Short answer
A phone photo is usually built from several frames, not one. The camera merges a burst to cut noise and extend dynamic range, then white-balances, converts colour, compresses the tonal range locally so shadows and skies both hold detail, and applies taste-driven hue and saturation tweaks. Some phones add scene or subject detection. The published details come mainly from Google's research papers; other manufacturers describe their pipelines only in general terms, so specifics vary and are often not public.
Phone sensors have tiny pixels that collect little light, so single frames are noisy and clip easily. Google's HDR+, introduced in 2014 and described in a 2016 paper, captures a burst of raw frames at the same deliberately short exposure so highlights do not clip, aligns them, and merges them into one image with cleaner shadows and more tonal depth. Night Sight extends the approach with longer exposures in dim light. Merging changes colour as well as noise: averaged frames show less colour speckle in shadows, so darker areas can be brightened without turning blotchy.
A merged frame holds more contrast than a screen can show, so the pipeline compresses it. HDR+ uses local tone mapping, brightening shadows while keeping local contrast, then dehazing, then a global S-shaped curve with the sRGB transfer function. These choices change perceived colour: lifting shadows lightens and desaturates dark colours, and a strong S-curve saturates midtones. Google's engineers describe the goal as natural-looking pictures of scenes with strong contrast, and for Night Sight describe deliberately keeping shadows dark so that a night scene still reads as night.
The HDR+ pipeline includes a step that is plainly aesthetic: hue-specific adjustments that shift bluish cyans and purples towards light blue and raise the saturation of blues and greens, to make skies and vegetation more appealing. Other manufacturers describe similar ambitions in marketing terms without technical detail. Some phones also identify scene types or subjects — skies, food, faces — and process them differently, and some let users choose a style that biases tone and warmth. The practical point is that phone colour is designed, not measured, and designs differ between brands and between software updates.
Details for manufacturers other than Google are summarised here only in general terms, because they are not published at the level of an academic paper.
Phone pictures tend to be more colourful than the scene, and the surplus is added in several places, not one. The colour-correction matrix sets overall chroma, and a camera-testing vendor notes that cameras are often tuned so that a test chart comes out at 110 to 120 per cent of its true chroma, treating anything above 120 as excessive. A contrast curve applied to each channel adds more in the midtones. Hue-specific tuning then raises chosen colours further, blues and greens in the published Google pipeline. Editing tools add a fourth kind, usually called vibrance, which lifts only the weakly saturated areas so that already vivid colours are not pushed too far. Last comes the screen: Pixel phones ship in an 'Adaptive' colour mode that Google describes as showing the most vivid colours, with a separate 'Natural' mode for the most accurate. A photo judged on the phone that took it has therefore been boosted twice.
The HDR+ authors name the limit themselves: too much range compression together with too much saturation gives a flat, cartoon-like picture.
When colour accuracy matters more than appeal — photographing a paint sample, a product, a skin condition for a clinician, a fabric for a supplier — reduce the processing the phone can apply. Use a raw or 'pro' mode if available; turn off scene optimisers and filters; include a grey card or white sheet in the frame; lock exposure on a mid-tone rather than letting the phone brighten the scene; and avoid mixed lighting. Then compare on a calibrated screen. You are still seeing an estimate, but a less opinionated one.
| Stage | What it does to colour | Documented by |
|---|---|---|
| Colour-correction matrix | Sets overall chroma; cameras are often tuned above 100% of a test chart's true chroma | Imatest documentation |
| Global contrast curve | An S-curve applied per channel raises saturation in the midtones | HDR+ paper (the curve); Colourwise reading (the effect) |
| Hue-specific tuning | Raises the saturation of blues and greens and shifts sky hues | HDR+ paper |
| Vibrance-type control | Lifts only weakly saturated areas, for a subtler increase | Grading-software documentation; phone makers do not say whether they use it |
| Screen colour mode | A vivid default mode displays every image more saturated than an accurate mode would | Pixel support documentation |
Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.
FactStrong evidence
Google's HDR+ captures bursts of constant-exposure raw frames, set low enough to avoid blown highlights, and aligns and merges them to reduce noise and increase dynamic range; it was introduced in 2014.
Source: Burst photography for high dynamic range and low-light imaging on mobile cameras (Hasinoff et al., ACM Transactions on Graphics 35(6), SIGGRAPH Asia 2016); Night Sight: Seeing in the Dark on Pixel Phones
FactStrong evidence
The published HDR+ pipeline includes hue-specific adjustments that shift bluish cyans and purples towards light blue and increase the saturation of blues and greens to make skies and vegetation look more appealing.
Caveat: Describes Google's pipeline as published in 2016; later versions and other brands may differ.
FactModerate evidence
Night Sight's designers added an S-curve to its tone mapping and deliberately kept dark areas dark so that low-light photographs still convey that they were taken at night.
Colourwise interpretationLimited evidence
Across manufacturers, smartphone colour is shaped by aesthetic tuning as well as accuracy, and those choices are rarely published in detail.
Based on: Inferred from the detailed Google publications and the absence of comparable technical documentation from most other manufacturers; not a survey of devices.
Caveat: A generalisation; some manufacturers publish more than others and pipelines change with software updates.
FactModerate evidence
Camera-testing documentation defines mean camera chroma as the average chroma of a camera's rendering of a ColorChecker relative to the reference values, reports that boosts to 110–120 per cent are common in compact digital cameras, and treats more than 120 per cent as excessive.
Caveat: The testing vendor's stated experience, not a published survey, and it refers to compact cameras. No comparable published figure for current phones was found.
FactModerate evidence
Pixel phones from the Pixel 3 onwards default to an 'Adaptive' screen colour mode that Google describes as showing the most vivid colours, and offer a 'Natural' mode described as the most accurate.
Caveat: One manufacturer's support page; modes and defaults change between models and software versions.
FactStrong evidence
The HDR+ authors limit how far their pipeline compresses dynamic range, stating that excessive range compression and saturation produce a flat, cartoon-like rendition.
Colourwise interpretationLimited evidence
A phone photo viewed on the phone that took it is usually saturated twice, once by the camera's tuning and again by a vivid default screen mode, so judging colour accuracy requires an accurate display mode or a calibrated screen.
Based on: Combines the documented saturation boosts in camera pipelines (HDR+ paper, Imatest documentation) with Google's description of its default screen mode. Not measured across devices, and it assumes the vivid mode is left on.
Caveat: Depends on the maker, the model and whether the user has changed the screen mode.
Reviewed 29 September 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.