Film vs Digital: Why Film Still Matters

The film vs digital photography argument has been running since roughly 2004, and for most of that time it was really an argument about nostalgia. It is not anymore. We run a developing lab in East Williamsburg, and the volume of film crossing our counter has gone up every year we have been open. Kodak says demand for film has doubled in five years. Harman bought new machines. Four different companies have released brand new 35mm cameras since 2024. So this post is not a defense of film. It is an honest comparison from a shop that handles both formats every day, including the sections where digital wins outright, the real arithmetic on what a single frame of film costs, and the specific things film simulation software still cannot do.

Why film is making a comeback

The film photography comeback has receipts now, not just vibes. In January 2025 Kodak's chief technical officer Terry Taber said film demand had doubled over the previous five years and that roughly half of Kodak's film workforce had been hired in the previous three. Kodak took its Rochester sensitizing line offline through November and December to upgrade the machine, stockpiling inventory first so shelves would not go empty. You do not shut a coating line for two months for a market you expect to die.

Kodak also spent the last year pulling distribution back in-house from Kodak Alaris, the entity spun off in the 2012 bankruptcy. Kodacolor 100 and 200 arrived in October 2025, Gold 200 and UltraMax 400 were rebadged that November, Ektachrome went direct in February 2026, and Ektacolor Pro and Ektapan landed in March 2026. Kodak's stated reason was to increase supply and create stability in a market where prices have fluctuated. Honestly, prices have not dropped yet.

Across the Atlantic, Harman Technology, which makes Ilford film, made its largest manufacturing investment since the 1990s: two custom-built converting machines, the first commissioned in three decades, more than doubling annual 35mm cassette capacity. Harman also launched a genuinely new color negative emulsion, Phoenix, followed by the reworked Phoenix II.

Then the cameras. Pentax shipped the half-frame 17 in June 2024, MiNT shipped the lidar-autofocus Rollei 35AF three months later, Lomography announced the LOMO MC-A in October 2025, and the Amsterdam startup Analogue has the aF-1 slated for late 2026. That is more new 35mm development than the previous fifteen years combined. Not all momentum, though: Pentax said in March 2025 it was still too early to say whether a second film camera was coming, which tells you the tooling costs remain brutal.

Film vs digital quality: where each one actually wins

Start with resolution, because people get it wrong in both directions. A carefully exposed, carefully scanned 35mm frame yields roughly 12 to 22 megapixels of useful subject detail depending on stock, lens and scanner. A modern 24-megapixel full-frame sensor generally out-resolves that, and the reason is contrast rather than line pairs. Film keeps resolving fine detail past the point a sensor gives up, but at progressively lower contrast, so it reads as soft. Digital wins the sharpness test even where film is technically recording more information.

Noise is not close either. A 2012 DxO Labs study compared Kodak Portra 160NC against a Nikon D3x and measured signal-to-noise at an 18 percent grey patch: 23.7 dB for the film, 43.9 dB for the sensor, roughly a seven-stop gap. Smooth areas look grainier on film even with noise reduction switched off entirely on the digital file. The D3x was a 2008 camera, and every sensor since has widened the margin. Anyone telling you film is cleaner in the shadows is selling something.

Dynamic range depends entirely on what you call usable. In that same study, at a marginal quality threshold the Portra measured 23.9 stops against the sensor's 13.7. At a good quality threshold it flipped: film 8 stops, sensor 10. Film's range is much longer but degrades the whole way down. A sensor's range is shorter but stays clean until it abruptly does not. Both camps quote half of that.

The genuine film advantage is highlight behavior. Color negative has soft limits at both ends of its curve. A sensor has a soft shadow limit and a hard highlight limit, because a photosite fills up and then it is full. Overexpose color negative by two or three stops and the highlights compress into a shoulder you can pull back in the scan. Overexpose a raw file by two and there is nothing left to recover. It is why we tell people to meter for the shadows on Portra 400, and why blown skies are rarer on the rolls we scan than on the digital files people bring in to compare. On color, neither format is more accurate. One decides in Rochester, the other in Lightroom.

