Phone photography that looks like a camera: a practical guide

Seven habits that move phone photos from snapshot to deliberate picture, each tied to what the phone's burst pipeline is actually doing when you press the shutter.

A person holding up a smartphone to photograph desert rocks and Joshua trees in Joshua Tree National Park
Photo: Joshua Tree National Park, Public domain, via Wikimedia Commons

The phone photography tips that make the biggest difference are about working with the phone's burst processing, not against it: wipe the lens, tap to set focus and pull exposure down for highlights, use the phone's real lenses instead of pinch-zoom, shoot RAW when the light is hard, use the 2x–3x lens for portraits, keep night mode for still subjects, and edit with curves and white balance rather than filters. Each habit maps to a specific step in how modern phones merge several frames into one photo.

Key takeaways
  • Wipe the lens before every deliberate shot; a smear costs contrast no software restores.
  • Pull exposure down for highlights: clipped pixels stay white in every burst frame.
  • 2x crop = 12.5 MP on a 50 MP sensor; in-between zoom values are crops and upscales.
  • RAW for hard or mixed light; a 48 MP ProRAW file is about 75 MB.
  • 333 ms handheld vs 1 s braced: night mode rewards a still phone and a still subject.

A phone photo is not one exposure. Since Google's HDR+ paper in 2016, flagship phones have captured bursts of frames, aligned them and merged them before you see anything.

That pipeline is why phones handle contrast and dim light so well for their sensor size, and also why some photos come out smeared, plasticky or oddly flat. The steps below are what I would tell anyone who wants their phone pictures to look deliberate.

The phone in my own kit is a Samsung Galaxy S25+, so I use it for concrete numbers, but every step applies to any recent iPhone, Pixel or Android flagship.

0.3 luxlight level Google's Night Sight was built to handle (Liba et al., 2019)
333 mslongest single frame Night Sight takes handheld
~75 MBsize of one 48 MP Apple ProRAW file
12.5 MPreal pixels left in a 2x crop of a 50 MP sensor

Why does a clean lens matter more than any setting?

A fingerprint on a phone lens spreads bright light across the frame as haze and streaks, lowering contrast in a way no software step fully removes. Phone lenses are a few millimetres wide and sit exactly where your fingers and pocket lint go, so a single smear covers a large share of the aperture.

Wipe the lens with a soft cloth or the inside of a cotton shirt before every deliberate shot, not after you see the haze. The symptoms to watch for:

  • a glow around streetlights at night;
  • a milky, low-contrast look in backlit scenes;
  • streaks radiating from bright points such as the sun or headlights.

Phone cameras are multi-element stacks of tiny moulded lenses; the X-ray scan below shows how much glass sits behind that small window.

The story

Sean Baker's feature film Tangerine was shot between 24 December 2014 and 18 January 2015, mostly around Santa Monica Boulevard in Hollywood, on three iPhone 5s phones, with cinematographer Radium Cheung. The crew ran the FiLMiC Pro app for control over focus, colour temperature and bit rate, steadied the phones with Tiffen's Steadicam Smoothee, and fitted Moondog Labs anamorphic adapters, prototypes Baker got free while the product was still on Kickstarter.

The whole film cost about $100,000, and the money saved on cameras went to extras and locations. It premiered at Sundance on 23 January 2015. The lesson for a phone shooter is the crew's: real glass, deliberate settings and a steady phone, not more pixels.

Wikipedia: Tangerine (film), production
A 3D X-ray microscope image of a mobile phone camera lens module showing the stacked lens elements inside the housing
A phone camera lens module imaged with 3D X-ray microscopy. Several aspheric elements are stacked in a barrel only a few millimetres across; anything on the outer window degrades all of them.Photo: ZEISS Microscopy from Germany, CC BY 2.0, via Wikimedia Commons

How do you lock focus and exposure on a phone?

Tap the subject to focus, then drag the exposure slider down until the brightest important area keeps its detail; on iPhone, a long press locks both (AE/AF lock), and on Samsung and Pixel the slider appears after a tap. Phones meter for the whole scene by default, which is why skies and faces near windows blow out.

Underexposing a little is not a hack; it is the phone's own strategy. Google's HDR+ pipeline deliberately captures every frame of the burst at the same, low exposure so highlights do not clip, then merges the frames to clean up the shadows and tone-maps them brighter.

Pulling the slider down hands the pipeline frames it can work with. Raising it in a contrasty scene does the opposite: clipped pixels are white in every frame, and averaging eight white frames still gives white.

