Image Contrast

Increase or reduce image contrast, with one-click auto-contrast

Contrast is the spread between the darkest and lightest tones of a picture. A hazy landscape, a photo through a window or a scan of a faded print has too little and looks grey and flat; a harsh midday shot can have too much. The image contrast tool pushes tones apart or pulls them together with one slider, and its auto-contrast option stretches the histogram so the image uses the full range from black to white.

A live histogram shows the effect as you work — contrast literally widens or narrows it — and a split preview compares the result with the original. The full-resolution image is edited locally in your browser, transparency survives, and a batch can be saved as one ZIP.

How to use the Image Contrast

1. Drop your images on the upload area, click to choose them, or paste one with Ctrl+V.
2. Move the Contrast slider: positive values add punch, negative values give a softer, matte look.
3. Choose the Pivot. Mid-grey matches CSS contrast(); Image average keeps the overall brightness steady.
4. For washed-out or hazy photos, tick Auto-contrast, then choose how many outliers to ignore and whether the channels are stretched linked or separately.
5. Compare with the split line, then press Download adjusted image or download all as a ZIP.

Image Contrast features

How the contrast formula works

Contrast scales the distance of every channel value from a pivot p: out = (in − p) × k + p. With the slider value c from −100 to +100, the factor is k = (100 + c) ÷ (100 − c). That has the same shape as the widely quoted formula F = 259 × (C + 255) ÷ (255 × (259 − C)) and gives nearly identical numbers, but it is symmetric: +50 triples the spread and −50 divides it by three.

With the mid-grey pivot (127.5), a level of 160 at +25 becomes (160 − 127.5) × 1.667 + 127.5 ≈ 182, and a level of 90 becomes ≈ 65. Values pushed past 0 or 255 are clipped, which is why strong contrast loses detail in deep shadows and bright highlights.

What auto-contrast does

Auto-contrast is a histogram stretch, also called levels or normalisation. It finds a low point lo and a high point hi in the histogram and remaps every value with out = (in − lo) × 255 ÷ (hi − lo). A foggy photo whose values only run from 40 to 210 is stretched so that 40 becomes black and 210 becomes white.

One stray black or white pixel would stop the stretch from doing anything, so Ignore outliers skips a percentage of the darkest and brightest pixels when choosing lo and hi; 0.5% is a good default. Linked uses one pair of values for all channels and keeps the colour balance. Per channel stretches red, green and blue independently: if one channel is weak — blue in a photo taken under warm bulbs, say — stretching it on its own removes much of the cast.

Tips and common mistakes

• Read the histogram first. If it already touches both ends, auto-contrast will change little and the slider is the better tool.
• Strong contrast also makes colours look more saturated and noise more visible. For portraits, stay under about +30.
• Negative contrast with the image-average pivot gives a faded, matte look without changing overall brightness.
• Per-channel auto-contrast can neutralise colours that are meant to be there, such as a golden sunset. Switch to Linked if colours shift.
• Brighten or darken with the brightness tool, crisp up edges with Sharpen Image, and for black-and-white work convert with Grayscale Image before adding contrast.