Reference
Every filter, in one place
Filters searches on six things: the colour that dominates a frame, how much of the frame it owns, how busy the picture is, what is in it, whether you have favourited it, and when it was taken. Any three of them can run at once.
Colour dominance
Twenty-four colours: twenty-one hues around the wheel, plus three neutrals that sit outside it. For each photo Filters measures the share of the frame every colour holds, so a query is a threshold rather than a guess.
Twenty-one hues
- Red
- Red-Orange
- Orange
- Amber
- Yellow
- Lime
- Green
- Spring
- Mint
- Teal
- Cyan
- Sky
- Azure
- Blue
- Indigo
- Violet
- Purple
- Magenta
- Pink
- Rose
- Crimson
Three neutrals
A snowfield, an overcast harbour and a night shot have no dominant hue — asking which colour owns them returns nonsense. They get their own filters instead, measured on lightness rather than hue.
- White
- Grey
- Black
The dominance threshold
Every colour filter carries a percentage you can move. It reads as a sentence — Orange ≥ 60% — and it is the difference between a photo that has orange in it and a photo that is orange.
Low thresholds are for finding an accent: a red coat in a grey street, a single yellow door. High thresholds are for finding a mood: the whole frame committed to one colour. Most useful queries sit somewhere between 35% and 70%, and because the match count updates live you can feel where the useful band is for your library rather than being told.
Same library, one slider
- ≥ 20%
- 1,204
- ≥ 40%
- 631
- ≥ 60%
- 188
- ≥ 80%
- 27
Illustrative counts from one library — yours will fall differently, which is the point of showing the number live.
Second-colour detection
The dominant hue tells you what a photo is mostly made of. The runner-up tells you what makes it interesting.
Two readings per photo
Filters reports the top two colours with their shares — 32% green + 10% purple. The green is a field; the purple is the lavender growing in it. Two photos can both read 32% green and be completely different pictures, and the second number is usually where that difference lives.
Why it changes a query
When you are assembling a set that has to hang together, matching on the accent matters as much as matching on the base. A run of green frames all carrying a little purple reads as deliberate. The same greens with random accents reads as a camera roll.
Busyness
How much visual detail a frame carries — blue sky at one end, dense foliage at the other. It is a property of the composition, not of the sensor: a clean shot is a simple shot, whether it was taken on a phone or a medium-format back.
Clean
Wide, uncluttered frames — an empty sky, a painted wall, a plate on a table. The shots that give a caption room to breathe.
Mid
A clear subject with a settled background. Portraits, street corners, most of what a camera roll is actually made of.
Busy
Dense visual detail edge to edge — foliage, market stalls, crowds, patterned interiors. Loud frames that need quiet neighbours.
Busyness earns its keep in combination. “Blue” returns skies and swimming pools and crowded festival shots in equal measure; “Blue and Clean” returns only the ones with room to breathe.
Subjects
Nine categories, recognised with Apple's on-device Vision framework the first time Filters reads a photo, then cached. The work happens once; every query after that reads the cache.
- People
- Dogs
- Cats
- Plants
- Food
- Vehicles
- Birds
- Nature
- Buildings
Subject recognition is good rather than perfect — treat it as a way to narrow a colour query, not as a strict inventory. Colour and busyness are measured from pixels directly and carry no such caveat.
Favourites and date ranges
The unglamorous filters that make the others usable.
Favourites
Filters reads the favourite flag you have already set in Photos. It never writes one. Combined with a colour, it turns “my good photos” into “my good photos, in green” — which is a shortlist you can actually use.
Date ranges
Any span, down to a single day. Useful on its own for a trip, and useful as a third circle when a colour query returns more than you can look through. On the free tier the range is capped at the last twelve months; Pro removes the ceiling.
Combining three at once
Filters are not applied one after another down a list. Up to three become overlapping circles, and the centre carries the live match count — so you can watch a query tighten as you move a threshold instead of running it and hoping.
Plated food shot against a plain surface — the frames that carry a caption.
Dense foliage and garden shots, the texture pictures in a nature set.
Your best minimal blue frames — skies, water, painted walls.
Recent bright, airy shots: the raw material for a light-toned grid.
Suggestions when nothing matches
Stack three tight filters and you will eventually ask for something your library does not contain. An empty result screen is a dead end; it tells you nothing about which of the three conditions was the problem.
So when a query returns zero, Filters offers adjustments that have results behind them — each one already counted, so tapping it never lands you on another empty screen.
No matches for Crimson ≥ 70% · Busy · Food
- Lower the threshold to 45% — 23 photos
- Try Rose instead of Crimson — 41 photos
- Drop the busyness filter — 12 photos
- Widen the dates to the full year — 8 photos
Your library is already sorted by colour. It just can’t say so yet.
Filters is finishing up for iPhone, iPad, and Mac. Free covers the last twelve months of your photos; Pro opens the rest for $19.99 once or $9.99 a year.
