A white plate looks yellow in your dinner photo. Your friend’s face looks cold beside a window. In a video, the room seems to change color when you turn toward the sofa.
White balance is one possible explanation. Your phone is trying to account for the color of the light, and that estimate does not always match the look you want. But changing white balance will not fix every color problem.
Quick answer: three things to try first
White balance adjusts the color balance of an image to compensate for the illumination. Start with these three checks before opening a Pro mode:
- Simplify the light. If a window and a room lamp both illuminate your subject, try switching off the lamp or moving the subject into one main source.
- Check Auto and remove an unwanted look. If you previously selected a WB preset, return it to Auto. Compare without a filter or style you intentionally added, then take another photo.
- Check a neutral reference in the saved image. Look at a gray or white object under the subject’s light, rather than judging only the food or someone’s skin. If the whole image looks warm mainly on your screen, also check your display settings.
These are comparison steps, not a test that certifies accurate color. If the result still has a shared color cast, an available WB control or a modest editing adjustment may help. If different parts of the scene have different casts, start with the lighting.
| What you see | Possible explanation | What to try | What this cannot prove |
|---|---|---|---|
| Neutral objects look yellow or orange | Warm illumination may be under-corrected, or a warm look may be intentional. | Try one dominant light source. If WB controls are available, move toward a cooler result and compare neutral objects. | A warm photo alone does not prove AWB failed. |
| Neutral objects look blue | Cool illumination may be under-corrected, or a fixed WB setting may not suit the scene. | Check for an old manual setting or preset. Try Auto, or adjust WB toward a warmer result. | Do not interpret every blue object as a WB problem. |
| Neutral areas look green or magenta | A tint imbalance may remain; the light spectrum or color rendering may also contribute. | If your editor offers Tint, try the opposite direction gently. Compare the same subject under a different light. | Making gray neutral does not guarantee that every object color is correct. |
| Color changes between shots or while reframing | AWB may react to different scene information; the lighting may also have changed. | Keep the scene, lighting and mode stable. If available, check the result, then lock WB or use a fixed setting. | WB lock preserves a setting; it does not establish color accuracy. |
| Color changes after switching lenses or modes | The camera or processing path may differ, and the field of view may include different lighting. | Compare one camera and mode first. Record the model, app, software and lighting before blaming AWB. | A color difference alone does not identify AWB as the cause. |
| One side looks warm and the other cool | Different areas may be receiving light from different sources. | Try switching off a competing lamp or moving the subject into one main source. | One global WB adjustment may not neutralize both areas. |
| The same saved photo looks warm mainly on your screen | Display settings may affect how you see the image. | On supported Apple devices, compare with Night Shift and True Tone temporarily off; also compare the original file on another display. | This check does not calibrate your screen or prove the photo is accurate. |
What white balance actually does
The light reaching your camera carries information about both the source illuminating the scene and the objects reflecting it. A neutral object can therefore record a yellowish or bluish cast.
Auto White Balance, or AWB, estimates how much of that color comes from the illumination and applies compensation. It has to distinguish a colored object from a neutral object under colored light. Google describes this ambiguity in its explanation of Night Sight: a camera cannot simply assume that a blue-looking patch has an inherently blue surface. 3
For the wider capture pipeline, see phone camera specs explained.
That is why a photograph can disagree with your memory of a room. Your impression of the scene and the camera’s chosen balance are not interchangeable.
It helps to separate three goals:
- Neutral references: white or gray objects have little visible tint.
- A convincing appearance: the photograph resembles how you remember the scene.
- An intentional mood: the warm restaurant lighting or sunset remains visible.
For a product listing, the first goal is a useful starting point. For a candlelit portrait, removing all warmth may make the picture less appealing. A neutral reference also cannot establish that every other color is reproduced correctly.
Color temperature vs tint: two different adjustments
Temperature adjustments move the image between cooler blue and warmer amber. Tint deals with the green–magenta direction. If a wall looks greenish, pulling the temperature slider toward blue may simply leave you with a cooler green wall. 2
Why Kelvin can seem backward
The color temperature of a light source and the Kelvin value selected for WB describe different things.
