Soft focus: the quiet listing killer VQS now gates on.

Blurry real estate photos rarely get caught in visual review, but under VQS they fail certification outright. Focus is a set-level gate, not a scored parameter: a soft frame does not drag the composite down, it fails the whole set. A 19-image set can average 98, with every image passing, and still earn no badge because two frames are soft.

Standard Vision Lab publishes VQS, the Visual Quality Score, an objective and versioned image-quality standard for real-estate photography. Every image in a set is scored 0 to 100 across the full set of measured checks. Focus is not one of those scored checks. It sits above them, as a gate on the whole set. That design choice is deliberate, and it matters for how you shoot and cull.

Why softness slips through visual review

Most listing photos are reviewed at the wrong size. Culling happens in a grid of thumbnails. Client sign-off happens on a phone or in a portal preview. At those sizes a slightly soft frame looks fine. Composition reads. Colour reads. Fine detail does not.

Softness also hides behind processing. Export sharpening can make a missed-focus frame look acceptable in a small preview while the full-resolution file stays soft. And human reviewers scan for what is easy to see, such as crooked verticals, blown windows and colour casts. Confirming sharpness in real estate photography means checking every frame at full magnification, and under deadline pressure that step is the first to go.

The result is familiar. A strong set ships with one or two blurry real estate photos in it, and nobody notices until the images are on a large screen or in print.

A gate, not a score

VQS measures a fixed set of checks per image across exposure, white balance, colour, geometry, artifacts and semantic checks, plus set consistency. Each check and each image carries a status: PASS at 90 or above, FLAG from 70 to 89, FAIL at 69 and below. Focus is treated differently. It is a set-level gate, not a scored parameter.

The reasoning is plain. A soft frame is not slightly worse than a sharp one. It is below the line the certification stands behind, and averaging it into a composite would let strong frames bury it. VQS already governs the SET score by the lowest image, on the principle that a listing is only as good as its worst photo. Set-level gates extend that principle: some faults fail a set outright, no matter what the rest of the numbers say.

The gate is also the honest position for now. VQS is a versioned standard, and promoting focus to a scored per-image parameter is the kind of change that arrives as a documented release, not a quiet adjustment. The version history lives on the releases page. Until then, the rule is simple: a soft frame fails the set.

The set that failed at 98

Here is the example worth memorising. A 19-image daytime set goes through the canonical runner. Every image passes. The mean score is 98. The lowest image scores 93. On composite alone, this set would certify comfortably; Certified runs from 90 to 99, and Platinum requires 100.

Two frames are soft. They trip the focus gate. The set FAILS and no badge is issued, because a FAILING set never certifies, regardless of composite.

That is the gate doing its job. The two soft frames are precisely the ones a buyer would enlarge, and the mark would have vouched for them. Refusing the badge protects what the badge means, for every photographer who holds one.

Shooting for sharpness

Sharpness problems are cheaper to fix on site than in review. A few habits close most of the gap.

• Work from a tripod for interiors. It removes camera shake from the equation and makes every other check easier to pass.
• Magnify on the camera, not just the laptop. Zoom into the focus point on the back screen before you move the tripod. A frame confirmed sharp on site never needs a reshoot.
• Watch shutter speed the moment you go handheld. Detail shots and tight spaces are where motion blur creeps in.
• Reshoot immediately when in doubt. A second frame costs seconds. A soft frame discovered later costs a return visit or a failed set.

Culling before the run

A certification run takes 12 to 20 original JPEGs from a daytime shoot, so the range gives you room to cut. Use it. Review every candidate at 100 per cent and drop any frame you would not defend at full size. Twelve sharp images certify. Twenty images with two soft frames do not.

Do not pad the set either. Every image is fingerprinted with SHA-256, and duplicates within a set or against earlier runs are removed automatically, so every run measures fresh work. The only way through the gate is a set of genuinely sharp frames.

The discipline transfers. Once checking at full magnification is part of your cull, blurry real estate photos stop reaching clients at all, certified or not.

Where to check your own work

If you want to see what the runner reports, the sample report shows every image and every measured value, and the demo lets you run a shoot. The runner is deterministic: the same image gets the same score, byte-identical, every time. Certified sets appear on the public verification register, and the plain-English guide to the whole standard is at VQS explained. Certification runs are available on a US$9.99 monthly subscription with one measurement run included each month, or as a US$129 single run with the badge valid for 12 months.