How to Fix Exposure: Shadows, Highlights, and Contrast
A step-by-step exposure workflow covering histogram checks, highlight and shadow recovery, and contrast control.
Published · · Photique editorial
What's Actually Broken: Exposure, Contrast, or Both
Before you touch a single slider, answer one question: what is actually wrong with this image? That sounds obvious, but most editing mistakes start right here. You see a dark photo and reach for the Exposure slider — when the real problem is flat contrast.
Exposure at capture is the amount of light recorded by the sensor. In an editor, the Exposure control changes the overall brightness of the recorded image. If the whole frame looks uniformly too dark or too bright, it is usually a sensible place to start, after ruling out a local lighting or contrast problem.
Contrast is the relationship between the bright and dark areas. A dull, hazy-looking image usually has low contrast: shadows that aren't deep enough, highlights that aren't bright enough, everything compressed into a gray middle. That is a different problem and needs a different fix.
Misread the diagnosis and lifting brightness on a flat image will not restore tonal separation. A large lift in underexposed shadows can also reveal noise already present in the file.
Reading the Histogram: What the Current Rendering Shows
The histogram shows how tones are distributed across the brightness range, dark values on the left and bright values on the right. There is no single correct shape. A central hump may suit an overcast portrait, while peaks near the edges may suit a high-contrast silhouette. The histogram does not judge the image, and it does not by itself prove that source data is missing: a RAW converter usually plots the current rendering, not every value captured by the sensor.
Clipping happens when values reach the limit of the available range. In a finished RGB image, clipping in all channels produces pure black or white without tonal detail. A single color channel can clip while luminance detail remains, causing a color shift instead. A pile-up at an edge is a reason to inspect the relevant area, not an automatic verdict on the whole frame.
Most RAW converters provide clipping warnings. Enable them, inspect important areas under the current processing settings, and check individual color channels. The command names and warning colors vary between applications.
ETTR — exposing as far to the right as important highlights safely allow — can reduce shadow noise. However, an in-camera histogram is normally derived from the JPEG preview and may not match the RAW file's exact headroom. Check channel warnings, account for scene contrast, and do not sacrifice important highlights merely to move the graph rightward.
Recoverable Shadows vs. Blown Highlights
This distinction matters more than any slider technique.
A RAW file often contains more usable shadow information than its initial preview shows. How far those shadows can be lifted depends on the sensor, ISO, exposure, and output size; noise and color errors emerge along with detail.
With highlights, distinguish near-white tones, clipping in one channel, and clipping in all channels. A RAW converter may reconstruct color or structure from channels that remain below their limit. If an important area reached the ceiling in every channel and contains no tonal information, however, it cannot be recovered faithfully; generative fill can only invent plausible new content.
A practical test: lower overall exposure and highlight controls while watching clipping warnings and individual channels. If natural structure returns, the file had headroom. If an area remains uniform across different processing settings, it is probably fully clipped. One slider at its extreme is not a universal test because controls and ranges vary by editor.
One important nuance: a sky that looks almost white is not the same as one that has clipped. Near-white frequently holds recoverable data. Run the test rather than making that call by eye.
RAW vs. JPEG: How Much Room You Actually Have
RAW is unprocessed sensor data. JPEG is a finished file — the camera has already applied its own processing, compressed the image, and discarded information. The correction latitude between the two formats is substantially different.
RAW usually provides more latitude for tonal and color correction, but there is no universally safe number of stops. The limit depends on the camera, ISO, original exposure, scene, and intended output. A strong shadow lift in JPEG is more likely to reveal noise, banding, or posterization because processing and compression have already discarded data.
That does not mean JPEG corrections are hopeless. The guiding rule is to work gently and locally. A heavy Exposure move across the whole frame degrades a JPEG quickly. A subtle tone curve, a subject mask on a face, a careful local correction on one dark area — these work far better. JPEGs tolerate precise, targeted adjustments and resist blunt global ones.
If a JPEG has hard clipping in areas that matter — a completely blown sky, a face that has gone fully to shadow — the honest answer is often that a reshoot will cost less effort than a rescue attempt.
A Step-by-Step Correction Order
Order matters. Moving sliders at random means each new adjustment is compensating for the last one, and you end up going in circles.