A lower color temperature generally describes redder, warmer-looking light; a higher one describes bluer light. A WB setting tells the camera what illumination to compensate for. With the scene unchanged, selecting a higher WB Kelvin value generally produces a warmer result; selecting a lower one produces a cooler result. 1
So, when you are adjusting WB, a blue cast usually calls for a move toward a warmer result. A yellow cast calls for a move toward a cooler result. Check the preview and saved image as you adjust.
Not every Warmth or Temperature slider uses Kelvin. Adobe Camera Raw, for example, uses different units for rendered JPEG/TIFF adjustments and RAW temperature settings. 2 Treat a slider’s labels as directions until you know what its numbers represent.
For LED lighting, the Kelvin specification is commonly a correlated color temperature, or CCT. It is not the physical temperature of the lamp. Two sources with similar CCT can still differ in their green–magenta balance or spectral output. 8 There is no single Kelvin recipe that guarantees natural skin, white plates and accurate clothing colors under every lamp.
Why Auto White Balance gets confused
AWB faces an estimation problem. A scene dominated by one color, dim illumination or strongly colored light can leave fewer reliable clues about the source. Google’s historical Pixel 3 Night Sight documentation describes a learning-based approach developed for difficult low-light balance; that example does not establish how every current phone handles the same scene. 3
For motion, capture duration and multi-frame limits, see when to use phone Night mode.
A practical inference is that changing the composition can also change the information available to AWB. If you turn from a neutral wall toward a brightly colored sofa, the estimate may shift. It need not do so on every device.
If that happens during a video, try a short comparison under stable lighting: start on the same framing, pan through the same movement, then repeat with a WB lock if your camera offers one. Check whether the change becomes less noticeable.
A color shift after switching lenses or modes is harder to diagnose. You may have changed the field of view, the camera or the processing path. Keep one camera and mode fixed before attributing the difference to AWB. Without matched samples, “the color changed” is an observation, not proof of a particular fault.
Why mixed lighting is difficult to fix
Imagine a person beside a window with a warm room lamp lighting the other side of their face. Adjusting the whole image until the lamp-lit side looks neutral can leave the window-lit side too cool. Favoring the window side can preserve warmth on the other.
A single global balance may not neutralize both areas. Research on mixed-illuminant correction addresses this with spatially varying treatment, so it would also be wrong to say every camera always applies one simple setting everywhere. 7
For your own photo, try changing the setup before chasing a slider. Turn off a competing lamp, move the person, or accept some warmth in the background while prioritizing the subject. These are practical consequences of the mixed-light problem, not guaranteed fixes.
| Lighting situation | What WB can help with | Better first step | Remaining limit |
|---|---|---|---|
| One reasonably consistent source | A broadly shared yellow/blue or green/magenta cast. | Try Auto, then an available WB adjustment; check a neutral area under the subject lighting. | Correcting a neutral area does not ensure every color is reproduced accurately. |
| Window light plus a warm room lamp | You can favor the subject or one region. | Reduce the competing source or move the subject into the preferred light. | A global WB can improve one area while leaving another visibly tinted. |
| Several LEDs that appear similar | Temperature and tint may reduce a shared cast. | Compare the subject under a single source before making larger edits. | Matching CCT does not guarantee matching spectra or color rendition. |
| Sunset or deliberately colored lighting | WB can shift the balance toward neutrality or retain a mood. | Decide whether you want neutral references or the atmosphere of the light. | A warmer result can be intentional; neutrality is not the only valid goal. |
| Moving between different light sources | Auto can update the estimate; manual WB can hold a selected look. | Recheck a fixed setting or lock after moving into new lighting. | A previously suitable locked WB may no longer suit the scene. |
LED scenes can present another limit. Neutralizing a gray patch does not necessarily make every colored object look right: lighting spectra and camera response matter too. 8 9 That is a technical inference about color reproduction, not a finding that all LEDs produce poor photographs.
How to improve colors in Auto mode
People under warm room lights
Start by deciding how much of the room’s warmth you want to keep. If both a neutral shirt and the surrounding wall look excessively yellow, try the same portrait under one main source.