Burst photography for high dynamic range and low-light imaging on mobile camerasSamuel W. Hasinoff et al. (Google Research) · ACM Transactions on Graphics / SIGGRAPH Asia · 2016

The paper behind HDR+, shipped on the Nexus 6, 5X and 6P. Instead of bracketing, it captures up to 8 frames at one constant, deliberately short exposure, works on raw Bayer data rather than the phone's processed output, aligns the frames with a fast FFT-based method and merges them with a Wiener filter, then applies HDR tone mapping. Exposure rises to 8 ms before gain is raised, with caps of 100 ms and 96× gain. On 12–13 MP sensors that means processing up to 104 MP per photo, in about 4 seconds on 2016 hardware. Limit: the paper judges quality mainly by comparison images, not a large perceptual study.

Read the paper (free)
a burst of raw frames, same short exposure, highlights protected align tiles matched to a reference frame merge noise averages out, moving parts rejected tone map shadows lifted, white balance set If frames are clipped, or the subject moves, merging cannot recover it: that is why exposure and stillness still matter.
What happens after you press the shutter on a modern phone, simplified from Hasinoff et al. (2016) and Liba et al. (2019). Your two inputs that survive the pipeline are exposure (keep highlights unclipped) and stillness.

Should you use digital zoom or the telephoto lens?

Use the phone's marked lens positions (0.6x, 1x, 2x, 3x, 5x and so on) and avoid the in-between values, because between lenses the phone is cropping and upscaling. A 2x crop keeps only a quarter of the sensor's pixels; a 4x crop keeps one-sixteenth.

Pixel count falls with the square of the zoom factor. On the Galaxy S25+, whose main camera is a 50 MP sensor of about 1/1.56 inch behind an f/1.8 lens, a 2x crop leaves 12.5 MP of real pixels, which is still a full 12 MP output; that is why many phones now label 2x as a usable "lens".

At 3x the phone switches to its 10 MP telephoto, which has more real pixels than a 3x crop of the main sensor (about 5.6 MP). The do megapixels matter post covers when those pixels turn into visible detail.

1x: whole sensor, 50 MP 2x crop: 1/4 area 3x: 1/9 Real pixels left 1x main: 50 MP 2x crop: 12.5 MP 3x crop: 5.6 MP 3x telephoto lens: 10 MP pixels = sensor MP ÷ zoom² (before multi-frame tricks)
Cropping is not free. On a 50 MP main sensor, 2x still yields a full 12 MP picture, but 3x from the main camera has fewer real pixels than the dedicated 10 MP telephoto.

Multi-frame tricks narrow the gap a little. Google's Super Res Zoom uses the tiny shifts of your hand between burst frames to fill in detail that one frame cannot hold, which is why a slightly shaky hand actually helps at moderate zoom.

Handheld Multi-Frame Super-ResolutionBartlomiej Wronski et al. (Google) · ACM Transactions on Graphics / SIGGRAPH · 2019

The method replaces demosaicing with a multi-frame merge: natural hand tremor shifts each raw frame by sub-pixel amounts, so a burst samples red, green and blue at more positions than one frame can, giving more resolution and less noise. It handles local motion and occlusion, runs at about 100 ms per 12 MP raw frame on a phone, and became the basis of Super Res Zoom and the default merge in Night Sight on Google's flagship. Limit: the gain is real at modest zoom but cannot replace optics at 10x and beyond, where a periscope telephoto wins.

Read the paper (free)

Use this to check what you just read. Put in your phone's sensor numbers and see how many real pixels you have at any zoom.

Real pixels at any zoom

Lens positionGalaxy S25+ hardwareReal pixels at that zoomUse it for
0.6x12 MP ultra-wide, f/2.212 MPInteriors, tight streets; keep faces away from edges
1x50 MP main, f/1.8, OIS, ~1/1.56 inch50 MP sensor (default photos are smaller)Everything in low light; landscapes
2xCrop of the main sensor12.5 MPHalf-body portraits, details
3x10 MP telephoto, f/2.4, OIS10 MPHead-and-shoulder portraits, compression
5x and upCrop of the telephoto3.6 MP at 5xRecord shots only

When should you shoot RAW on a phone?

Shoot RAW when the light is hard or mixed (midday sun, backlit faces, sodium streetlights, stage lighting) and you plan to edit; stay in the normal HEIF or JPEG mode for everyday shots, moving subjects and anything you will share straight away. RAW keeps more tonal and colour data, so you can recover highlights and fix white balance without the banding that a processed JPEG shows.