- Start with the histogram. Identify whether the problem is exposure, contrast, or clipping — before anything else.
- Exposure — use it only for a global brightness shift across the whole frame. Do not try to use it to fix localized areas.
- Highlights and Whites — shape the bright tones while watching important detail and clipping. Their interaction varies by editor, so there is no fixed order or value.
- Shadows and Blacks — shape dark areas and the black point. Watch noise, color, and depth as well as visible detail.
- Contrast — choose the control that suits the adjustment. A general Contrast slider is convenient for small changes; a tone curve gives finer control over separate ranges. An S-curve is one option, not a required destination.
- Local masks last — sky separately, face separately. Work globally first; masks are for what remains after the global pass.
| Symptom | First tool | Practical limit |
|---|---|---|
| Whole image too dark | Exposure +, then Shadows | Shadow clipping in JPEG |
| Blown sky | Highlights −, Whites − | True clipping — no data to recover |
| Flat, hazy look | Tone curve or Contrast | Color shift at heavy contrast settings |
| Dark face, bright background | Subject mask + Exposure | Noise at high ISO |
When Shadows and Highlights Are the Point, Not the Problem
Not every blown or crushed area needs fixing. Some images depend on breaking the standard exposure rules.
High-key portraits use a white or near-white background, soft shadows, and an overall sense of lightness — that is the genre. Pulling the background down to a neutral gray destroys the feel of the image.
Low-key and silhouette work relies on deep shadows and minimal detail, the subject defined by outline alone. Lifting the shadows to reveal texture erases exactly what makes it work.
Night scenes and candlelit interiors carry dark corners as part of their atmosphere. Correcting to a technically balanced exposure turns night into afternoon.
Before any correction, ask one question: does the image work as a whole, or has one area simply slipped out of control by accident? If it works, the correct edit may be no edit at all.
Edit, Reshoot, or Leave It
Edit when the file is RAW, clipping is absent or limited to non-critical areas, and the overall exposure shift is modest.
Reshoot when you are working with a JPEG that has hard clipping in important areas; when lifting the shadows would require pushing past the point where noise takes over the image; or when the scene had extreme dynamic range and no bracketing was used.
Leave it when deep shadows or blown highlights are deliberate choices — a silhouette, a high-key portrait, a low-key mood shot. Correcting those images toward technical neutrality removes the reason they were taken.
Checklist Before You Move a Slider
- [ ] Identified the actual problem: exposure, contrast, or both?
- [ ] Checked the histogram for true clipping
- [ ] Confirmed the file format: RAW or JPEG?
- [ ] Considered whether the "flaw" might be an intentional choice
- [ ] Starting with global adjustments; local masks come last
- [ ] Comparing before and after on a calibrated display when possible, with the histogram as a technical aid
If the image still is not coming together, you can upload one photo to Photique. It critiques composition, light, color, and technique; identifies strengths and weaknesses; provides an edit recipe; and can create an AI Edit or visualize a Reshoot. It cannot faithfully recover data the camera never recorded.
Frequently asked questions
What is the difference between fixing exposure and fixing contrast?
Exposure is overall brightness — if the whole frame is uniformly too dark or too bright, Exposure is the right slider. Contrast is the relationship between bright and dark areas. A flat, hazy image usually has low contrast, not a pure exposure problem, and needs a curve or Highlights/Shadows adjustment rather than a global Exposure move.
Can blown highlights be recovered in editing?
Only if usable data was recorded. Lower exposure and highlight controls while checking clipping warnings and individual channels. If natural texture returns, recovery is possible. If the area remains uniform across processing settings, it is probably clipped in all channels; one slider at an extreme does not prove that on its own.
How much exposure correction is safe on a JPEG?
JPEGs have significantly less latitude than RAW files. Heavy global Exposure moves quickly introduce visible noise and posterization. Work gently and locally on JPEGs — a subtle tone curve or a masked local adjustment will hold up far better than a blunt global shift.
When should I reshoot instead of trying to fix an image in editing?
Reshoot when working with a JPEG that has hard clipping in important areas such as a face or key subject, when recovering shadows would push noise to an unacceptable level, or when the scene had extreme dynamic range and no bracketing was used. Rescue attempts in those situations often cost more time than a new shoot.
Photique