Assignment: Landscape, Sky, Foreground and Filters

The intention

Make two landscape negatives of the same subject — one on paper, where the sky is going to be white and you will compose for that, and one on panchromatic sheet film behind a filter whose factor you have applied and then verified — and settle, with numbers, what a filter is actually worth on a camera working at f/200 and beyond.

Part IV promised that the filter numbers for the filters this course uses would be given in Part VII. This is that page. It also closes a question the earlier assignments kept postponing: on a blue-sensitive material there is no filter route to a dark sky at all, and the reason is worth more than the remedy.

Learning objectives

By the end of this session you should be able to:

• Quote a filter factor from the sheet for the film in the camera, and explain why the same filter carries different factors on three films from the same maker.
• Compound a filter factor with a reciprocity correction in the right order, and say why the combined cost of a red filter on a long pinhole exposure is nearly fourteen times rather than eight.
• Design a bracket that measures a filter factor without the reciprocity correction contaminating it, and run it.
• Predict the sky, the foliage and any red object on paper and on filtered panchromatic film, and attribute each prediction to a spectral sensitivity rather than to a rule of thumb.
• Place a horizon and a foreground on a flat back knowing what the cosine-fourth law will do to an even sky.
• Log a changing light as a range rather than as a number.

Prerequisites

Portrait and Self-Portrait precedes this page in the sequence. Part IV’s spectral sensitivity lesson owns why a filter has a factor at all, and gives the ILFORD tables this page extends to Kodak’s panchromatic films. Part VI’s exposure and reciprocity lesson owns the correction, and its processing lesson owns sheet-film loading, development and the maker’s time–temperature table. Do not attempt the film route until you have processed a sheet by that page’s method at least once.

Safety classification

Level A. The paper route uses the same three baths as every previous assignment. The film route adds a film developer, handled by the owning lesson’s controls. No chemical on this page is new to the course.

• Substances. Paper developer, film developer, stop and rapid fixer at working dilution, indoors.
• Energies. The sun, for a long outdoor session, and the terrain you cross to reach the subject.
• Procedures. A tripod on open ground; exposures of minutes during which you are not looking at the camera; sheet film loaded and unloaded in the dark.
• Waste. Used fixer is silver-bearing and is collected; used film developer is collected by the route the processing lesson sets out.

What is not a hazard here, and why. The filters are the new object on this page and they are inert: a gelatin or polyester filter is a dyed sheet, it is not consumed by the exposure, it emits nothing, and it is handled dry with clean fingers on its edges. The reason it wants care is optical rather than chemical — a fingerprint on a filter 6 mm behind the pinhole is directly in the beam — so the control is a lens tissue and a sleeve, not a glove. The real hazards on this page are all outdoors and all about where you put yourself, which is why the section below is about ground and weather.

Hazards

Terrain. These are the course’s own conditions for this assignment, stated as conditions rather than as guidance from an authority. Plan the route and the return in daylight even if the exposure runs into dusk. Set the tripod back from any edge, and not on wet rock, loose scree, a mud bank or an undercut lip. Do not go closer to an edge to improve a foreground. For coastal work, check the tide before you go and know where the water will be in two hours. Tell somebody the route and the time you expect to be back, and take a means of calling for help that works where you are going.

Weather and daylight. A landscape session is planned around the light, which means it is planned around the weather. Check the forecast, carry more layers than the walk needs, and set a turnaround time before you leave.

Sun. ICNIRP’s shadow rule applies: shadow shorter than you means the ultraviolet is at its strongest, which under clear skies in late spring and summer is the four hours around midday. Note that this is exactly the period this assignment most wants you outdoors, so the protective measures are not optional decoration.

Never sight the sun. Placing a flare hood means judging where the sun is relative to the frame. Do that from the shadow the hood casts, or by holding a card up and watching where its edge shadow falls, and never by looking towards the sun along the camera’s axis. The hazard is retinal and it is not reversible.

