Underexposed, moody or backlit: diagnose before you brighten
Three different problems look identical on a thumbnail, and the right amount of brightening differs for each.
- Under-exposed. Not enough light reached the sensor. The data is in the file but sits low in the scale. This is the case brightening helps most.
- Dark by design. The scene really was dim, such as a candle-lit room. Cambridge in Colour calls images with most tones in the shadows 'low key'. Brighten this too far and you remove the mood it had.
- High contrast. A dark subject in front of a bright window or sky. Lifting the whole picture overexposes the bright part before the subject looks right.
What brightening can and cannot recover
A tone curve moves existing brightness values to new ones, and it can't invent information. Cambridge in Colour is blunt about the bright end: detail can never be recovered once a region becomes solid white. Their view of the dark end is gentler. Under-exposure is usually more forgiving than over-exposure, though lifting dark tones costs you something in noise, because the grain recorded in the shadows is brightened along with the picture. A photo that's dark and also grainy at the start will be grainier when you finish, and no curve repairs that.
Which curve suits which kind of dark photo
| Cause | Start with | Why |
|---|---|---|
| Slightly dim indoor shot | Gentle | A small lift of the middle tones is enough and keeps contrast. |
| Properly under-exposed | Balanced | Lifts the middle tones and the shadows a little, so faces open up. |
| Very dark, mostly shadow | Strong, at partial strength | A steep curve brings out the most detail but also the most noise. |
| Bright background, dark subject | Gentle, then crop | A global curve cannot fix uneven light. Crop out the bright area if you can. |
Brightening a photo step by step
- Make a copy and work from the original file.
- Open the image brightener and load the picture. It shows a before and after preview.
- Start with Balanced at about two thirds strength.
- Raise it until the subject looks natural. Stop when skin or sky starts to look flat.
- Check the report: average brightness and the share of pure-white pixels before and after. A jump in white pixels means highlights are being lost.
- Download the full-size result and compare it with the original on a bigger screen.
Why highlights survive a curve but not a brightness slider
A simple brightness slider adds the same amount to every pixel, which is why a sky that was nearly white to begin with goes completely white while the face is still murky. The brightener here scales color together with brightness and never lets a channel pass its maximum, so bright areas get less lift than dark ones, and the curve keeps pure white at white. The idea is the same as gamma. Stored values don't match how bright things look to the eye, so lifting the middle of the scale makes dark areas readable without moving the ends much. Cambridge in Colour's gamma article explains why dark tones get a disproportionate share of the stored values in the first place.
Does brightening make the file bigger or worse?
Dimensions stay the same. The file can grow, because more visible detail and more visible noise compress less well. If you save as JPEG or WebP you're compressing again, so choose a quality of 90 or higher. The result is encoded fresh from pixels, so camera and location fields from the original aren't copied into it. For sending the finished picture somewhere with a size limit, compressing an image for email covers that part.
Compare the three looks on one photo
Run all three looks on the same photo and put the results side by side before you commit. The script reads the file in this tab and makes no network requests of its own, so the picture is not uploaded. If the result is noisy, a smaller copy hides the grain, and the image resizer can make one.
Questions people ask
How do I fix a photo that is too dark?
Lift the middle tones with a curve, raise the shadows only as far as needed, and check that the bright areas keep their detail. Work on a copy so the original stays intact.
Is a pure black area recoverable?
No. Pure black holds no detail to bring back, and brightening just turns it grey. Shadows that are dark but not pure black usually have something left to lift.
Do the bright areas clip when I brighten?
It can with a crude brightness slider. The tool here limits each color channel and reports the pure-white share before and after, so you can see whether highlights were lost.
Does brightening reduce image quality?
It makes noise easier to see, and saving as JPEG or WebP adds a round of compression. Both effects are small at high quality settings, and larger on a very dark source.
Does this use AI?
No. It uses fixed tone curves, applied the same way to every pixel, so the same photo and settings always give the same result, and nothing about the picture is analyzed or sent anywhere.
Does it work on a phone?
Yes in current mobile browsers, though large photos are slow on older phones. A smaller copy brightens faster and is fine for checking the settings first.
Sources