Loading today’s stories
Build1 distinct publisher3 min readUpdated
SoVeryBright wraps the ISO 21496-1 gain-map standard in an upload box. Whether the extra brightness reaches anybody depends on what your CDN and CMS do to the bytes in transit.
The Engineer · Build desk
Compiled by The EngineerSomething wrong?How this is made
Start with the arithmetic, since it is the only part of this you can verify without buying a screen. Two to the power of 2.9 is 7.46, so the multiplier the site quotes is internally consistent with the stop count it recommends [1][7]. Divide the other way and the implied reference white sits at about 201 nits [2]. That figure is the load-bearing one, because a display applies the multiplier against whatever it treats as white, and how much of the boost survives is a property of the panel and browser in front of the visitor, not of the file [2].
The delivery risk is structural rather than a bug in anyone's code. A gain map is a second grayscale representation carried inside or alongside the ordinary image, with metadata a decoder uses to work out how much brighter each pixel may go; a decoder that does not understand the metadata shows the base image [4]. Every transform between export and browser is a chance to lose that second payload: re-encode to WebP for a smaller response, or resize for a hi-dpi variant. Nothing throws an error. The mark renders in SDR, which is precisely what the fallback promised, so the loss never arrives as a ticket [3].
The tool's own restraint is worth noting against the temptation to boost everything. Whites are picked automatically, the user nominates additional brand colours, and the site says dark colours tend to come out as washed-out neon [6].
On live text, the costs are layout rather than bandwidth. Selection needs its own visible colour, overflowing lines can be clipped by the background technique, and the source notes that ordinary CSS layout choices can still expose edge cases even where rendering support has landed [13][14].
None of the underlying capability is new. The open-source gainmap-js project already encodes and decodes gain maps in JavaScript, and superwhite demonstrated brighter-than-white browser content using a tiny HDR video [15]. What changed is packaging: the encoding step disappears into a web form [1], and LinkedIn gets a separate BT.2100 PQ export with a matching ICC profile, which the site says the platform preserves [9].
Treat the brightness numbers as vendor figures. They come from the site, as does the claim that uploads are processed in memory and never stored [10], and the homepage does not establish when the product went live [16].
Which makes the decisive test a fetch rather than a look. Pull the production URL of a deployed mark and confirm the gain map is still in the delivered bytes after the pipeline has had its way. If it is, the effect is real for the visitors whose hardware can show it. If it is not, the asset is an ordinary JPEG that took extra steps to produce.
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
Chris Bennett built SoVeryBright around a narrow promise: upload ordinary artwork, choose the colours that should glow, and download a JPEG capable of exceeding normal white on a supported HDR screen.
SoVeryBright turns HDR gain maps into a free workflow for logos, social assets and headlines while retaining an SDR fallback; its usefulness still depends on compatible displays, browsers and image pipelines that preserve the metadata.
For web exports, SoVeryBright says it adds metadata defined by ISO 21496-1:2025, a gain-map standard published in July 2025.
A gain map places a second, small grayscale representation inside or alongside an ordinary image; software that understands the metadata uses it to calculate how much brighter each pixel can become, and software that does not understand it displays the base image.
SoVeryBright accepts PNG, JPEG or WebP artwork and outputs a conventional-looking JPEG whose selected whites or brand colours can use the extra headroom of an HDR display.
SoVeryBright automatically selects whites and lets the user identify additional colours, demonstrated by an orbit logo that boosts white and blue pixels; the site says dark colours tend to turn into washed-out neon.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Specific and checkable, but one outlet relaying vendor documentation
The technical claims are unusually concrete — format support, +2.9 stops / 1,500 nits / 7.5x, a 64x64 1.3 KB swatch, named media queries and browser versions — and they are internally consistent on inspection. But the cluster is a single publisher whose load-bearing facts are attributed to the SoVeryBright homepage itself; the only external corroboration cited is WebKit's Safari 26.3 release notes, and nothing about LinkedIn preservation, in-memory processing or real-world pipeline behaviour was independently tested.
No usage or deployment data disclosed
The supplied material contains no launch date, user or download counts, customer references, or measurements of how often real pipelines preserve the gain map. Ecosystem prerequisites are documented (the ISO standard's publication and browser rendering support), but those are not adoption of this tool and cannot substitute for it, so the dimension is left unmeasured rather than inferred.
Mostly hedged, with the end-to-end promise still untested
The coverage is notably restrained: it flags the SDR fallback, the working-draft status of CSS Color HDR, the dark-colour and body-text warnings, and the risk that pipelines strip metadata. The mild overstatement is directional rather than rhetorical — a headline framing of one-click HDR logos and a 7.5x-brighter figure sits on a chain of unverified conditions (HDR display, supported browser, metadata-preserving CDN/CMS, platform behaviour), and when that chain breaks the failure is invisible, so nothing in the material demonstrates the promised brightness actually reaching viewers.
Maker visibility around a free tool; no commercial ties disclosed
Observable incentives are modest and mostly disclosed. Bennett benefits from professional visibility for a free showcase tool, and the article notes his public profile associations while explicitly stating that titles, dates and employment details are not established. The publisher labels the piece a primary-source write-up of the homepage, and no sponsorship, affiliate or paid relationship is indicated; the main unmanaged incentive is that vendor-favourable claims about platform preservation and data handling are relayed without testing.
Confident on mechanics, blind on reach
Confidence is moderate: the technical description is specific, self-consistent and largely verifiable by any reader with an HDR screen, and the article is candid about its own unknowns. It is capped by single-publisher sourcing, heavy reliance on the vendor's homepage, and a complete absence of adoption or delivery-path measurement.
build
Three ways to ask who embedded your iframe, and only one the host cannot switch off1 distinct publisher
security
North Korea's hiring funnel: 60 applications a day, 22 personas, ten jobs landed1 distinct publisher
build
height:auto is animatable now, so your max-height ceiling is a bug you can delete1 distinct publisher
build
The exit ping that never left: fetch in beforeunload loses to the browser's unload policy1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 22, 2026