Required PPE

• Sun protection — hat, sleeves, sun cream, and shade between sheets — is the protective measure for the outdoor half of this session, and it is the one most often skipped.
• Footwear with grip and a spare layer are protective equipment on a landscape page in the way gloves are protective equipment in the darkroom.
• Back indoors: nitrile gloves, eye protection and one pair of tongs per tray, plus the controls the film-processing lesson specifies for its developer.

Ventilation

Not among the outdoor controls. Indoors, an openable window or an extractor rather than a sealed cupboard, which is ILFORD’s general instruction; the film developer’s own sheet governs its handling, and at these dilutions none of the baths generates a vapour that calls for extraction.

Materials

Item	Quantity	Notes
Variable-contrast RC paper	10 sheets	The paper route, plus the filter-on-paper test.
Panchromatic sheet film	6 sheets	The film route and the factor bracket. Any film whose maker publishes a filter table.
Yellow filter, Wratten 8 or equivalent	1	Gelatin, polyester or a cut square from a lighting gel.
Deep yellow or orange filter, Wratten 15 or equivalent	1	Note what its maker actually calls it; see the caution below.
Red filter, Wratten 25 or equivalent	1	The one that does most and costs most.
Card for a flare hood	1 sheet	Matt black on the inner face.
A lux meter or calibrated phone app	1	Readings before and after every exposure.
Prediction sheet, folded	2	Predictions on the left, observations on the right.

Chemicals

Chemical	Quantity	Form
Paper developer concentrate	100 mL	Diluted 1+9 to make 1 L; 60 seconds at 20 °C for RC paper.
Film developer	per its own sheet	Dilution, time and temperature from the maker’s table, as the processing lesson sets out.
Stop bath concentrate	50 mL	Diluted 1+19 to make 1 L. A citric acid stop.
Rapid fixer concentrate	300 mL	Diluted 1+4; 30 seconds for RC paper, and the sheet’s own time for film. An ammonium thiosulfate fixer.
Water	about 20 L	Dilution and washing.

Equipment

The camera with its flat back at two focal distances, 50 mm for the foreground-dominated frame and 120 mm for the compressed one; the pinhole plates and register; the filter carrier or a way of holding a gel flat behind the pinhole plate; a tripod; a lux meter; a timer; sheet-film holders and a changing bag; and the processing kit.

Estimated cost

Cost band £, rising towards ££ if you buy filters new. Gel offcuts from a theatrical lighting supplier are a fraction of the price of photographic filters and are perfectly usable behind a pinhole, with the caveat below about knowing what you actually have. Dated prices live in the laboratory planner.

Waste streams

Used fixer is silver-bearing, and so is the first change of wash water; both are collected in the labelled container rather than poured away, because silver in a thiosulfate complex is a soluble metal load. Used film developer is collected by the route the processing lesson sets out. Local regulations govern where the containers go, and they differ between authorities even within the United Kingdom.

Preparation

1. The sky problem, stated as a decision rather than a defect

Part IV establishes the mechanism and this page will not repeat it: an undyed silver halide emulsion responds in the blue and the ultraviolet and stops around 500 nm, so a blue sky is, to a paper negative, the brightest thing in the landscape by a very large margin, and green foliage is nearly as dark as a shadow. Expose for the land and the sky goes past the shoulder of the curve and records as undifferentiated maximum density, which on the working positive is paper-base white with no cloud in it.

You have exactly two routes, and this assignment makes you take both.

Route A, paper, accepting the white sky. Compose so that the sky is doing a job rather than failing at one: put the horizon high and let the sky be a small bright shape; or put it low and let a large white field carry a silhouette. What you may not do is compose as though the sky will have tone, and then be disappointed.

Route B, panchromatic film with a filter. The film records red and green as well as blue; the filter subtracts blue from what reaches it; so the sky, which is mostly blue, darkens relative to everything else. This is the route that has numbers, and the numbers are below.

⁇Is the white sky a fault the material has, or a rendering the material offers?

The same physical fact has been read both ways by named parties a hundred and thirty years apart.

