Safelight: Spectra, Distance, Duration and the Fog Limit

Ask most printers why their safelight is safe and the answer is a colour. Red, or amber, or “it came with the darkroom”. None of those is an answer, because a safelight is a quantity of light, not a colour of it, and the quantity that matters is the one arriving at the paper, multiplied by the time it arrives for.

ILFORD’s own definition says so without saying so: a safelight is the illumination that does not cause a significant visible change to a material during use. The clause at the end is the whole subject. Kodak put it more bluntly still — no safelight protects a sensitised material indefinitely. This page is about what determines how long “during use” is allowed to be, so that the next build can be designed to a number and the test after it can measure one.

What fog actually is, and where you see it first

Safelight fog is an increase in density above base plus fog, caused by illumination that was believed to be beneath the material’s notice. The definition is easy. What is not obvious is where it shows up, and getting that wrong is why bad safelights survive for years.

The intuition says fog appears in the white borders of the print, where no enlarger light fell. It does not — not first. Kodak state the sequence explicitly: excessive exposure to a correct safelight, or normal exposure to an incorrect one, degrades the highlights and lowers print contrast, and this happens before actual fogging is visible in areas that received no white-light exposure such as the borders. ILFORD describe the same failure from the printer’s side: safelighting can appear safe while causing low-level fogging that is seen not as fog but as a general loss of photographic quality, particularly reduced contrast and a lack of clear highlights.

So the first symptom of a failing safelight is not a grey border. It is a print that will not sing — milky whites, a maximum black that is not quite black, and a tonal scale that has lost its ends. The reason is on the characteristic curve. A small uniform exposure adds to whatever exposure a region of the paper already received. In the highlights, where the paper is sitting in the toe and the curve is shallow, a small addition lifts the density out of the toe and destroys the separation between the lightest tones. In the shadows, where the paper is already near its maximum, the same addition does almost nothing. The result is a compressed scale — which reads as dull, not as fogged.

📷What a printer notices, in order

First: highlights that are not clean, and prints that need more contrast than they used to. Then: a maximum black that has gone slightly grey. Only last, and often only on an unexposed offcut: a border that is no longer paper-white. By the time you see the third, the first has been costing you prints for months.

What the paper can actually see

A material’s spectral sensitivity is what makes a safelight possible at all. ILFORD state the base case plainly: all chloro-bromide black-and-white emulsions are blue sensitive with a slight sensitivity to green light, and to make an emulsion sensitive to anything else, sensitising dyes have to be added. That single sentence is the licence for every red and amber lamp ever sold.

Four material classes matter here, and they are not interchangeable.

Chloride and chlorobromide papers. Blue and near-ultraviolet, with the slight green tail ILFORD name. Slow, too: ILFORD describe their variable-contrast RC papers as roughly equivalent to a film of ISO 3 to 6.

Variable-contrast papers, which are a special case worth getting right. ILFORD’s own account is that MULTIGRADE paper is coated with a mixture of three separate emulsions, each of them a basic blue-sensitive emulsion carrying a different amount of green sensitising dye. All three have the same inherent contrast and the same speed to blue light; what differs is their speed to green. Expose with blue and all three respond together, which gives a narrow exposure range and high contrast. Expose with green and only the more heavily dyed components respond at first, which gives a much wider exposure range and low contrast. That is how the filters work.

And it is why green leakage from a safelight is worse than it looks. Green light on a variable-contrast paper does not merely fog it; it softens it, because a green exposure is by construction a low-contrast exposure. A lamp leaking green adds veiling density and pulls the grade down, and a printer chasing the second symptom by dialling in more magenta never finds the first.

Orthochromatic materials. Dye-sensitised into the green. ILFORD’s ORTHO Plus sheet allows either total darkness or an ILFORD 906 deep-red lamp with a 15 W bulb at not less than 1.2 m, and notes that under it reds appear much darker than normal — which is the same fact from the eye’s side.

Panchromatic film. Sensitised across the visible spectrum, and therefore at the top of Kodak’s list of materials that must be handled in total darkness. There is no lamp. A deep-red safelight is not a dim safelight for panchromatic film; it is a light leak.

