Every ScreenDetect device page is a source-backed screen profile: a plain-English read on what a display is good for, what tends to bother people, and what to check before you buy, return, or repair it. It is built from published specifications, independent measurements, and documented owner experience, and it always keeps those three apart. This page explains exactly how one is written, who stands behind it, and what we refuse to fake.
What a screen profile is, and is not
A profile is not a hands-on lab review. We do not receive review units, and we do not run our own benchmarks. When a figure comes from someone else's testing, we say so and point to them, rather than restating a measurement as if it were ours.
That distinction matters. Most device pages online blur the line between a manufacturer spec sheet, a reviewer's measured result, and a writer's guess. We keep those three separate on purpose, because a buyer deciding on a screen deserves to know which claims are solid and which are still open.
Who writes and reviews these
ScreenDetect is written and edited by Jacob Dymond, founder and editor, who also builds the on-screen diagnostic tools the profiles link to. Every profile is drafted, source-checked, and reviewed against the standards on this page before it is published, and the same person is accountable for corrections.
We stay in one lane deliberately: we are not a review site chasing scores, and we do not pretend to independent lab access we do not have. Our expertise is turning scattered, uneven screen information into a decision you can act on, and telling you plainly when a claim is not settled. You can read more about the person and independence behind the site on the about page.
How a profile gets written
Every profile follows the same repeatable path, and we do not skip steps to fill a page:
- Research the specific model. We gather the exact panel, size, resolution, refresh, brightness, and coating details for that model, not a close sibling.
- Tier the sources. Each fact is assigned to a source tier (below) so its weight is explicit before anything is written.
- Draft in plain English. We explain what each detail means for real use, not marketing language.
- Label confidence. Every meaningful claim is marked official, review-reported, needs context, or verify-yourself.
- Route to the inspection lane. Each likely symptom is linked to the specific ScreenDetect test that helps confirm it.
- Review and publish. The draft is checked against these standards, then published with visible author and date.
- Revisit when it matters. Material changes are logged, not quietly edited in.
How we use software and AI
We build our own tools, including AI, to do the parts a person is slow at: pulling the exact specifications for a specific model, gathering independent measurements scattered across manufacturer pages, lab reviews, and support docs, and organizing that raw material into a structured first draft. This is about speed and coverage in the fetching and assembly stage, not about handing over judgment.
From there a person takes over. A human decides what each number actually means, which claims are solid enough to state, what needs a caveat, and which ScreenDetect test to send you to. We do not publish machine output unread, and we never invent a spec or a measurement to fill a gap. The confidence labels below exist for exactly this reason: they keep an AI-assisted draft from quietly passing off a guess as a fact. If a detail cannot be sourced, it stays out or gets marked verify-yourself.
Where the information comes from
We work from a clear source hierarchy, strongest first:
- Manufacturer and official pages for confirmed specifications, model numbers, and supported modes.
- Independent lab and review outlets for measured behavior the manufacturer does not publish, such as measured brightness, response time, or flicker.
- Support docs, retailer listings, and owner reports for setup quirks, common complaints, and return-window patterns.
When sources disagree, we do not silently pick one. We note the disagreement and, where it matters, send you to check the specific thing yourself on your own unit.
How we label confidence
Not every fact deserves the same weight, so we mark them:
Swipe sideways to compare columns.Scroll table to view all columns.
| Status | What it means |
|---|---|
| Official | Confirmed on a manufacturer or first-party source |
| Review-reported | Measured by an independent outlet, credited to them |
| Needs context | Real but varies by unit, region, firmware, or setting |
| Verify yourself | Best confirmed on your own screen, and we show you how |
Anything we cannot stand behind, we leave out or move into the "known, needs context, and unknown" section rather than guessing to fill a gap.
A worked example. Samsung lists the Galaxy S26 at up to 2,600 nits peak brightness. An independent lab measured roughly 1,806 nits in HDR under its own test window. Both are true, and they are not the same claim: we mark Samsung's 2,600-nit figure as an official peak claim, and the lab's number as review-reported under a stated method. We never merge them into a single confident "2,600 nits" and we never treat a peak claim as sustained full-screen brightness.
How we handle disagreement and uncertainty
Two credible sources often report different numbers because they used different methods, windows, or units. When that happens we show the range and the reason rather than averaging them into a false single figure. When a detail cannot be settled from the outside at all, we say it is unknown and hand it to you as a verify-yourself check, because your own eyes on your own unit during the return window are the honest tiebreaker.
Corrections and updates
Screens change after launch: firmware adjusts behavior, new lab data lands, prices move. When something material changes, we update the page and record what changed in its update log, rather than quietly editing and bumping a date. If a page has not been re-checked recently, the visible "last updated" date tells you exactly where it stands. Spot something wrong? Email hello@valiancelabs.com. Any confirmed mistake is corrected within 24 hours, usually much sooner.
How ScreenDetect is funded
Some pages may carry clearly labeled affiliate links or ads. No manufacturer, retailer, or advertiser pays for, previews, or approves what a profile says, and funding never changes a confidence label, a caveat, or which test we send you to. If a screen has a real weakness, it stays in the profile regardless of any commercial link on the page.
What we add that a spec sheet cannot
A spec list tells you the numbers. It does not tell you what to inspect the day a screen arrives, which faults are easy to confuse, or when a quirk is normal versus a reason to return. That is the part we own: the inspection lane. Each profile routes you to the specific ScreenDetect test for the symptom in question, so the page ends in an action rather than a shrug.
Coming later: real inspection data
ScreenDetect's tests are run by real people on real screens. In time, and only once the sample size is large enough to be honest about, we plan to fold anonymized, aggregated inspection results into these pages so buyers can see what owners actually find, not just what a spec sheet promises. We would rather ship that late and accurate than early and misleading.