P. H. Emerson, 1889, treats it as a fault with a technical remedy. Answering the charge that photography “is false in local colour, putting all the lights and darks of natural colouring out of tone”, he replies that “with the aid of orthochromatic plates it does no such thing” — that is, the colour-blind rendering is an error, and the answer is a better plate. In the printing chapter he allows one repair and only one: cloud printing is “the simplest form of combination printing, and the only one admissible when we are considering artistic work”, while Rejlander’s fuller combination printing he compares to “cutting out figures and pasting them into white spaces left for that purpose in a picture-book”.

ILFORD, on a current data sheet, treat the same behaviour as a rendering to be chosen. ORTHO Plus is sold with the statement that its lack of red sensitivity means reds appear much darker than normal, and that is offered as a usable effect. Nobody at HARMAN is claiming the film is more truthful than a panchromatic one; they are claiming it looks like something.

What the course uses, and why. It refuses to call the paper rendering either true or false, because both words import a claim about what a photograph is for. What it insists on is that you predict the rendering before you expose and say in the log which of the two positions you were taking. A white sky you predicted is a decision. A white sky that surprised you is a mistake, whatever Emerson would have called it.

2. The filter numbers, at last

A filter factor is the number you multiply the exposure by to compensate for what the filter removes. Kodak print factors per film, not per filter, and comparing three of their own sheets shows why that is not fussiness.

KODAK WRATTEN filter	TRI-X 400, daylight	T-MAX 400, daylight	T-MAX 100, daylight	T-MAX 400, tungsten
No. 8 (yellow)	2	1.6	1.5	1.3
No. 11 (yellowish green)	4	4	—	3
No. 12 (deep yellow)	2.5	2	—	1.3
No. 15 (deep yellow)	2.5	2	2	1.3
No. 25 (red)	not read*	8	8	4
No. 47 (blue)	not read*	10	8	20
No. 58 (green)	not read*	6	6	6
Polarising	2.5	3	2.5	2.5

Quoted from the three Kodak publications named in the sources, with the T-MAX 100 column taken from the table Part VI already prints. Kodak’s own warning goes with it: “Filter factors for other Kodak black-and-white films are different.”

*Three cells the course will not fill in. In the copy of the TRI-X sheet this course read, the daylight and tungsten columns of the red, blue and green rows run together in the extracted text, so the digits could be read more than one way. A number that might be an 8 or might be an 85 is not a citation, so those cells stay empty rather than being guessed at from the neighbouring films. If you have the original sheet, read them off it and fill them in your own copy of the table.

Three things to read out of what is here.

Yellow varies and red does not. The No. 8 yellow costs a full stop on TRI-X, two-thirds of a stop on T-MAX 400 and about half a stop on T-MAX 100 — a spread of half a stop between three films from the same maker, arising from where each emulsion’s blue sensitivity sits relative to the filter’s cut. The No. 25 red is 8 on both T-MAX films despite their two-stop speed difference, because a deep red filter removes so much of what any ordinary panchromatic emulsion uses that the residual differences between them stop mattering.

The tungsten column is the proof that a factor is not a property of a filter. Red costs 8 in daylight and 4 under tungsten; blue costs 10 in daylight and 20 under tungsten. Same filters, same film, and the numbers move in opposite directions, because tungsten light has more red in it and less blue. A filter factor belongs to a filter, a light source and a material together.

Extended-red film is a different animal again. Part IV gives ILFORD’s table for SFX 200, where the No. 25 red costs only 2.8, because SFX still responds beyond 700 nm where the filter is transparent and the eye is not. Do not carry a factor across from that table to an ordinary film.

✕What this course could not verify, and will not guess

There is no manufacturer factor here for a true orange filter on an ordinary panchromatic film. Kodak’s TRI-X and T-MAX sheets tabulate No. 8, 11, 12, 15, 25, 47 and 58; they call No. 15 a deep yellow, not an orange, and No. 16 and No. 21 — the numbers usually sold as orange — do not appear. ILFORD’s SFX sheet does give a No. 21 orange at 2.4, but that is an extended-red film and its factors transfer to nothing else.

