The "1-inch sensor" printed on a spec sheet is not 25.4mm across the diagonal. It is roughly 16mm. The label does not describe the sensor's physical size at all — it inherits the outer glass diameter of broadcast vidicon camera tubes used from the 1950s through the 1980s, where the light-sensitive area inside measured about two-thirds of that diameter. That two-thirds relationship is where 16mm comes from.
Which means reading the gap between 1/1.3" and 1/1.4" as "0.1 inch" tells you nothing. Converting the label into sensor area is what reveals how different two sensors actually are. Every figure below is derived from published specifications and arithmetic.
Two constants convert the label into millimeters
The industrial camera convention is stated cleanly by Tucsen: for formats of 1/2" or larger, diagonal (mm) = format in inches x 25.4 / 1.5875. Below 1/2", the divisor changes to 1.4111. The first formula makes one inch equal 16.00mm; the second makes the equivalent reference 18.00mm.
The split also traces back to the tubes. In smaller tubes the glass wall took up proportionally less room, so the usable imaging area occupied a larger share of the diameter — and that ratio hardened into convention. Apply a single proportion across all formats and you will be off by more than 10% at the small end.
For a 4:3 sensor, a diagonal d gives width 0.8d, height 0.6d, and area 0.48d². Running the common smartphone formats through that:
| Format | Divisor | Diagonal (mm) | 4:3 width x height (mm) | Area (mm²) | vs 1-inch |
|---|---|---|---|---|---|
| 1" | 1.5875 | 16.00 | 12.80 x 9.60 | 122.9 | 100% |
| 1/1.3" | 1.5875 | 12.31 | 9.85 x 7.38 | 72.7 | 59% |
| 1/1.4" | 1.5875 | 11.43 | 9.14 x 6.86 | 62.7 | 51% |
| 1/1.7" | 1.5875 | 9.41 | 7.53 x 5.65 | 42.5 | 35% |
| 1/2.55" | 1.4111 | 7.06 | 5.65 x 4.24 | 23.9 | 19% |
| 1/3.06" | 1.4111 | 5.88 | 4.71 x 3.53 | 16.6 | 14% |
Two things stand out. The 1/1.3" and 1/1.4" formats look adjacent on paper but differ by about 16% in area — 72.7mm² against 62.7mm². And a flagship main sensor at 1/1.3" has more than three times the area of a budget 1/2.55" unit. The spacing of the labels badly compresses the real gap.
For scale: 35mm full frame is 36 x 24mm, or 864mm². Even a smartphone "1-inch" sensor is one-seventh of that.
Checking the math against real sensors
The conversion can be verified in reverse. Multiply the horizontal pixel count by the pixel pitch to get the width of the imaging area, divide by 0.8 for the 4:3 diagonal, then multiply by 1.5875 and divide by 25.4 to recover the format label.
A 200MP sensor is 16,384 x 12,288 pixels; a 50MP sensor is 8,192 x 6,144. Feeding in Samsung's ISOCELL HP2 (1/1.3", 0.6μm) and HP9 (1/1.4", 0.56μm) as listed by LensXP, plus Sony's IMX989 (1", about 1.6μm) as reported by SamMobile:
| Sensor | Published format | Pixel pitch (μm) | Horizontal pixels | Computed width (mm) | Computed diagonal (mm) | Recovered format |
|---|---|---|---|---|---|---|
| ISOCELL HP2 | 1/1.3" | 0.60 | 16,384 | 9.83 | 12.29 | 1/1.30" |
| ISOCELL HP9 | 1/1.4" | 0.56 | 16,384 | 9.18 | 11.47 | 1/1.40" |
| Sony IMX989 | 1" | 1.60 | 8,192 | 13.11 | 16.38 | 1/0.98" |
HP2 and HP9 match their published labels to two decimal places — confirmation that the 16mm rule is not a loose approximation. The IMX989 comes back as 1/0.98", about 2.4% larger than the nominal 16.00mm, meaning Sony's "1 inch" runs slightly generous. That kind of spread is normal in the convention. The label is a classification, not a measurement.
A sensor format is not a name derived from measuring anything. It is a classification number inherited from the camera-tube era.
What really splits is the area of a single pixel
In low light, what matters is less the total sensor area than how much light one pixel collects. Square the pitch to get per-pixel area: HP2's 0.6μm gives 0.36μm², while the IMX989's 1.6μm gives 2.56μm² — a 7.1x gap. Total sensor area differs by only 1.7x; the per-pixel figure blows out because HP2 packs four times as many pixels into that area.
High-resolution sensors close the gap by merging pixels. Samsung Semiconductor describes its Tetra²pixel scheme as merging up to 4 or 16 neighboring pixels. On a 200MP 0.6μm sensor that means 1.2μm at 50MP, or 2.4μm at 12.5MP.
| Mode | Effective pitch (μm) | Per-pixel area (μm²) | Output resolution | vs IMX989 native 1.6μm |
|---|---|---|---|---|
| 0.6μm, full resolution | 0.60 | 0.36 | 200MP | 0.14x |
| 4-pixel merge | 1.20 | 1.44 | 50MP | 0.56x |
| 16-pixel merge | 2.40 | 5.76 | 12.5MP | 2.25x |
Merge sixteen and each output pixel covers more area than a native pixel on the 1-inch sensor — but resolution drops to 12.5MP. That trade is why a high-resolution sensor's low-light files come out smaller than its daylight files.
Read a spec sheet in this order
- Pixel pitch before format label — pitch plus pixel count yields sensor size by calculation. The reverse does not work.
- Remember the 1/2" boundary — comparing 1/1.7" and 1/2.55" on a single proportion is wrong; the reference is 16mm on one side and 18mm on the other.
- Check the binned output resolution — "200MP" and the 12.5MP file actually saved are numbers from different layers.
- Look up the telephoto and ultrawide separately — a 1/1.3" main paired with 1/3"-class companions is common, a gap of more than 4x in area.
- Do not trust the spacing of the labels — 1/1.3" to 1/1.4" is 16% of area; 1/1.4" to 1/2.55" is 2.6x.