A guide
There is no standard for this category, so here are the five checks we apply to our own work — written so you can apply them to anybody's.
Updated 9 September 2026
Start here
Five things you can check in about thirty seconds, without any equipment.
Most baby sensory video is made by eye, which is a reasonable way to make something and a poor way to check it. These are the checks we apply to our own videos, written so you can apply them to anybody's — including ours.
One
Pause on a still frame and look at the darkest and lightest parts. You want something close to true black against something close to true white. Mid-grey against off-white gets described the same way and is a much weaker signal for a young baby, whose contrast sensitivity is a fraction of an adult's.
A quick trick: screenshot the frame and turn your phone's brightness right down. A genuinely high-contrast pattern stays clearly readable. A soft one collapses into mush, which is closer to what a six-week-old sees at normal brightness.
For reference, our own videos measure 18.96:1 by the standard contrast formula, against a theoretical maximum of 21:1 for absolute black on absolute white.
Two
This is the check almost nobody makes, and it matters more than contrast. A pattern finer than a baby can resolve is not a stimulus at all — it averages out into flat grey, however bold the colours are.
Newborn acuity is around 20/400 on behavioural measures. Practically: if a pattern element is smaller than a large coin held at the screen, a very young baby probably cannot separate it from its neighbour. Fine stripes, small polka dots and busy detailed scenes tend to fail this even when they look striking to you.
Bigger and simpler is genuinely better in the first months, which is the opposite of what looks impressive in a thumbnail.
Three
Smooth pursuit — tracking a moving object with the eyes rather than in jumps — begins to develop around six to eight weeks and only approaches adult quality at four to five months. Before that, a baby watching something move quickly is not tracking it; they are losing it and reacquiring it.
Watch a moving element and try to follow it with your own eyes without moving your head. If you find yourself flicking rather than gliding, it is far too fast for a newborn.
Our own ceiling is about 8 degrees of visual angle per second for the 0–3 month videos, rising to roughly 13 by the oldest age band. More useful than the ceiling, though, is how often nothing moves at all: in most of our newborn videos the median motion is zero.
Four
Rapid full-screen brightness change is the property associated with photosensitive seizure risk, and the widely used limit is three flashes in any one second. Plenty of baby sensory content sits well inside that limit. Some, especially the very fast-cut compilations, sits closer to it than you would expect.
If a video makes you blink or feels busy to watch as an adult, that is a reasonable signal to trust. There is a longer piece on this here.
Five
Babies settle into a pattern and then gradually lose interest in it, and that whole arc is the useful part. A video that cuts every few seconds never lets it happen, and mistakes novelty for engagement.
Time a few scenes. For a newborn video, something in the range of twenty seconds to a minute and a half per pattern is generous rather than slow. Ours run 20 to 90 seconds at the youngest tier.
And
View count.
A video with millions of views has demonstrated that it holds attention, which is not the same as being well made for an infant visual system. Fast cuts and busy detail hold attention very effectively. They just do it by being startling.
If it is useful, our full specification is published with the measured values and the research each figure comes from, so you can hold our videos to the same five checks.