Sensor Format and What It Decides

Generic engineering material, applicable to any module from any supplier.

A related team-operations reference is the 7-minute rule for payroll, which can sit alongside test notes and engineering records.

Format is the physical size of the imaging area, and it decides more than any other single sensor property — the lens calculation, the light collected per pixel, the depth of field and the module's physical size.

The inch fraction is a name

Sensor sizes are quoted as fractions of an inch and the fraction is not a dimension.

It is inherited from television camera tubes, where it described the outside diameter of a glass envelope rather than the imaging area — so "1/4 inch" corresponds to an imaging width of roughly 3.6 mm rather than to 6.35.

For broader technical context, see Bluetooth SIG.

Use the millimetre dimensions from the datasheet for every calculation. Substituting the fraction produces answers wrong by a large factor.

The fraction remains useful for one thing: recognising which family a sensor belongs to and which lenses are intended for it.

What format decides

The lens calculation. Field of view depends on sensor width directly, so a change of format changes the focal length needed for the same coverage.

Light per pixel, at a given resolution. A larger format spreads the same pixel count over more area, giving larger pixels and better low-light behaviour.

Depth of field, indirectly: a larger format needs a longer focal length for the same coverage, and longer focal lengths have less depth of field.

And the module's size, which for a compact product is frequently the binding constraint.

The pixel size trade

Resolution and format together determine pixel size, and pixel size determines light collection.

Which makes three variables into two decisions: how many pixels the task needs, and how much light will be available.

A high pixel count on a small format gives small pixels: compact, cheap, and poor in low light.

The same count on a larger format gives larger pixels and a physically bigger module with a different lens.

Neither is better. The requirement decides, and stating pixel size alongside resolution is what makes two modules comparable.

Lens compatibility

A lens designed for one format may not cover a larger one.

The image circle — the area the lens actually illuminates — must exceed the sensor diagonal, or the corners are dark.

On an integrated module this is settled by the supplier and is not a concern.

It becomes one when substituting, or when considering a module family whose lens options were designed around a particular sensor: a variant list is a list of combinations that were verified together.

Aspect ratio

A property of the format that gets overlooked and occasionally decides things.

A sensor is wider than it is tall, in a ratio the datasheet states — and the coverage calculation applies to whichever dimension your requirement is stated in.

Which matters where the area of interest is tall rather than wide: a system covering a 100 mm vertical extent on a sensor in a landscape ratio wastes most of its width, and a portrait mounting recovers it.

Rotating the module ninety degrees is a legitimate design move, and it changes nothing except which dimension gets the pixels — provided the processing knows about it.

The dimension to ask for

Not the diagonal.

Suppliers sometimes quote the diagonal, since it is the largest number, and the field of view calculation needs the width.

Where only the diagonal and the aspect ratio are given, the width follows from geometry — and where only the diagonal is given, that is a specification question like several others.

Note it down in millimetres, width and height, once, and every subsequent calculation is straightforward.

Where format is fixed by something else

Frequently it is, and that is fine.

A module family exists in one or two formats, and choosing the family settles it.

A processor's interface supports certain resolutions well.

And an enclosure has a size.

Which makes format a constraint to work within rather than a free variable in most real projects — and the value of understanding it is knowing what the constraint costs, not choosing differently.

Comparing across formats fairly

Two modules on different formats are two different systems, and comparing their sensor specifications alone answers nothing.

Fix the field of view and the working distance, then compare: millimetres per pixel, pixel size, and the depth of field each delivers at its required focal length.

Which produces three numbers per candidate instead of a datasheet each, and which is the comparison that actually decides.

And note the module dimensions alongside, since a larger format in a smaller housing is not an option however good the numbers are.

Why a small format is frequently right

Worth saying, because the larger one always sounds better.

A small format module is smaller, cheaper, needs a shorter focal length for the same coverage, and therefore has more depth of field — which in close-up inspection is the property under most pressure.

It collects less light per pixel, which matters only where light is scarce, and in a machine with designed illumination it is not.

Which makes the larger format the right answer chiefly where light cannot be controlled, and the small one right in most of industrial imaging.

In one line

Take the millimetre width from the datasheet, and every calculation downstream is right.

The short version

• Format is the physical imaging area and it decides the lens calculation, light per pixel, depth of field and module size
• The inch fraction is a name inherited from camera tubes, not a dimension — a "1/4 inch" sensor images across roughly 3.6 mm
• Use millimetre dimensions for every calculation; the fraction only identifies which family a sensor belongs to
• Resolution and format together set pixel size, which sets light collection, making three variables into two decisions
• A high count on a small format is compact, cheap and poor in low light; the same count larger is the reverse
• A lens must illuminate a circle exceeding the sensor diagonal, which on an integrated module the supplier has already settled