Production I: making the work

Production is the stretch of the term when the plan meets reality. For most capstones it runs from about week 4 to week 9, with the main shoots finished by week 7 so that the later weeks can be used for reshoots and gaps found in the edit. This lesson is about working methods during production: keeping a series consistent while letting it develop, working with the people in the images, capturing the information you will need later, reviewing as you go and tracking whether the project is on course. The technical skills themselves were taught earlier: exposure and light in PHOT-101 and PHOT-102, sensors and noise in PHOT-201, portable and studio lighting in PHOT-210 and PHOT-211, portraiture in PHOT-220 and fieldwork in PHOT-230. Here the question is how to deploy them across many shoots so that the result reads as one body of work.

Making a series rather than single images

A series is not a collection of good pictures on the same subject. It is a set of images that share an approach, so that each one gains meaning from the others. The approach includes decisions about distance, framing, lens, light, colour, format and the relationship between photographer and subject. Fixing some of these decisions early creates consistency; leaving others open allows variety and discovery. The art is in choosing which to fix.

Amara fixed several variables for her portraits: a 35 mm equivalent lens, the subject at about two metres, eye level, the subject facing the camera in their usual working position, one battery LED as key light from the subject's left, and a white balance set to match the ambient light so that the night colour stayed visible. She left open the pose, the props and the background, which came from each workplace. For the empty interiors she fixed a tripod height of 1.5 metres, a slightly wider lens and long exposures without added light, so the spaces carried the colour of their own fixtures.

A coverage plan for the series

A shot list for a series is less specific than for a commercial job, but it should still define the kinds of images the body of work needs. Amara's coverage plan was a grid: five sites across the top and image types down the side (portrait, interior, detail, transition such as arrival or departure, and a wider establishing view). The target was at least one strong portrait and one strong interior from each site, with details and transitions optional. Filling in the grid after each review shows where the gaps are long before the edit.

Working with people

Most capstones involve people, and the quality of the relationship shows in the images. Before the first shoot at a site, introduce yourself and the project to everyone who might be photographed, not only the manager who approved access. Explain what you will do, how long it takes, how the images will be used and that people may say no or stop at any time. Ask about routines so that you can work around them; a baker cannot pause when the dough is ready. Respect time limits. Amara's test shoot had shown she had about three minutes with each worker, so she set up the light before asking, made a few frames, showed the person the back of the camera and finished within the time.

Reciprocity matters. People give their time and their image; offer something in return that suits the setting and avoids pressure, such as a print, a copy of the book dummy, an invitation to the opening, or images the workplace can use internally if the agreement allows. Do not pay in ways that compromise consent or documentary integrity, and be honest about what the project can and cannot do for them. Keep promises: if you said you would bring prints in two weeks, bring them in two weeks. PHOT-401 lesson 5 and PHOT-320 lesson 11 discuss consent, representation, trauma and vulnerability in more depth; apply those principles whenever subjects are minors, unwell, in crisis or in a dependent relationship with someone who approved your access.

Technical consistency at night

Night and low-light work raises technical problems that can break a series if they are handled differently from shoot to shoot. Mixed light sources (LED, fluorescent, sodium vapour, screens) produce colour casts that vary within a frame. Low light forces a choice among higher ISO and more noise, slower shutter speeds and motion blur, and wider apertures and less depth of field. A portable light adds control but can flatten the mood if too strong. Decide on a consistent treatment and record your settings so that you can reproduce them.

Amara exposed portraits at ISO 1600 to 3200 with the LED as key and ambient as fill, and kept shutter speeds at 1/60 second or faster for people. She accepted visible ambient colour casts as part of the series and corrected only skin tones toward a natural hue in post-production. For interiors she used ISO 100 to 400 on a tripod with exposures of several seconds, which kept noise low and smoothed any movement. The noise characteristics of her camera and the denoising methods from PHOT-201 and PHOT-420 lesson 8 told her where her ceiling ISO lay.

Capturing information at the shoot

Many problems in the later stages trace back to information not captured at the shoot. Record, for each session: date, place, time, people present and their names and roles, releases signed, any conditions or requests, technical notes and the file number range. A small notebook or a phone note template works; voice memos are quick but must be transcribed. Write captions while the facts are fresh: who, what, where and when, in the subject's own words where possible. IPTC metadata (lesson 7) can be added at ingest from these notes, so a consistent template saves hours.

Daily and weekly review

Review every shoot within a day: ingest and back up (lesson 7), make a contact sheet, flag candidates and write a log entry. Once a week, lay out the strongest candidates from all shoots, physically or on screen, and look at them together. Weekly review is where you see whether the series is cohering, whether one site or image type dominates, and whether the fixed variables are still serving the work. It is also where you notice that you have been making the same picture again and again, which is a common sign that it is time to change something.

Iteration is part of the method. If a choice is not working, change it deliberately and record why. Amara found after week 5 that her empty interiors looked like real estate photographs because they were too evenly lit and too tidy. She moved the camera lower, included traces of work such as a mop bucket or a coat on a chair, and lengthened exposures to show the movement of fans and screens. The change made the interiors feel inhabited even without people, and the log records the week and the reason.

Tracking yield and progress

Track the number of strong candidates against the number the final output needs. Amara needed about 20 images on the wall and about 35 in the book, and she expected to discard perhaps half of her strong candidates in the edit for reasons of sequence, repetition or consent. That meant a target of about 60 strong candidates by the end of production. Counting candidates each week and comparing with the target shows early whether more shoots, different shoots or a change of approach are needed.

import random
random.seed(4991)
yield_per_shoot = [3, 2, 4, 1, 3, 5, 4, 2]  # strong candidates from the first eight shoots
mean = sum(yield_per_shoot) / len(yield_per_shoot)
have, target, nights_left = sum(yield_per_shoot), 60, 10
runs, hits = 10000, 0
for _ in range(runs):
    total = have + sum(random.choice(yield_per_shoot) for _ in range(nights_left))
    hits += total >= target
print(f'mean per shoot {mean:.2f}, have {have}, expected at end {have + mean * nights_left:.0f}')
print(f'chance of reaching {target}: {hits / runs:.0%}')
needed = (target - have) / mean
print(f'nights needed at the mean rate: {needed:.1f}')

mean per shoot 3.00, have 24, expected at end 54
chance of reaching 60: 8%
nights needed at the mean rate: 12.0

The resampling simulation draws future shoots from the yields Amara has already seen. With ten nights left, the expected total is about 54 candidates and the chance of reaching 60 is low. She has three options: add shoot nights, raise the yield per night by concentrating on the image types that work best, or lower the target by accepting a smaller book. She raises the yield by spending interior nights at the sites where interiors had worked and adds two nights in week 8.

When production goes well

Momentum is valuable but brings its own risks. When shoots are going well it is tempting to keep shooting past the schedule because each night produces something new. Remember the backward schedule from lesson 1: the last main shoot date is set by output lead times. More images also mean a longer edit. If you find a promising new direction late in production, write it down as a possible next project rather than trying to fold it into this one.