So if your filter is orange, the course has no published number for it and will not invent one. Measure it by the bracket in section 4, which takes one extra sheet and gives you a factor that is true for your filter, your film and today’s light, which is better than a table anyway.

The same caution applies to lighting gels. A theatrical gel is specified by its transmission for lighting purposes, not by a photographic factor, and two gels of the same colour name can differ substantially. Buying gel is a good economy; assuming a Wratten factor applies to it is not.

3. Why paper and a red filter is a safelight

Now apply the same logic to the paper route, and notice that it produces an answer nobody has to measure.

Paper is blue-sensitive and stops around 500 nm. A deep red filter passes essentially nothing below that. Putting a red filter in front of a blue-sensitive material is the definition of a safelight — it is what a safelight is, and ILFORD’s safelight sheet matches red and light-brown filters to exactly these materials for exactly this reason.

That gives the honest statement in place of a number. The course has not measured a filter factor for a red filter on paper, and no manufacturer publishes one, because the combination is designed to produce no useful exposure. It is not infinite: Kodak’s K-4 makes the point that the colour sensitivity of an emulsion does not end abruptly at a wavelength and that most materials retain some sensitivity to what a recommended safelight transmits, which is why safelight specifications carry a distance and a time. But the factor is large enough that the exposure would be measured in hours and the picture would be a fogged grey.

Three consequences, and they close the question the part opened.

1. On paper, a yellow filter is expensive and does the wrong thing. It removes blue, which is what paper uses; the sky stays the brightest thing in the frame relative to everything else, and you have paid a large unknown factor for nothing. The direction is certain even though the size is not.
2. On paper, a blue filter is nearly free and equally useless, because it passes what the material was going to use anyway.
3. Therefore there is no filter route to a dark sky on a blue-sensitive material. Either accept the white sky, or change the material. That is the whole of it, and it is why this assignment is two negatives rather than one.

🧪A one-sheet test that settles it for your own paper

Expose two paper sheets of the same scene at the same time: one bare, one behind the red filter with thirty times the bare exposure. Process them together.

The prediction. The bare sheet is a normal negative. The filtered sheet, at thirty times the exposure, is still substantially thinner than the bare one — because thirty is nowhere near the true factor.

What it is worth. It converts an argument into an observation, it costs one sheet, and it gives you a lower bound on the factor that you can quote in your own log: “greater than 30 on MULTIGRADE RC in daylight, measured”. That is a real result, of a kind nobody publishes.

4. The bracket that separates the factor from the reciprocity correction

Here is the difficulty that makes filter work on a pinhole camera different from filter work on a lens camera. A filter factor is applied to the metered time. The reciprocity correction is applied to the whole metered time, after the factor, because every published correction is indexed on the total. So the two multiply, and because the correction is itself a power law, the combined cost is larger than the factor.

🧮DO THE MATHS: a red filter that costs three stops and takes four

The reading. A meter gives f/16 at 1/60 s for ISO 125 film — ILFORD’s rating for FP4 PLUS.

To the pinhole’s aperture. From f/16 to f/250 is a factor of (250/16)² = 244.1, so the metered time at f/250 is (1/60) × 244.1 = 4.07 s.

Unfiltered, corrected. ILFORD publish P = 1.26 for FP4 PLUS, so T_c = 4.07^1.26 = 5.86 s.

With the No. 25 red, factor 8. Apply the factor first: 4.07 × 8 = 32.6 s metered. Then correct: 32.6^1.26 = 80.5 s.

Now compare the two exposures you would actually give: 5.86 s bare against 80.5 s filtered. That is a ratio of 13.7, not 8. The filter’s nominal cost is three stops; its real cost on this exposure is 3.8 stops.

Why. Raising the time by a factor of eight moves you further into the region where the correction bites, and the correction is a power greater than one, so it amplifies whatever it is given. Written out, the ratio is 8^1.26 = 13.7 — the factor raised to the exponent. On any material with a published exponent, the true cost of a filter on a long exposure is the factor raised to the power P. That is a general result and it belongs in your log’s filter field.