Apple's ProRAW, on iPhone 12 Pro and later Pro models, is a computational RAW: it keeps Smart HDR, Deep Fusion and Night mode merging but stores the result as a DNG. Apple puts a 12 MP ProRAW file at about 25 MB and a 48 MP one at about 75 MB, so a day of RAW shooting fills storage quickly.

ProRAW does not work with Live Photos, Portrait mode or video. Most Android flagships offer a similar RAW or DNG option in Pro mode or a separate RAW app.

FormatWhat you getTypical sizeBest for
HEIF or JPEGFully processed, 8-bit, tone-mapped2–5 MBSharing, moving subjects, everyday
Single-frame RAW (Pro mode)Sensor data, little or no multi-frame mergingVaries by phoneTripod and bright scenes
Apple ProRAW (12 MP)Merged multi-frame data as DNG~25 MBHard light you will edit
Apple ProRAW Max (48 MP)As above at full resolution~75 MBLandscapes for large prints or crops

The reasoning behind exposing for highlights and pulling shadows is laid out in RAW, dynamic range and ETTR, and fixing colour casts is covered in white balance and colour temperature.

Which phone lens is best for portraits?

Use the 2x or 3x position, roughly 50–70 mm full-frame equivalent on most phones, and step back. Distance, not focal length, sets perspective: shooting a face from 50 cm with the 1x lens enlarges the nose relative to the ears, while the same head-and-shoulders framing at 3x puts you about three times farther away, which flattens the face naturally.

Portrait mode blur

Portrait mode's blur is synthetic: the phone estimates depth and blurs by layer, which is why hair edges and glasses sometimes look cut out. Turn the blur strength down to a lower simulated f-number setting and it becomes much more convincing. The full mechanism is in how computational photography works, and the optics of distance and compression are in focal length and perspective compression.

Close-up of the three rear camera lenses on an iPhone 13 Pro: ultra-wide, wide and telephoto
Three fixed lenses on an iPhone 13 Pro: ultra-wide, wide and telephoto. The marked zoom positions in the camera app are where you get each lens's real optics.Photo: 人工知能, CC BY-SA 4.0, via Wikimedia Commons

Why does night mode smear moving subjects?

Night mode collects light over 1 to 6 seconds of frames, so anything that moves during that time is blurred, ghosted or partly rejected by the merge.

Use night mode forSwitch it off for
Streets and buildingsChildren and pets
Still people who can hold a poseDancers and performers
Landscapes, the sky at duskTraffic and riders

Google's Night Sight paper explains the trade. The phone measures handshake and scene motion before the shot ("motion metering") and picks the number of frames and exposure per frame.

Handheld, it caps each frame at 333 ms; if it detects the phone is braced or on a tripod, it allows 1-second frames. More stillness means longer frames and less noise, which is why propping the phone on a wall or railing improves a night photo more than any setting.

Propping the phone on a wall or railing improves a night photo more than any setting.This post, from Liba et al. (2019)
Longest single-frame exposure in Google burst pipelines
See the data as a table
Per-frame exposure (ms)
HDR+ (2016) maximum100
Night Sight, handheld333
Night Sight, braced or on tripod1000
Night Sight paper demo, long exposure4000

Source: Hasinoff et al. 2016 (HDR+); Liba et al. 2019 (Night Sight) · The 4,000 ms value is the paper's demonstration of 25 frames of 4 s each on a stabilised camera, not a default setting.

Handheld Mobile Photography in Very Low LightOrly Liba et al. (Google Research) · ACM Transactions on Graphics 38(6) · 2019

The paper behind Night Sight describes a pipeline that produces usable photos at light levels as low as 0.3 lux, where human vision loses colour. It uses motion metering to choose frame count and exposure, a robust merge, a learned white balance for low light and local tone mapping. Total capture takes 1 to 6 seconds, with processing under 2 seconds. In one 0.06 lux test, handheld capture used 13 frames of 333 ms, while a tripod allowed 6 frames of 1 second with visibly less noise. Limit: it is optimised for static scenes; fast motion still blurs.

Read the paper (free)
A night city scene photographed on a smartphone, with lit buildings and streetlights
Night scenes with static buildings are where multi-frame night modes shine; people walking through the same frame are what they handle worst.Photo: Enoch Leung from Canada, CC BY-SA 2.0, via Wikimedia Commons

How should you edit phone photos: filters or curves?