What each material can see, as the manufacturers describe it

• Paper: blue + near-UV (340–500 nm) — ILFORD: all chloro-bromide emulsions are blue sensitive with a slight sensitivity to green. Slow, too - roughly a film ISO of 3 to 6
• Ortho adds green (500–590 nm) — the region a sensitising dye opens up; also where a variable-contrast paper's dyed components work, so leakage here lowers contrast as well as raising fog
• Panchromatic: the rest (590–760 nm) — sensitive across all three bands at once, so nothing is left for a lamp: Kodak put panchromatic film at the head of the total-darkness list

Drawn to teach, not measured. The band edges are the conventional colour boundaries and the material classes are placed from the manufacturers' own verbal descriptions - ILFORD's statement about chloro-bromide emulsions, the ORTHO Plus deep-red recommendation, and Kodak's total-darkness list. Nothing here is read off a published sensitivity curve. The coloured strip approximates where the visible spectrum falls and is a reading aid only; the wavelengths in the labels carry the information. The bands and curves are drawn to show the relationship, not measured.

What three emitters put into that axis

• Blue pump peak (440–470 nm) — a phosphor white LED is a blue die under a phosphor; the unconverted blue sits inside the paper band above, and red acetate in front of the lamp does not remove it
• Amber (575–615 nm) — on the boundary between the paper band and the ortho band: brighter to work by, less margin. Kodak are open that they recommend amber for paper because printers judge density better under it
• Deep red (635–675 nm) — clear of the paper band entirely, and narrow because the band comes from the semiconductor rather than from a filter. The course specifies no part: take your own emitter's dominant wavelength and half-width from its own datasheet

Drawn to teach, not measured. The emitter bands are drawn where those families of LED are conventionally found; no datasheet figure is quoted, because the course verified no specific deep-red part. Read this diagram against the one above: the two share an axis on purpose. The coloured strip approximates where the visible spectrum falls and is a reading aid only; the wavelengths in the labels carry the information. The bands and curves are drawn to show the relationship, not measured.

Safelight filters as a technology, and the trap in their names

A traditional safelight is three things, and Kodak name them: a housing that stops the bulb’s white light escaping anywhere but through the filter, the filter itself, and the bulb, whose wattage is part of the recommendation rather than a detail. ILFORD’s filters are one piece of glass coated with coloured gelatin, plus a piece of clear glass, bound up with a diffuser.

The filter families exist because material classes do. Both manufacturers publish a table and the tables map onto each other.

Material class	ILFORD filter	Kodak filter
Blue-sensitive papers, including variable-contrast papers	SL1 (orange), 902 (light brown)	OC (light amber)
Fast blue-sensitive materials such as line film	904 (dark brown)	1 (red)
Slow orthochromatic materials	915 (light red)	1A (light red)
Fast orthochromatic materials and recording materials	906 (dark red)	2 (dark red)
Very slow panchromatic materials	907 (dark green)	3 (dark green)
All panchromatic materials, colour papers — with extreme care	908 (very dark green)	10, 13 (dark amber, amber)

The name of the filter is not a permission slip for a different material. A 906 is a deep red for orthochromatic film and it is also a very conservative lamp for paper; a 908 is designed for the maximum possible efficiency with panchromatic material and ILFORD say in terms that it must be used with extreme care and that fast panchromatic materials must not be exposed to its direct light for any appreciable length of time. The label describes what the filter passes, not what your paper will tolerate.

Three further facts about filters change how you use them.

They fade. ILFORD say a filter used for several hours a day should be changed each year and the installation date recorded; Kodak put it at every three months for eight to twelve hours a day. A faded filter transmits more of the colours it was bought to absorb — which is the one direction of drift that cannot be noticed by looking.

Apparent colour is only a partial indication of transmission. That is Kodak’s phrase, and their warning is specific: coloured bulbs and improvised safelights may look right and still emit light, or other radiant energy, that fogs an emulsion.

