The number the box leaves off

The widest lens in this catalog, at 170 degrees, catches the car coming out of a side street that the narrowest at 125 clips off the edge of the frame — and it pays for that by spreading the same pixels across 45 more degrees of scene. Both halves of that trade are real, and only one of them is on the box. Front cameras here run from 13.7 pixels per degree up to 27.4, a 2.0× spread, every one of them sold as high definition.

The figure is a comparison index, not an optical measurement, and this site says so wherever it appears: manufacturers quote a diagonal angle almost universally, so dividing horizontal pixels by it understates every camera by roughly the same amount. That is exactly why the ordering survives the approximation. The long version, with the caveats spelled out.

Cameras are sorted by where the lenses point rather than by what the marketing calls them: 5 types, from a single forward channel through dual and triple setups to the cabin-facing cameras rideshare and delivery drivers buy and the mirror units that replace the rearview glass entirely. Channel count is the spine of the catalog because it changes the wiring, the card wear and the price bracket all at once.

53 cameras from 14 brands are tracked. A brand or a type gets its own page only once at least 2 cameras qualify, which is why 9 brand pages exist rather than 14 — a hub holding one camera is a dead end with a heading on it, and there are 5 type pages under the same rule.

Each figure was read off a manufacturer’s own product page where one exists, and off a specialist retailer’s spec table where the maker has since pulled the page. Where two sources disagreed, the manufacturer’s number won. Where nobody publishes a figure at all, the field stays empty and the page drops the row instead of guessing at it.

That last rule leaves visible holes, and they are worth looking at. 12 of the 53 cameras here will not tell you which image sensor is behind the lens, and 9 publish no operating temperature range at all — on a device whose whole job is to sit against glass in direct sun. A blank cell is a fact about the manufacturer, so it is left blank.

Pixels per degree, the power-cell flag and the parking tier are computed here; everything else on a camera page is a published specification. The derived ones are set in the retroreflective green borrowed from a licence plate wherever they appear, so you never have to wonder which is which.

The power-cell flag is the plainest of the three: 46 of 53 cameras store their shutdown charge in a supercapacitor, which cannot swell the way a lithium pouch cell in a baking car does, and it is shown with the maker’s temperature ceiling beside it because a supercapacitor camera whose maker will not state a ceiling is not obviously the safer buy. The parking tier exists because “parking mode” on a box covers at least four different behaviours, and the gap between recording the impact and recording the approach is the whole value of the feature. How each figure is calculated.

The “what owners say” section on a camera page summarises public owner discussion — the swollen cells, the cards that die in four months, the firmware update that broke buffered parking — read from a local archive of several years of it, and labelled as owner reporting rather than as testing. It is written for 37 of the 53 cameras here, which is every one with enough real discussion behind it to summarise honestly.

The other 16 say so on their own pages instead of showing an empty heading. And every camera’s description, pros and cons in this catalog were written from its specifications rather than from owner reports, so each page states that basis at the top. Nothing here puts “owners report” in front of something nobody reported.

Not one of the 53 cameras shows a rating, because not one clears the gate: a star only renders if it has a cited source that is not a retailer’s own aggregate. Reprinting a storefront’s stars would be both a licensing problem and a claim about owners nobody here surveyed. Sorting therefore offers front pixels per degree, approximate price, release year and name — each one naming the specification it orders by, because “best first” over an empty rating column would be a lie told by a dropdown.

Two people keep this site running. No office, no test lab and no staff beyond the two of us, and neither of us has driven a car with most of these cameras on the windshield. That is the reason the site is built around published figures and arithmetic anyone can check rather than around a verdict: the arithmetic is honest at this scale, and a fabricated test is not.

There is not one retailer link, buy button, price check or affiliate tag anywhere on dashcamlist, no advertising, and no arrangement of any kind with a camera manufacturer. No money moves, so nobody can buy a position. The 8 shortlists are each a filter and a sort over the same catalog file, stating the specification they are ordered by, and the 69 head-to-head pages are curated pairings a shopper would plausibly cross-shop rather than every possible combination. What the site collects is short for the same reason.

The catalog carries one date and it is on every page: specifications and prices were recorded on 2026-09-23. Prices move fastest — dash cam street pricing swings with bundle deals, which is why every price here is written with a tilde and is a band, not a checkout total.

Specifications drift more quietly. Manufacturers swap image sensors mid-production without touching the model number, and firmware updates change what a parking mode does months after launch. When a figure here stops matching the maker’s own page, that is a bug in this catalog, and the fix goes into the one file every page is built from. Corrections go here, or straight to chelsea@hagadornwebdev.com.