What a frame of film actually costs

Here is the arithmetic nobody does before they buy their first camera. All prices below are ours as of August 2026, for a 36-exposure roll of 35mm, developed in our own lab. Standard developing is $18 a roll, and a 4K scan upgrade is $12.

So a frame runs between about 78 cents and $1.32 depending on stock and whether you want scans. Lab membership starts at $18 a month and drops developing to $16 and 4K scans to $6, saving $8 on every scanned roll, so it pays for itself somewhere past two rolls a month before you count the rolling developing credits it includes. Below that, pay per roll.

Now the digital side, honestly. The marginal cost of a digital frame is effectively zero. The cost is all up front and it depreciates. Say $1,200 on a body, a card and a spare battery. At $0.78 a frame, that same $1,200 buys about 1,540 frames of film, or roughly 43 rolls. Shoot fewer than 43 rolls over the life of that camera and film is genuinely cheaper. Shoot 10,000 frames on the digital body over three years and your cost per frame is 12 cents, at which point film is not remotely competitive.

Two things skew it. Film cameras are not free either, and our point and shoot inventory starts around $230 and runs past $2,000 for a Contax T2. But film bodies mostly hold or gain value while digital bodies lose most of theirs in five years. Neither changes the conclusion: film is cheap to start and expensive to shoot heavily, digital is expensive to start and free to shoot heavily. Pick on volume, not feelings.

Why learning photography on film teaches exposure faster

We say this on the shop floor several times a week. If you want to understand exposure rather than understand which way to drag the slider, learn it on a manual film camera.

Film removes the feedback loop you have been using as a crutch. No histogram, no back screen, no highlight blinkies. You have a meter reading and a decision, and you will not learn whether the decision was right for four or five days. The discomfort is the point. It forces you to build an internal model of how light behaves, because you have nothing else.

One ISO per roll helps too. Load Ilford HP5 400 and your sensitivity is locked for 36 frames, leaving aperture and shutter speed as the only variables. You will learn how they trade against each other faster than you would with auto ISO quietly covering for you. Ilford rates HP5 Plus at EI 400 and supports metering it from 400 to 3200 with adjusted development, which is the second lesson once the first lands.

The honest caveat: digital's instant feedback is a faster loop for anything you can verify visually. Flash placement, focus accuracy, motion blur at a given shutter speed. Learn those on digital.

For what to buy first, our guide to the best beginner 35mm cameras covers the specific bodies we hand people when they ask this question in the shop, and everything currently on the shelf lives in 35mm cameras.

The look, and what film simulations still can't do

Film simulation has gotten genuinely good. Fujifilm's in-camera recipes, VSCO's profiles and Dehancer's plugin are not the crude LUTs they were ten years ago. Dehancer builds its profiles from analog darkroom prints using colorimetry and densitometry rather than sampling scanner output, and its grain engine responds to exposure and tonal value instead of sitting flat on the image. If you need a film look on a digital file for a client deadline, use it. It works.

Three things still separate a simulation from an emulsion, and all three are structural rather than cosmetic. Halation is the first. When a bright highlight hits film, light passes through the emulsion, reflects off the film back or pressure plate, and returns up through the layers. The red-sensitive layer sits deepest, so that returning light is filtered by the layers above it before re-exposing, which is why halation reads red-orange rather than white. Simulations approximate this with a radial glow, which is not what is physically happening, and it shows on scenes with mixed highlight shapes. Even reviewers who like Dehancer note its halation tool needs per-image tuning, which is a polite way of saying no preset gets it right.

Grain structure is the second. Film grain is silver halide crystals with random sizes and positions, and it is the smallest detail in the frame by definition. Digital noise is a uniform grid, because pixels are a uniform grid. On a 45-megapixel file the grain you add is larger than the real detail beneath it, so you have to throw away resolution to make it convincing. Grain is also tonally dependent and present everywhere, where overlay blend modes leave extreme highlights and shadows suspiciously clean.

Spectral response is the third. A film's color comes from three dye layers fixed at manufacture, so when a scene contains light those layers respond to unusually, like tungsten on CineStill 800T, the result came from chemistry rather than a curve applied later. A simulation only transforms RGB values a sensor already recorded. It cannot recover a response the sensor never had.