What a filter really costs once the reciprocity correction is applied after it

• No reciprocity failure — the naive answer
• FP4 PLUS, published P = 1.26
• HP5 PLUS, published P = 1.31
• A severe material, P = 1.6

Show the numbers behind this plot

Four rising curves showing the true multiplier on the corrected exposure against the filter factor printed by the manufacturer, over factors from one to ten. The lowest and straightest is the reference case of a material with no reciprocity failure, where the true multiplier simply equals the published factor: a factor of two costs two, four costs four, eight costs eight and ten costs ten. Above it is ILFORD FP4 PLUS with its published exponent of one point two six, where a factor of two really costs two point four, a factor of two point five costs three point one, a factor of four costs five point six, a factor of eight costs thirteen point seven and a factor of ten costs seventeen point eight. Slightly higher again is ILFORD HP5 PLUS with its published exponent of one point three one, where the same factors cost two point five, three point three, six point two, fifteen point two and nineteen point nine. The steepest curve is a material with a severe exponent of one point six, in the region Kodak's own TRI-X table implies, where a factor of two costs three point zero, four costs nine point two, eight costs twenty-seven point nine and ten costs thirty-nine point eight. The gap between the reference line and the others is the whole point: for a mild filter the difference hardly matters, but for a deep red filter of factor eight the exposure is nearly twice the naive answer on FP4 and more than three times it on a severe material. A marker sits on the FP4 curve at a factor of eight and a true multiplier of thirteen point seven, the worked example on this page.

Series	Published filter factor	Actual ratio of corrected exposures, filtered to bare
No reciprocity failure — the naive answer	1.00	1.00
No reciprocity failure — the naive answer	2.00	2.00
No reciprocity failure — the naive answer	2.50	2.50
No reciprocity failure — the naive answer	4.00	4.00
No reciprocity failure — the naive answer	6.00	6.00
No reciprocity failure — the naive answer	8.00	8.00
No reciprocity failure — the naive answer	10.00	10.00
FP4 PLUS, published P = 1.26	1.00	1.00
FP4 PLUS, published P = 1.26	2.00	2.40
FP4 PLUS, published P = 1.26	2.50	3.11
FP4 PLUS, published P = 1.26	4.00	5.66
FP4 PLUS, published P = 1.26	6.00	9.42
FP4 PLUS, published P = 1.26	8.00	13.70
FP4 PLUS, published P = 1.26	10.00	18.20
HP5 PLUS, published P = 1.31	1.00	1.00
HP5 PLUS, published P = 1.31	2.00	2.48
HP5 PLUS, published P = 1.31	2.50	3.26
HP5 PLUS, published P = 1.31	4.00	6.15
HP5 PLUS, published P = 1.31	6.00	10.50
HP5 PLUS, published P = 1.31	8.00	15.20
HP5 PLUS, published P = 1.31	10.00	20.40
A severe material, P = 1.6	1.00	1.00
A severe material, P = 1.6	2.00	3.03
A severe material, P = 1.6	2.50	4.26
A severe material, P = 1.6	4.00	9.19
A severe material, P = 1.6	6.00	18.10
A severe material, P = 1.6	8.00	27.90
A severe material, P = 1.6	10.00	39.80

Computed as factor raised to the power P, using the exponents ILFORD publish; nothing here is measured. It holds only where the whole exposure is long enough for the correction to apply — under a second, ILFORD state that no compensation is needed and the naive line is the right one. The curve is drawn to show the shape, not measured from a real material. Your own materials will differ, and measuring them is what the sensitometry part of the course is for.

The bracket that isolates the factor follows straight from that last sentence. If both exposures are one second or shorter, ILFORD’s own statement is that no compensation is required, so the ratio you measure is the factor and nothing else.

🧪Measuring a filter factor cleanly, in four sheets

Hypothesis. The published factor for this filter, on this film, in this light, is F.

The control. The unfiltered exposure, made within a minute of the filtered one on the same scene in the same light.

The one variable. The filter.