The recommended filter is not always the safest one. Kodak are candid about this. A red filter often has less effect on paper than the amber OC they recommend for it, but most workers judge print density better under amber, so the recommendation trades a little protection for a working condition. Knowing that a manufacturer’s recommendation contains a judgement, not just a measurement, is part of reading it properly.

Why an LED is a better safelight, and how it can still be a bad one

A filtered tungsten lamp is a broad, hot source cut down by an absorber. A coloured LED emits a narrow band from the semiconductor itself. Nothing is being absorbed, so nothing has to be absorbed perfectly, and the failure mode of a fading filter does not exist. That is the whole case for building one.

Three ways to lose the advantage.

A phosphor-converted white LED behind a red filter. A “white” LED is a blue die with a phosphor on top that re-emits part of the blue as broadband yellow. Its spectral power distribution therefore contains a blue pump peak, and that peak sits squarely inside the paper’s strongest sensitivity. Putting red acetate in front of it is exactly the filtered-lamp arrangement again, with the added problem that the thing being filtered is a narrow, intense line rather than a smooth continuum: an acetate that looks opaque to blue by eye can pass enough of a peak to matter, and the eye is the worst possible judge of what is getting through, because the eye is what the filter was designed to satisfy.

An RGB strip set to red. There are blue and green dies on that strip. They are off, which is a software state, and a strip that is off is one power glitch, one controller reset and one wrong address from being on. Physical separation beats a setting.

A “red” lamp with no datasheet. Kodak’s warning about improvised safelights is fifty years old and still the right one. A red-painted bulb, a red gel, a bicycle lamp: each of them may look identical to a deep-red LED and emit something quite different, and there is no way to tell by looking.

🔬The datasheet figures that decide it, and the one this course could not source

Buying an emitter honestly means reading four numbers off its own sheet: dominant wavelength, spectral half-width (the width of the emission band at half its peak), radiant flux against forward current, and how the wavelength moves with current and with junction temperature. Cree’s XP-E2 family sheet is this course’s worked example of where those figures live — it prints a relative spectral power distribution from 380 to 780 nm and relative chromaticity charts against both current and junction temperature.

The course names no deep-red part. No specific deep-red emitter available in the United Kingdom was verified against its datasheet during the writing of this part, so no part number, dominant wavelength or half-width is printed here as fact. What is printed instead is the specification you hold a candidate against, which does not expire when a product line does.

Infrared, and the far side of the red edge

Beyond the deep red the paper stops responding altogether, which is why deep red is such efficient safelighting: almost every photon the lamp emits is a photon you can see by and the paper cannot use.

This is also the argument against the classic arrangement. A tungsten bulb behind a red filter emits the great majority of its power as heat and near-infrared, of which the eye uses none; the filter then throws away most of the visible; and what survives is a dim red light produced at appalling efficiency, in a hot housing, from a source that is still emitting a little of everything. A deep-red LED at a fraction of a watt can put the same illuminance on the tray with none of that.

Infrared is not entirely harmless, and Kodak name the specific case: plywood that appears opaque may admit infrared, which is a caution about darkroom construction rather than about the lamp. For infrared-sensitive material there is no safelight at all, and Kodak’s list puts high-speed infrared films in the total-darkness group beside panchromatic film.

The quantity: illuminance times time

Now the arithmetic that makes a working time a number.

Over the range where reciprocity holds, the photographic effect depends on the product of illuminance and time rather than on either separately.

H = E × t

Exposure at the paper

H is the exposure the paper receives from the safelight, E is the illuminance at the paper’s position in lux, and t is the time the paper spends there in seconds. Two consequences follow immediately, and they are the two levers you actually have.

Distance is a lever, and a strong one. For a source small compared with its distance, illuminance falls with the square of distance — the inverse-square law.

E(d) = E₀ × (d₀ ÷ d)²

Illuminance against distance

E₀ is the illuminance measured at a reference distance d₀, and E(d) is the illuminance at some other distance d. Since the permitted time is inversely proportional to the illuminance, the permitted time goes as the square of distance: double the distance and you buy four times the working time. Moving a lamp from 1.2 m to 1.7 m is the cheapest safelight upgrade in existence, and it costs a bracket.