Edit with exposure, white balance and a gentle tone curve first, then local adjustments, and skip preset filters. A filter applies the same heavy colour and contrast shift to every photo whatever its light, and on an already tone-mapped phone image that often crushes shadows and shifts skin tones.

  1. Straighten and crop. Level horizons; crop to the composition, not to a fixed ratio. Square suits a centred subject, a point the forum thread on square photos makes for screens viewed in either orientation.
  2. Set white balance. Neutralise a grey or white surface, then warm or cool to taste.
  3. Exposure and highlights. Pull highlights down before lifting shadows.
  4. Tone curve. A soft S-curve adds contrast without the harsh clipping of a contrast slider.
  5. Reduce sharpening and "clarity". Phones already sharpen; adding more makes halos.
  6. Check skin. Look at faces at 100% for orange or grey casts and correct with the colour mixer.

Composition matters more than any of this. The case for and against the grid overlay in your camera app is in does the rule of thirds work.

Silhouette of a fisherman standing in a boat on the Brahmaputra River at sunset, with an orange sky
A silhouette on the Brahmaputra at sunset. To get this on a phone, tap the bright sky and pull exposure down; the phone's instinct is to lift the boat and wash out the sky.Photo: Inu Etc, CC BY-SA 4.0, via Wikimedia Commons

Do phone photos now carry AI labels?

Some do. Samsung's Galaxy S25 series, launched in January 2025, attaches C2PA Content Credentials to images edited with its generative AI tools, and Google's Pixel 10 phones, launched in August 2025, attach Content Credentials to every JPEG taken in the Pixel Camera app. The credential records how a file was made and edited, and it can be checked with free viewers.

Credentials are stripped by many platforms and screenshots, so they help but do not prove anything on their own. For using phones on the road, see the motorcycle travel photography guide, and for photographing people respectfully, street photography in India. My full camera kit is on the gear page.

Questions people ask

Do megapixels matter on a phone?

They matter for cropping and for how far the phone can zoom without a telephoto, and less for everyday detail, because 50 and 200 MP sensors usually bin pixels into 12–50 MP photos. The top answer on the Stack Exchange thread makes the key point: at the same output size, noise does not depend on pixel count, so more megapixels mostly buys cropping room.

Are smartphone light meter apps accurate?

Good ones are close enough for film and manual cameras. On the Stack Exchange thread, one user found a phone meter matched a Canon 5D Mark II's metering, and a meter-app developer said reflected readings can be within about a tenth of a stop when used correctly, with limits in dim light and no incident readings without an add-on dome.

Why do people like square photos?

Square looks the same on a phone held upright or sideways, and it has a long history in 6×6 cm medium format. Instagram made it the default for years. It suits centred, symmetric subjects; use it as a composition choice, not a habit.

Sources

  1. Hasinoff et al.: Burst photography for high dynamic range and low-light imaging on mobile cameras, ACM TOG (2016; constant short exposure, 8 frames, 100 ms cap)
  2. Liba et al.: Handheld Mobile Photography in Very Low Light, ACM TOG (2019; 0.3 lux, 333 ms and 1 s frames, 1–6 s capture)
  3. Wronski et al.: Handheld Multi-Frame Super-Resolution, ACM TOG / SIGGRAPH (2019; Super Res Zoom)
  4. Wikipedia: Tangerine (film) (three iPhone 5s, FiLMiC Pro, Moondog Labs adapters, $100,000 budget; source for the story)
  5. Apple Support: About Apple ProRAW (file sizes, supported models, limits)
  6. SamMobile: Galaxy S25+ launched (2025; 50 MP 1/1.56-inch f/1.8 main, 12 MP f/2.2 ultra-wide, 10 MP f/2.4 3x telephoto)
  7. TechCrunch: Galaxy S25 will support Content Credentials (2025)
  8. Google: How Pixel and Android bring C2PA Content Credentials to images (2025)
  9. RedShark News: Smartphone camera news from MWC Barcelona 2026 (2026)
  10. Photography Stack Exchange: Do megapixels matter with modern sensor technology?
  11. Photography Stack Exchange: How well do smartphone light meter apps work?
  12. Photography Stack Exchange: Why do people like square photos?
phone photographycomputational photographyHDR+Night SightProRAWdigital zoomSuper Res ZoomContent Credentials
Sapta

Sapta (Saptarshi Sarkar) is a mechanical engineer turned marketer in Bengaluru and a solo touring rider. He writes the seven Saptaverse blogs. Work with him or read more of Artblog.

Keep reading in Artblog

From the other blogs