The design. Use the fastest film you have and the largest hole, to force both exposures under a second. With HP5 PLUS at ISO 400 and the 0.40 mm plate at f = 50 mm, which is f/125, bright sun at 99,200 lux gives a bare exposure of 125² × 310 / (99,200 × 400) = 0.12 s. Multiply by a factor of 8 and the filtered exposure is 0.98 s — just inside the no-compensation region.

The sheets. One bare at 0.12 s. Three filtered, at the published factor and one stop either side: 0.49 s, 0.98 s and 1.96 s. Process all four together.

The reading. Whichever filtered sheet matches the bare one in density gives the measured factor. The resolution of the method is one stop unless you interpolate by eye, which is honest to about half a stop — so this measures a factor of 8 to something like 6 to 11, not to 8.0. Say so in the log, and note that a factor known to half a stop is entirely good enough, because a factor is being applied to an exposure you are bracketing anyway.

What it cannot do. It cannot give you the factor at the light level you will actually use in the evening, because the factor depends on the illuminant. Measure it in the light you photograph in, or carry the uncertainty.

5. Mounting a filter behind the pinhole, and the flare hood

A pinhole camera has nowhere to put a filter in front of the aperture, so it goes behind the plate, between the plate and the film, held flat against the inside of the front panel. Cut it larger than the plate’s window by a good margin so that its edge is nowhere near the light cone.

Filter behind the pinhole plate, and the flare hood in front of it

1. 1Flare hood, depth L, matt black inside — clear width must exceed 2.54 L at f = 50 mm, 2.03 L for the short sheet dimension, 1.06 L at f = 120 mm
2. 2Pinhole plate in its carrier — unchanged; the register letter still governs
3. 3Filter, flat against the inner face — cut generously oversize; held at the edges; kept clean, because it is inside the cone
4. 4Film plane at f — the filter changes nothing about the geometry — only the spectrum

Compute the hood before you cut it: 2 tan θ, where θ is the half-angle to the corner, is the clear width per unit depth. A hood that vignettes is worse than no hood, because it puts a hard edge on the frame where the cos-fourth falloff was gradual.

The hood’s geometry is one line of trigonometry. At f = 50 mm the corner of a 127 mm sheet sits 51.8° off the axis, so a hood of depth L needs a clear width of at least 2 tan 51.8° = 2.54 L across the long dimension. A 20 mm hood therefore needs a 51 mm clear opening. At f = 120 mm the same sheet’s corner is only 27.9° off axis and the requirement falls to 1.06 L, so a deep hood is practical on the long setting and nearly useless on the short one. Compute it before you cut it: a hood that intrudes into the field puts a hard black edge on the picture, which is a far worse fault than the flare it was meant to prevent.

6. Foreground, horizon and the corner of an even sky

Two geometric facts from Part VI decide landscape composition on this camera, and both are numbers rather than opinions.

The near-far stretch belongs to the focal distance. Image size is h′ = h f/u, so at f = 50 mm a rock 0.5 m from the pinhole is rendered at ten times the scale of the same rock at 5 m. That exaggeration is the pinhole landscape’s most recognisable device, and it is available only at the short setting: at f = 120 mm the same pair of rocks would need the near one at 1.2 m to fill the same fraction of the frame, and the whole effect softens.

The corner falloff is worst exactly where the sky is most even. The cos-fourth law puts the corner of a 4 × 5 sheet at f = 50 mm 3.73 stops down, and Part VI notes that real falloff is generally worse than cos⁴ because of the plate’s own thickness. In a textured foreground nobody notices; in a clear sky it reads as two dark upper corners. So the composition question is which corner of the frame the sky occupies, and the honest choices are: fill the sky with cloud, so the falloff hides in the modelling; put the horizon high, so there is little sky in the corners; or use the long setting, where the corner is only 1.09 stops down.

