 0.5 m target at 50 m?
About 0.57° or finer, from 0.5 divided by 50 in radians. A 0.6° unit meets it at 50 m; a 0.96° unit meets 0.5 m only inside about 30 m. Always convert the target size and range into a required angular resolution before comparing models, because "resolution" without a range is not a specification.
Why does one unit quote 120 m at 750 kHz and only 40 m at 1.2 MHz?
Because absorption rises with frequency. The higher band gives a finer beam — 0.6° against 1° — and a finer range resolution, 2.5 mm against 4 mm, at the cost of usable range. Dual-frequency units exist so one housing covers both a close-in inspection mode and a longer detection mode, with the mode chosen per task rather than per purchase.
FLS vs Multibeam Echosounder: Detection or Measurement?