Same scene, film and digital

When you compare a matched pair, ignore the overall color impression. Grading can move a digital file anywhere. Look at four things instead, because those come from the capture medium itself.

Start at the brightest area. On the digital frame, find where detail stops and note how abruptly it stops. On the film frame, follow the same highlight and watch whether it compresses into tone or clips flat. Second, look at a large area of even tone, a wall or a sky, which is where film's grain is most visible and the sensor's noise advantage most obvious. Third, find a hard edge between a bright source and a dark background and look at the color of the fringing. Fourth, zoom to 100 percent on brick or fabric or hair and compare how much contrast survives at the smallest scale.

Scanning both every day, we can tell you what you will find: the digital frame is cleaner and sharper, the film frame handles the bright end more gracefully. That is exactly what the measurements predict. Neither one is winning. They are failing differently, and which failure mode you prefer is the actual decision in front of you.

When digital wins outright

Shooting sports or fast action, shoot digital. Continuous autofocus tracking on a current mirrorless body is not something a 1980s SLR competes with, and burst rate is the difference between getting the peak moment and not.

Shooting a paid event, shoot digital. A wedding is 2,000 frames minimum with one attempt at every moment in it. That is 55 rolls, north of $2,000 in materials before you have made a dollar, plus a five-day wait to find out whether the ceremony exposures worked.

Shooting in very low light, shoot digital. That seven-stop signal-to-noise advantage is most decisive exactly where film is weakest. Pushing Portra 800 two stops is a real option and we do it in our lab regularly, but a modern sensor at ISO 12800 will still be cleaner.

Client needs files tomorrow, shoot digital. Need video, shoot digital. None of this is a concession. It is the correct tool for those jobs, and a shop that told you otherwise would be wasting your money.

Frequently Asked Questions

Is film photography making a comeback?

Yes, and it is measurable rather than anecdotal. Kodak reported in early 2025 that film demand had doubled in five years and that about half its film staff were hired in the prior three. Harman more than doubled its 35mm cassette capacity. Four companies have launched new 35mm cameras since 2024.

Is shooting film worth it in 2026?

It depends on volume. At roughly $0.78 to $1.32 per frame including developing, film is worth it if you shoot deliberately and value the process and the highlight rendering. If you shoot thousands of frames a year or work to client deadlines, digital is the better answer.

Is film better quality than digital?

Not by most technical measures. Modern sensors beat 35mm film on resolution, noise and shadow detail by wide margins. Film's real quality advantage is highlight rolloff and organic grain structure. "Better" only means something once you name the attribute you care about.

How much does it cost to develop a roll of film?

At our Brooklyn lab, standard developing for 35mm or 120 is $18 a roll as of August 2026, with a 4K scan upgrade at $12. Turnaround is typically four to five business days. Membership starts at $18 a month and brings developing to $16 and scans to $6.

Can film simulations replace shooting real film?

For most commercial purposes, close enough. For three things, no: true halation physics through the emulsion layers, grain that is structurally the smallest detail in the frame rather than an overlay, and the fixed spectral response of dye layers.

Should a beginner start with film or digital?

Start with film if you want to understand exposure, because losing the back screen forces you to build a real model of light. Start with digital if you want volume, immediate feedback, or you are learning flash. Most photographers we serve end up shooting both.

How to start shooting film

Treating film vs digital as a fight is mostly a symptom of the internet. In Ilford's global survey of more than 6,800 film photographers across 100 countries, only 23 percent shot film exclusively. Everyone else ran both. That matches the shop floor: people carry a digital body for the work that has to be right and a film body for the work they want to feel like something.

If you want to try it, keep the first attempt cheap and boring. Buy a used point and shoot or a manual SLR, load a $10 roll of Kodak Gold 200, shoot it in daylight, and develop it before you buy anything else. One roll will tell you more than another month of reading comparisons. If it lands, our CineStill 400D review is a reasonable next stop for a second stock.

Ready to try film? Browse beginner-friendly bodies in 35mm cameras or the full used inventory in cameras, pick up a roll or three from film, and send it to us when you are done. Everything we develop is processed in-house at our East Williamsburg lab, and we take mail-in orders worldwide. Details and current pricing live on the film processing page.