Where to put the horizon, at the two focal distances, given the corner falloff

1. 1f = 50 mm — corner 58.4°, 3.73 stops down — horizon high; foreground object at 0.5 m carries the frame; sky kept out of the corners
2. 2f = 120 mm — corner 27.9°, 1.09 stops down — horizon low; an even sky is survivable here and nowhere else on this camera
3. 3The falloff wedges — shaded where the loss exceeds about two stops — visible on a plain sky, invisible in cloud or texture

Corner angles and losses from Part VI's cos-fourth table for a 101.6 × 127 mm sheet. The compositions are the course's own recommendation, not a rule: what is not negotiable is that you decide where the sky goes knowing the number.

7. Logging a light that will not sit still

A landscape exposure of minutes is an exposure during which the light changes. Kodak’s own daylight table, read through the constant Part VI derived, puts numbers on how far it can move: bright sun with distinct shadows is about 99,200 lux, weak hazy sun 46,900, cloudy bright 24,800, and heavy overcast or open shade 12,150. That is a range of three stops between two conditions that can occur ten minutes apart.

So the light field in the log stops being a number and becomes a range with two timestamps: measure at the start of the exposure, measure at the end, and record both. If the two readings differ by more than half a stop, the exposure you gave was not the exposure you computed, and the entry should say so. This is the single field most likely to explain an inexplicable negative three weeks later.

8. Write the predictions, then fold the sheet

On the prediction side, before anything is exposed: the composition and where the sky sits; the measured lux and the computed times for both routes; the filter and its published factor, the exponent you will use and the compounded cost; the predicted tone of the sky, the foliage and any red object on each material; the corner falloff in stops for the focal distance you chose; and the bracket times. Fold it.

The making

Stage 1 — Arrive, place, measure, log (30 minutes)

Set up, level the camera, place the horizon by the template, measure the light, and write all ten before-the-shutter fields for every sheet you intend to expose. The order matters: everything computed before the first exposure means the light has changed as little as possible between the computation and its use.

Stage 2 — The paper route (40 minutes)

Two sheets at the short setting with the foreground doing the work, at the computed time and one stop over. One further sheet at the long setting for the comparison. Fit the flare hood if the sun is within about 20° of the frame edge, and judge that from the hood’s shadow rather than by looking.

Stage 3 — The film route with the filter (50 minutes)

Reload with sheet film. Expose bare, then filtered at the compounded time, then filtered one stop either side. Re-measure the light between every sheet and log it. If cloud crosses during an exposure, write down when and for how long — that observation is worth more than the sheet.

Stage 4 — The factor bracket (25 minutes)

Run the four-sheet design from section 4 in the brightest light available, with the largest hole, so that every exposure is at or under a second. This is the only part of the session where the pictures do not matter.

Stage 5 — Process (40 minutes)

Paper first, by the method the earlier assignments used. Then the film, by the processing lesson’s route and the maker’s own time–temperature table.

One development decision belongs here. ILFORD state that contrast is increased with long exposures, because the shadows sit at a lower illuminance and suffer the failure more, and their remedy is to pull the development. Kodak put numbers on the same instruction for TRI-X: 10 per cent less development at one metered second, 20 per cent at ten, 30 per cent at a hundred. If your filtered exposure was 33 metered seconds, a reduction of about a quarter is the defensible reading of that table. Log it as your arithmetic on their instruction, not as a figure they printed.

Stage 6 — Read, then unfold (15 minutes)

Rank the tones on both negatives before unfolding the predictions.

Expected observations

On the paper negative. A sky at or near maximum density, with no cloud separation. Foliage dark, sometimes nearly as dark as shadow. Any red object — a barn door, a post box, a coat — much darker than it looks. Two visibly darker upper corners if the sky is plain.

On the bare film negative. A sky that is lighter than the eye expects, because a panchromatic film is more blue-sensitive than the eye is and a clear sky is a bright blue field. Cloud barely separated from it.

On the filtered film negative. The sky substantially darker relative to the land, cloud standing out of it, distant haze reduced, and green foliage a little darker than on the bare sheet. The stronger the filter, the more of all four.

On the factor bracket. One of the three filtered sheets close to the bare one. If none of them is close, the factor is further from the published number than a stop, which is a result worth reporting rather than a failure.