If the room also receives window light, compare with the lamp off. Look at the saved result before changing anything else. Avoid making a person’s face your only neutral reference: skin is not gray, and a setting that makes it less yellow may make the rest of the scene too blue.
Food and products
For a menu picture, a listing or a record of an item’s color, choose a stable setup you can repeat. Put a neutral reference beside the item under the same illumination for a comparison frame. Use a genuinely neutral matte reference where possible; an ordinary sheet of paper or a colored reflection is not a calibrated gray target. In an editor with a WB picker, a suitable neutral gray area can provide a starting point; avoid a glossy reflection or an overexposed white patch. 2
Then look at the item itself. A neutral reference is helpful, but it does not guarantee that a red fabric or green ingredient is reproduced faithfully.
Window light with the room lamps on
Take one comparison with the competing lamp off. If the color split becomes less distracting, changing the lighting may be enough. If you want to keep the lamp’s atmosphere, decide which part of the scene matters most rather than trying to make every surface equally neutral.
Sunset and golden-hour scenes
Keep a version that preserves the warmth you liked. You can also make a more neutral edit for comparison, but a warm sky is not automatically evidence of incorrect white balance.
The useful question is whether the photograph serves your purpose. A catalog image and a sunset memory do not need the same treatment.
These steps work without manual exposure settings. Move to WB controls when you need a more repeatable balance or Auto keeps choosing a look you do not want.
When manual white balance or a WB lock helps
Manual WB lets you select a balance instead of leaving the choice to Auto. A WB preset assumes a particular type of illumination. An AWB lock holds the balance the camera has already calculated. 4
WB lock caution: locking an unwanted Auto result preserves that unwanted result.
Check the color first, then lock if the lighting is stable. Recheck it when you move from the window into a room lit by another source.
| Option | When it helps | Benefit | Limit |
|---|---|---|---|
| Auto WB | Everyday shooting, especially when illumination changes. | The camera estimates the needed balance as conditions change. | Strongly colored or ambiguous scenes can be difficult; results can vary. |
| WB preset, if offered | A stable scene where the preset is a useful starting point. | A repeatable assumed illumination target. | Daylight or Tungsten is not an exact match for every lamp or scene; presets vary by device. |
| Manual WB, if offered | Stable lighting, repeated product shots, or a deliberate warm/cool look. | You can choose and retain the balance instead of letting Auto decide. | Kelvin alone may not address tint. Availability depends on the phone, app, camera and mode. |
| AWB lock, if offered | A stable scene where you already like the Auto result, including a video clip. | It prevents the AWB setting from continuing to change. | It can preserve a bad setting. Reconsider the lock when lighting changes; AE/AF lock is not automatically WB lock. |
Availability varies by phone, app, camera and mode. Samsung’s documentation describes WB color-temperature adjustment in supported Pro/Pro video workflows; check that the control exists for the camera you are using. 10
Apple documents a video setting at Settings → Camera → Record Video → Lock White Balance. 5 Use that path when the option is present; it does not establish a still-photo WB lock or support in every video mode.
Some phones also offer native manual photo WB. Apple’s iOS 27 guide documents manual WB for iPhone 18 Pro models in the Main-camera context. 6 That is a scoped example, not a control available on every iPhone.
Do not assume an AE/AF lock also locks WB. Exposure, focus and white balance are separate controls; only rely on the app’s documented behavior.
Fixing white balance after capture: JPEG/HEIF vs RAW
A saved photo with a modest shared cast may still be worth editing. Keep the original and make a comparison version.
Start with small temperature changes, then adjust tint if a green or magenta cast remains. Check several areas rather than making one surface neutral at the expense of everything else.
JPEG and ordinary phone HEIF photos
A JPEG already contains the camera’s rendered look. You can change its colors, but that does not give you the same freedom as processing RAW data. Adobe’s RAW/JPEG guidance explains this difference. 11
For ordinary iPhone HEIC images, Apple likewise describes HEIC and JPEG output as processed, display-ready photographs. 13 Adobe documents HEIF/HEIC support in Lightroom and Camera Raw, with compatibility depending on the application and platform. 15
The practical implication is to treat ordinary phone HEIF photos as rendered images when planning a WB correction. This does not mean every HEIF file has the same bit depth, HDR behavior or editing limits as JPEG. The format name alone does not tell you how much correction will look good.