Wattage is the other lever, and it is the one the manufacturers use. Kodak’s table asks for a 15 W frosted bulb for direct illumination and 25 W for indirect, and specifies 7½ W where the material is more sensitive. ILFORD’s general recommendation for their black-and-white papers is an SL1 or 902 filter with a 15 W bulb at not less than 1.2 m, safe for up to 4 minutes.

What inverse-square predicts for the working time, from one anchor point

• Predicted permitted time, ∝ distance²

Show the numbers behind this plot

A single rising curve on axes of distance in metres against permitted time in minutes. The curve passes through the anchor point at 1.2 metres and 4 minutes, which is ILFORD's published recommendation for their papers under an SL1 or 902 filter with a 15 watt bulb. From there it rises as the square of distance: about 2.8 minutes at 1.0 metre, 4 minutes at 1.2, 6.9 minutes at 1.6, 9 minutes at 1.8, 11.1 at 2.0 and 17.4 at 2.5. Markers sit at 1.2 metres and 4 minutes labelled as the published anchor, and at 2.4 metres and 16 minutes labelled as four times the time for twice the distance. The message of the plot is that the curve is arithmetic rather than measurement: it shows what the inverse-square law predicts if a single published recommendation is treated as a threshold, which it is not, and every point on it must be confirmed by the fog test before it is used.

Series	Distance from lamp to paper (m)	Predicted permitted time (minutes)
Predicted permitted time, ∝ distance²	0.80	1.78
Predicted permitted time, ∝ distance²	1.00	2.78
Predicted permitted time, ∝ distance²	1.20	4.00
Predicted permitted time, ∝ distance²	1.40	5.44
Predicted permitted time, ∝ distance²	1.60	6.94
Predicted permitted time, ∝ distance²	1.80	9.00
Predicted permitted time, ∝ distance²	2.00	11.11
Predicted permitted time, ∝ distance²	2.20	13.44
Predicted permitted time, ∝ distance²	2.40	16.00
Predicted permitted time, ∝ distance²	2.60	18.78
Predicted permitted time, ∝ distance²	2.80	21.78
Predicted permitted time, ∝ distance²	3.00	25.00

Exact arithmetic from the inverse-square law, anchored on ILFORD's recommendation of up to 4 minutes at not less than 1.2 m — but that recommendation is a manufacturer's working limit, not a measured fog threshold, so the curve is a prediction and not a permission. Its real use is the shape: the payoff for moving a lamp back is quadratic, which is why distance is the first thing to change when a test fails. 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.

Where the simple product stops being true

Three ways, and each of them makes the lamp less safe than the arithmetic says.

Pre-exposed paper is far more susceptible than unexposed paper. ILFORD state it directly: paper is more sensitive to safelight fogging after it has been exposed in the enlarger than before, which is why their test procedure has an “after” strip and a “before” strip and why they call the after strip the more critical one. The two effects even have names — latensification for light arriving after the printing exposure and hypersensitisation for light arriving before it.

The mechanism is the latent image itself. A grain that has already absorbed some photons is partway to a developable speck; a few more photons finish the job. A grain that has absorbed none has to be started from nothing, and the sub-threshold events it collects decay. So safelight photons landing on an already-exposed sheet are worth more, per photon, than the same photons landing on a fresh one.

Kodak call the combination a “super-additive exposure”, and their word for the consequence is cumulative: virtually all exposures add together and can change both density and contrast. The practical form of that is unforgiving — the safelight exposure your paper gets while you are composing on the easel, the safelight exposure it gets in the tray, and the enlarger exposure between them are not three independent events. They are one exposure with three sources.

✕The word that means two different things

Kodak’s “super-additive exposure” is not the superadditivity of Metol and hydroquinone in a developer. One is an exposure effect on the film or paper, the other is a chemical synergy between developing agents. The course keeps them apart and so should you.

Reciprocity does not hold at the very low illuminances a safelight produces. