Somewhere around year three or four, photographers tend to hit a problem they didn't see coming and can't name once it arrives. By then the technical side has stopped being the issue. Exposure happens without thinking. Anyone looking over your shoulder would say the pictures are good, and they'd be right about that.

Then you put a folder of your own work next to photographs you actually admire.

Nothing is wrong with yours. That's the trouble. They're clean and competent, and they could have come from anybody with the same camera and roughly the same amount of practice behind them.

Very little gets written about this stage, probably because nothing you can buy will shift it. Not a lens, not a preset, not another tutorial on settings. The pictures aren't failing technically. They just aren't taking a position on anything.

What voice actually is

Voice is the pile-up of decisions you keep making, mostly below the level of conscious thought.

Take one frame and count the choices in it. Most photographers can name two or three — subject, aperture, maybe the crop. The rest happen too fast to notice, and there are dozens. Whether you press now or wait four more seconds is a decision. So is standing where you stood instead of two paces left, which changes what ends up in the corners while your attention is on the middle. So is walking away, which nobody counts as a decision at all.

Photographers with a developed voice make those calls in a pattern somebody could pick out across twenty prints. Almost none of it is deliberate.

So you can't apply voice to a photograph. It shows up across a body of work, once there's enough of it to compare against itself. That's also why it takes so long. Ten frames won't show you anything. Two hundred made over a year of proper attention probably will, though even then you may need somebody else to point at it before you see it yourself.

Expression and voice are not the same animal.

Mixing these two up is where most photographers get stuck.

Style is the surface layer. Grade, contrast, crop, palette, the framing trick you reach for on autopilot. It lifts cleanly off one photographer and onto another.

Saul Leiter is the clearest case of this happening. He shot through rain-streaked and steamed-up glass, worked with a telephoto to compress the scene where most street photographers reached for something wide, and used expired color film that pushed everything toward muted, almost pastel tones. His faces are usually obscured or turned away. That look is now everywhere — there are editing presets named after his palette, and no shortage of photographers hunting reflections in wet windows.

Almost none of it lands the way his does, because what's been taken is the surface. Leiter worked the few blocks around his Manhattan apartment for over half a century, and by his own account, a window covered in raindrops held more interest for him than photographing someone famous. The technique came out of that position. Reverse the order, and you get the look with nothing underneath it.

So when photographers ask me how to find your photography style, the answer tends to disappoint them a bit. Go after the voice. Style follows on its own, and it's normally less dramatic than anyone was hoping for.

Start with whatever keeps stopping you.

Before technique, attention.

What makes you stop walking? Not what you think deserves a photograph, and definitely not what does numbers online. The question is narrower than that: what does your eye keep going back to when nothing is asking it to?

For some photographers it's hands, and they can't explain why. Others get caught by the half-second after a conversation ends, when a face drops back to neutral before anyone has decided what to say next. There are people who spend years photographing buildings losing ground to weeds. It sounds like a limitation until you see forty of them together.

That's not arbitrary. It comes out of your life — where you're from, what you did before this, what you've had cause to look at closely for years. Someone who worked a decade in kitchens reads a service counter differently. Someone raised on a farm sees a field differently from a visitor photographing the same field in nice light.

You don't have to be able to explain the pull, either. Josef Koudelka has said he never really worked out why he started photographing Roma communities, and wasn't especially interested in finding out. He just kept going back, on and off, from 1962 until 1971. The not-knowing didn't slow him down.

Most photographers can already tell you what they keep noticing. They wave it off as too ordinary to build on.

Stay put until the easy pictures run out.

First ten frames of anything are the obvious ones. So are the next twenty, if we're honest. The work worth keeping usually starts once you've used up every picture a stranger would take.

Go back to the same place enough times, and you get pushed past the surface, mostly out of necessity — the shot from the entrance is already made, so you have to find something else or waste the trip. What accumulates over those visits is a working knowledge of the place. You learn which corner has usable light at four, and that becomes the reason you show up at four.

Koudelka spent nine years on the Roma work, moving between settlements in Czechoslovakia, Romania, Hungary, France and Spain with a rucksack and a sleeping bag, sleeping outside. Leiter, at the opposite extreme of scale, stayed on the same handful of streets for fifty years. Neither of them was chasing coverage. They were both after the thing you only reach by being somewhere repeatedly.

Forty photographs of one subject across a few months will also show you what you do without thinking about it. The distance you work at, mainly, and whether people are in the frame or carefully kept out. You may find you've been consistently going for the quiet beat before a moment peaks rather than the peak itself — a lot of photographers do, and most have no idea until they see the evidence laid out.

Working by structured photography assignments gives that sustained attention a shape. Most of us can't hold it together without one.

Go through what you've already shot.

Before developing anything new, audit what's there.

Pull your last hundred frames. Not the portfolio hundred — all of them, including the ones you'd never show a soul. You're not grading quality here. You're hunting for repetition.

• Distance. Close in, or a step back?
• People. Do they turn up at all? If they do, are they aware of you, and do they hold the middle of the frame or drift to the edges? Plenty of photographers shoot people constantly without ever really photographing a person.
• Light. Soft or hard.
• Mood. Is there a register that keeps recurring — stillness, tension, humor, unease?
• Edges. What do you keep cutting out? This one usually tells you more than the other four put together. What a photographer refuses to include says as much as whatever ends up in the center, and most people have never once looked at their own work this way.

Most people discover they've been making the same three or four decisions for years and never named a single one. Naming them helps. It won't hand you a voice, but you'll know where you're standing. Going back over your recurring compositional choices will surface a few of these immediately.

Selection does more work than shooting.

Ask a photographer about editing, and you'll get a conversation about sliders. The kind which forms a voice is the older sense of the word: deciding what lives and what goes.

While you're shooting, you can put off the question of what any of it is for. Selection won't let you. A set of twelve pictures that belong together is making some kind of statement, whether or not you intended one. The same twelve chosen purely on individual merit will sit there saying nothing, and it's a strange thing to look at — every frame strong, the whole thing inert.

So take fifty frames from one stretch of work and cut down to twelve. Not the twelve best individually. The twelve that belong to each other. At some point your single strongest image will have to go because it fights the rest of them, and being able to do that without agonizing over it for a week is a reasonable sign something has shifted. You're judging against more than technical quality now.

Koudelka's nine years produced far more than any book could carry. Working with the publisher Robert Delpire, the 1975 edition of Gypsies came down to sixty photographs. A later edition, built from a different edit, ran to a hundred and nine. Same negatives, different set, and the two books do not say the same thing. That's what selection is: not finding the good ones, but deciding what the work is about and then paying for that decision in images you liked.

Then put them in an order and notice which sequence sits right. Make yourself explain why. Visual depiction starts roughly there, and the explanation is generally one fairly plain sentence regarding what you're interested in.

Study other photographers; don't harvest them.

Looking at work matters enormously. Looking at it badly is worse than not looking.

Copying takes the output — the grade, the framing, the subject. Studying takes the thinking behind it. When a photograph won't let you go, sit with it and work backward to the decisions. Why from that position, and why that instant rather than one a second earlier? What's outside the frame that the photographer decided you didn't need to see?

Whatever you come away with is rarely a technique. Usually it's a way of reading a situation, and that travels into completely different subject matter without producing anything derivative, because you took the reasoning and left the picture where it was.

Worth doing with work you can't stand, too. Getting precise about what irritates you sharpens a creative vision about as fast as admiration does.

Don't pin yourself down too early.

The moment a pattern shows up, there's a real pull to lock it in. You're a close-range high-contrast monochrome photographer now, and that's the whole brief from here on.

Resist it. Spotting a pattern early is information, not a contract, and treating it like one is how promising work goes stale inside a year.

Shoot everything from a distance? Spend a month up close. Live on available light? Try controlling it for a few weeks. Never photograph people? Go photograph people badly for a while. Some of these will just confirm what you already do, which is worth having, because a preference you've tested holds up better than one you fell into. Others crack open something you'd never have gone looking for.

Monochrome makes the point well. It only becomes a genuine decision once you know what the color was contributing and conclude the image works better on tone and form. Until then it's a default with an aesthetic bolted on.

Exercises worth doing properly

These dig out what's already in your work rather than bolting anything on.

• One subject, one month. Something within easy reach. Photograph it over and over for four weeks, and — this is the part everyone skips — keep going after it stops being interesting. The fortnight where you're bored and out of ideas is usually where the work turns.
• The hundred-frame review. Read your last hundred for repetition. Write down five patterns. Don't judge them yet.
• Fifty to twelve. Cut a set down, sequence it, then write one sentence regarding what it's for.
• Deliberate opposites. Two weeks working against your defaults.
• The exclusion list. One session, writing down what you chose not to photograph and why. Usually more revealing than the keepers, and almost nobody does it.
• Ten frames, one thread. Ten images from different periods. Find what links them. If nothing does yet, that's information rather than failure.

How you'll know it's happening

Voice arrives slowly enough that you'll miss the moment. The markers are recognizable, though.

Decisions come faster, and you can explain them when somebody pushes. Around the same time, you'll start binning technically strong frames because they don't belong with the rest, which feels wasteful for about a year and then stops feeling like anything.

The one that surprises people is the resistance — standing in front of something easy and not wanting to shoot it, because it isn't yours. That usually shows up before you could say what “yours” means.

Then there's the one that matters most, which is the work holding together even when the subject changes completely. That consistency is the thing itself — whatever stays put while everything else moves.

No timetable applies. What speeds it up is volume of considered work, plus somebody honest enough to tell you an image isn't as good as you think it is. Koudelka had that — while he was making the Roma pictures, he showed them to Markéta Luskačová and Dagmar Hochová, both photographers, specifically for criticism. He was nine years into the strongest work of his life and still putting it in front of people who would tell him the truth.

That pairing is roughly what structured project work exists for. If you'd rather not do it alone, The Artist's Lens runs through building meaningful photographic projects, taking you from individual pictures toward a coherent body of work using assignments and critique.

What it asks of you

None of this shows up as a revelation. It accumulates, slowly, out of sustained attention to subjects you actually care about and an honest look at what you've already got. Mostly it needs enough work over enough time that a pattern has somewhere to appear.

Leiter was in his seventies before most people had heard of him. Koudelka photographed for nine years before the first book existed. Neither of them picked a style and then executed it — they photographed what interested them, carefully, for a very long time, and the consistency only became obvious in hindsight. Often to other people before it was obvious to them.

Start with what keeps stopping you. Stay with it long past the point where it's easy. Then look hard at what you've made and be honest about what it's telling you.