RAW and computational RAW
RAW generally gives you more flexibility in choosing white balance during editing. Smartphone RAW modes can differ, however. Apple ProRAW combines RAW information with iPhone processing; it is not simply an untouched sensor dump. 14
Use an editor that supports the actual RAW format. A RAW capture can help with a difficult adjustment, but it cannot promise accurate color under every source, resolve all mixed lighting with one slider, or recover information that was never recorded. A RAW preview may hide detail the file still retains, but if the relevant capture data is truly clipped or lost, changing WB cannot recreate it.
If you regularly need substantial color corrections for products, RAW may be useful when your phone and workflow support it. For a casual dinner photo, a small edit to the original rendered image may be sufficient.
For the broader format decision, see when phone RAW is worth using, including editing, file management and sharing costs.
When the problem is processing or your display
If neutral areas look reasonable but a particular color still feels exaggerated, stop pulling the entire image warmer or cooler. Check an applied filter, style or editing profile. Adobe notes that a remaining cast can involve color rendering beyond WB. 2
If your complaint concerns contrast, oversaturated colors or an aggressively sharpened look, WB may address only part of it. Those symptoms deserve separate checks rather than another large temperature adjustment. Our phone HDR guide covers tone mapping and HDR viewing; why iPhone photos look overprocessed covers texture, contrast and rendering controls.
Also compare the same saved file with your display settings. On supported Apple devices, Night Shift warms the screen, while True Tone adapts its color and intensity to ambient light. 12 Temporarily switching those settings off can help you separate viewing conditions from the image you are editing.
That comparison does not calibrate the display. Nor does a photo looking different on another screen prove which view is accurate. It simply prevents you from making a large edit to compensate for a viewing effect.
Verdict
For most people, the best first move is to improve the light and check Auto again. Use manual WB or a lock when you need consistency under stable illumination. Use editing to refine the result, while recognizing that one adjustment may not solve mixed lighting or every color-rendering problem.
Choose neutrality when it matters to the task. Keep warmth when it belongs to the scene you want to remember.
Sources and disclosure
This is a research-based guide, not a matched phone-camera test. Photographic illustrations are AI-generated and demonstrate concepts; they do not establish a device’s color accuracy. Device controls depend on model, camera, mode, app and software version. The mixed-lighting and LED advice is conditional practical guidance based on the cited mechanisms.
1. Nikon, White balance (ZR manual). Accessed October 7, 2026.
2. Adobe, Make color and tonal adjustments in Camera Raw. Accessed October 7, 2026.
3. Google Research, Night Sight: Seeing in the Dark on Pixel Phones (2018). Accessed October 7, 2026.
4. Android Developers, CaptureRequest. Accessed October 7, 2026.
5. Apple, Change video recording settings on iPhone. Accessed October 7, 2026.
6. Apple, Use advanced camera tools on iPhone (iOS 27). Accessed October 7, 2026.
7. Afifi, Brubaker, Brown, Auto White-Balance Correction for Mixed-Illuminant Scenes (2021). Accessed October 7, 2026.
8. ETC, An Array of Choices Part 3: Mint and White. Accessed October 7, 2026.
9. EBU Tech 3353, Development of a Standard Television Camera Model (2012). Accessed October 7, 2026.
10. Samsung, Camera modes in Samsung galaxy smartphones. Accessed October 7, 2026.
11. Adobe Learn, When to use raw vs JPEG (2026-06-11). Accessed October 7, 2026.
12. Apple — Adjust the brightness and color temperature on your iPhone or iPad. Accessed October 7, 2026.
13. Apple Developer — Capture and process ProRAW images (WWDC21). Accessed October 7, 2026.
14. Apple Support — About Apple ProRAW. Accessed October 7, 2026.
15. Adobe — HEIC and HEVC media files support in Lightroom and Camera Raw. Accessed October 7, 2026.