Evaluation and reconciliation

1. Did the compounded exposure land? Compare the filtered sheet you judge correct with the time you computed as F^P × the bare corrected time. State the discrepancy in stops.
2. What factor did the bracket actually measure, and how does it compare with the published one? Give your answer with its half-stop resolution attached.
3. Did the sky behave as predicted on each material? Rank sky, cloud, foliage and the red object on both negatives and compare with the written ladder.
4. Measure the corner falloff. Judge the density at the centre and at both upper corners against a grey scale, convert to stops, and compare with the cos-fourth prediction for your focal distance. Part VI’s experiment page establishes that by-eye judging resolves about half a stop, so a difference of 3.7 stops is well within reach and a difference of 1.1 is marginal.
5. How far did the light move during the session, in stops, from your two-timestamp readings, and which negative does that explain?

The critique

One paragraph per submitted picture, naming causes. Answer in order: the intention as written; which optical or spectral fact produced the look; whether the exposure landed and by how much; whether the filter did what the factor said it would; where the falloff is visible and whether it helps; and the single change you would make.

Then one further sentence that belongs only to this page: say whether the picture would be better or merely different on the other material. Most students find that the honest answer for at least one of their two negatives is “different”, and noticing that is worth more than preferring one.

Troubleshooting

What you see	Most likely cause	What to check, and what to change
Filtered film negative far too thin	The correction applied before the factor, or not at all	Recompute as (bare metered × F) raised to P; the order is not interchangeable
Filtered negative thin by roughly a stop and no more	The published factor is not right for your filter or light	Run the four-sheet bracket; a gel is the usual culprit
A hard dark band along one or two edges	The flare hood is intruding into the field	Measure the hood: clear width must exceed 2 tan θ times its depth
Overall veiling greyness, worst near one edge	Flare from a bright source just outside the frame	Fit or deepen the hood; check the inner face of the panel is still matt black
Sky white and blank on film as well as on paper	Exposed for the sky’s brightness rather than the land’s, or no filter fitted	Check the log’s filter field; on film a filter is the only remedy, on paper there is none
A soft bright patch in the middle of the frame	Fingerprint or dust on the filter, which sits inside the light cone	Clean at the edges, hold at the edges, sleeve it between exposures
Film negative flat and grey overall	Long-exposure contrast rise over-corrected, or development short	Compare against the pull you applied; Kodak’s 10/20/30 per cent table is the reference

Clean-up, storage and disposal considerations

Sleeve the filters flat and dry. Dry the negatives and sleeve them the same day, labelled to the log. Used fixer and the first change of wash water are silver-bearing and are collected; used film developer goes to its own container by the processing lesson’s route. Developer and stop at these dilutions carry no silver. Check your local regulations, which govern the route your containers eventually take and differ between authorities.

What to submit

• One landscape negative and working positive from either route, selected as the picture.
• The paired negative on the other material, as evidence rather than as a picture.
• The factor bracket, four sheets, with the measured factor and its uncertainty written on the sleeve.
• The log entries for every sheet, with the light recorded as a range with two timestamps.
• The folded prediction sheet, unfolded, with observations written beside predictions.
• A critique paragraph for the selected picture, in the form above.

Further work

The four-filter sequence. The same scene on film, bare and behind yellow, deep yellow and red, all within twenty minutes. Mounted in order, it is the clearest demonstration in the whole course of what spectral sensitivity means, because only one thing changed and you can see exactly what it did to each colour in the scene.

The polariser question. A polarising filter darkens a blue sky by a mechanism that has nothing to do with spectral sensitivity, and its factor of about 2.5 appears on all three Kodak sheets. It works on paper too, because it removes polarised light regardless of wavelength. Whether a polariser behind a pinhole behaves as it does behind a lens is not something this course has verified, and it is a good candidate for your own controlled test: same scene, same time, rotated through 90° between two sheets.

Your own sky ladder. Photograph a clear sky, a hazy sky and a cloudy sky on paper at the same exposure and keep the three sheets together. It is a reference set for how much density a sky contributes to a paper negative, and after this it will save you a bracket every time.