Single RAW to HDR: How to Fake It When You Missed the Brackets (2026)
We’ve all been there: the light was perfect, you fired one frame, and only later realized you never shot brackets. Or you’re reworking an old photo from before you knew what bracketing was. A single RAW file holds more dynamic range than the JPEG preview suggests — typically 1–2 extra stops in the shadows — and with the right workflow you can squeeze a convincing pseudo-HDR out of it.
The honest framing first: this is an emergency technique. Nothing here creates information that wasn’t captured. But for rescuing shadows, balancing a contrasty scene, or giving an old photo new life, it works surprisingly well.
When single-RAW HDR makes sense
- Missed brackets: the moment is gone; the RAW is all you have.
- Moving subjects: real brackets would ghost badly (crowds, water, foliage in wind) — one frame sidesteps ghosting entirely.
- Old photos: reprocessing shots from before you started bracketing.
- Camera limits: some cameras and phones simply can’t auto-bracket.
Method 1: Export virtual exposures, then merge
The most effective approach: develop the same RAW three times at different exposures, then merge the results as if they were brackets.
- Open the RAW in your editor of choice — RawTherapee, darktable, and Lightroom all work, and the first two are free.
- Export three 16-bit TIFFs: one at −2 EV, one at 0 EV, one at +2 EV. Keep white balance, noise reduction, and lens corrections identical across all three — only the exposure changes.
- In Luminance HDR, open the HDR Wizard and load the three TIFFs (same steps as real brackets).
- Skip alignment — the frames are pixel-identical — and leave anti-ghosting off. There is nothing to deghost.
- Merge, then tone map gently: Drago or Mantiuk ’08 with conservative settings. All three files share one noise floor, so aggressive operators amplify noise fast — see our natural HDR settings guide for safe starting points.
Method 2: Tone map the single image directly
Luminance HDR can apply tone mapping operators straight to a single LDR or RAW image — no merge step at all. This is the fastest option when you only want a subtle shadow lift:
- Open the single image in Luminance HDR.
- Choose Tone Mapping and pick Drago or Reinhard ’05.
- Keep the strength low. You’re enhancing, not reconstructing — push it and you’ll get the classic overcooked look plus amplified noise.
Think of this as a smarter shadows/highlights slider rather than true HDR.
Real brackets vs. single RAW: honest limits
| Real brackets (3–7 frames) | Single RAW (virtual exposures) | |
|---|---|---|
| Highlight recovery | Excellent — the darkest frame holds the brights | None — clipped is clipped |
| Shadow noise | Low — brighter frames carry clean shadows | Higher — one noise floor, amplified |
| Dynamic range gained | 4–8+ stops | 1–2 stops, mostly in the shadows |
| Ghosting risk | Yes — needs deghosting | Zero — it’s a single frame |
| When to use it | Any planned HDR work | Missed brackets, moving subjects, old photos |
The table says it plainly: if you can shoot brackets, always shoot brackets — our bracketing guide shows how. Single-RAW HDR is the backup plan, not the plan.
Frequently asked questions
Can you make HDR from a single photo?
You can make a pseudo-HDR: export virtual exposures from the RAW and merge them, or tone map the single image directly. It convincingly extends shadow detail, but it cannot recover clipped highlights — for that you need real brackets.
Is single-image HDR real HDR?
Honestly, no. True HDR combines multiple captures to record a wider brightness range than the sensor captures in one shot. Single-image techniques only redistribute the data from one capture. They look HDR-ish, which is often enough — but the physics differ.
How many virtual exposures should I export?
Three (−2, 0, +2 EV) is the sweet spot. Five adds little, because the extra frames contain no new information — they’re interpolated from the same RAW data. Save the disk space.
Will virtual exposures add noise?
They amplify existing noise rather than adding new noise. The +2 EV virtual frame is just the shadows pushed two stops, noise included. Keep tone mapping gentle, and consider mild noise reduction on the merged result.
What about AI single-image HDR tools?
AI tools hallucinate plausible highlight detail where none was captured. That can look impressive for casual use, but the “recovered” detail is invented — fine for social media, risky for client work where accuracy matters. For a free, honest approach, the virtual-exposure workflow above keeps you in control of what’s real.
Luminance HDR is free and open source (GPLv2). Download it here.