You come home from a night out with forty frames on the card. Thirty-eight are fine, one is hopeless, and one — the good one, the one where everyone is actually looking at the camera — is just slightly smeared. Not ruined. Not crisp. Just soft enough that you can see it at a glance on a phone screen and feel it immediately on a laptop. That narrow middle ground is where most handheld night photography lives, and it is also, fortunately, the most fixable kind of blur there is.
Slight motion blur sits in a useful zone: enough smear to soften eyes, hair, lettering and edges, but not so much that the subject dissolves into a streak. That distinction matters, because the repair strategy depends entirely on which kind of blur you are looking at, and on whether the detail you want was ever recorded in the first place. This guide covers how to diagnose a night shot, what a realistic rescue looks like, and how to work through a batch of handheld evening photos without wasting time on frames that cannot be saved.
Why Night Shots Blur — and How to Read the Smear Before You Touch It
Night photography is a negotiation with light, and the camera almost always loses the first round. Indoors, on a street after sunset, or under a single warm bulb, there simply is not enough light reaching the sensor to allow a fast shutter. The camera compensates by holding the shutter open longer — often somewhere between a tenth and a thirtieth of a second — and by raising ISO. At those speeds, the hand is never still enough. Heartbeat, breathing, the small downward punch of pressing the shutter button: all of it lands inside the exposure and shows up as a directional smear across every edge in the frame.
Two other problems ride along with it. First, high ISO forces noise reduction, which smears fine texture in skin, hair and fabric even when the shot is technically steady. Second, in low light the autofocus hunts, and it sometimes locks a few centimetres behind the eyes — a miss that looks almost identical to shake on a small screen but behaves very differently when you try to correct it. Before any correction, diagnose. Zoom to 100% and study a hard edge: a door frame, a window mullion, the edge of a sign, an eyelash line, the rim of a glass.
What you see tells you what you have. Streaks running in one direction across the whole frame mean camera shake — your hand moved during the exposure. Streaks on the subject only, with a crisp background, mean subject motion: someone turned their head, laughed, or shifted their weight while the shutter was open. Softness that spreads equally in every direction, with no directional bias, usually means missed focus rather than movement. And blocky squares around high-contrast edges are compression artifacts, not blur at all — a different problem that responds to different tools.
One more check is worth doing before you commit to a repair: is the photo actually dark and sharp, or is it bright and soft? A sharp but underexposed night shot is an exposure problem, and lifting shadows will do far more for it than any blur correction. Reading the frame correctly at this stage is the difference between a two-minute fix and twenty minutes of fighting an image that was never blur-limited in the first place.
What a Blur Fix Can Realistically Recover — and When to Walk Away
The core mechanic of a good correction is reversal: it looks at the directional smear in the frame and pushes back against it, reconstructing the edge that the movement spread out. It works best on mild hand tremor with a recognizable subject — the exact scenario of a handheld night portrait where the person is clearly themselves, the clothing is visible, the background is intact, but everything carries a faint one-way softness. Under those conditions, a careful pass can tighten eyes, restore the texture of hair and bring back the crisp line of a collar or a sign.
It cannot, however, invent detail the camera never recorded. Heavy streaks, faces that have become smooth blobs, motion-blurred backgrounds that were never in focus to begin with, pixelation and JPEG blocking, genuinely dark in-focus shots and damaged scans all fall outside the recoverable range. Pushing a correction too hard on those files does not produce sharpness; it produces the crunchy, over-processed look that reads as artificial at any size. Knowing where the line is saves both time and disappointment.
Matching the tool to the diagnosis matters just as much as the diagnosis itself. A frame with a directional smear across the entire image calls for a shake-oriented correction. A frame where the background is tack-sharp and only the person moved calls for a motion-oriented approach. A portrait that is simply a little soft around the eyes — slight missed focus rather than movement — is best handled with a gentle face-focused pass. When the smear is mild and clearly directional, running the frame through a dedicated unblur photo pass is the sensible first attempt, because it targets exactly that failure mode instead of applying a blunt global sharpen.
Practical housekeeping makes the process cheaper and cleaner. Work on a copy, never the original. Check the file format and size before you start — uploads of JPG, PNG or WebP up to 4 MB are the working range, so a 30 MB RAW export needs resizing first. Runs are credit-based with no free trials, which is a good argument for culling before you process: flip through the burst at 100% and pick the single sharpest frame rather than processing all six. Finally, judge every result at full size and on a face. If the person comes back looking like someone else — different bone structure, waxy skin, altered features — the correction has gone too far. Discard that result rather than trying to soften it back into shape.
A Real-World Workflow: From Restaurant Table to Streetlight Street
Consider three frames from a single evening out. The first is a portrait across a restaurant table, lit by one warm pendant lamp. The background — banquette, menu, glassware — is clean and sharp; the subject’s head has a faint horizontal smear because they laughed mid-exposure. That is subject motion, and it responds well to a motion-oriented correction, with the caveat that the amount of travel in a laughing head is usually larger than hand tremor, so expectations should stay modest. The second is a candid shot under a streetlight where the entire frame, background included, streaks in one consistent diagonal direction. That is camera shake, the classic handheld night failure, and it is the most cooperative candidate of the three.
The third is a night market snapshot where nothing is streaked but the eyes are a touch soft while the collar and the stall behind are reasonably defined. That is a light focus miss, and a soft-face pass is the right instrument — provided the correction is kept subtle. In all three cases the workflow is the same: pick the best frame from the burst, examine it at 100%, note the direction and distribution of the smear, choose the mode that matches, and then compare before and after on eyes, hair and any text in the frame. Text is an honest judge — if signage lettering has become sharper and still legible, the correction is working.
Once the blur is addressed, the frame usually needs one more thing: the light. Night handhelds tend to come back with crushed shadows on faces, dull skin and warm artificial light that has drifted orange or green. A texture-and-light pass — lifting shadows on faces, keeping warm light warm, restoring detail in hair, skin and fabric, and bringing colour back toward natural — finishes the image without altering the pose, the crop, the furniture or the background. That separation of jobs is worth respecting: correct the blur first, correct the light second, and never ask one step to do both.
The strongest defence, of course, is at capture. Brace your elbows against your ribs or a table, exhale, and squeeze rather than stab the shutter. Tap focus on a bright, high-contrast edge near the subject’s eye. Shoot a short burst — three to five frames — because the second or third is frequently the steadiest. Keep the shutter speed at or above roughly one divided by your focal length where the light allows, and use whatever ambient light exists rather than forcing a flash that flattens the scene. A frame that arrives slightly soft but well exposed and correctly focused will always be easier to rescue than one that is bright, blurry and beyond repair.
A Gothenburg marine-ecology graduate turned Edinburgh-based science communicator, Sofia thrives on translating dense research into bite-sized, emoji-friendly explainers. One week she’s live-tweeting COP climate talks; the next she’s reviewing VR fitness apps. She unwinds by composing synthwave tracks and rescuing houseplants on Facebook Marketplace.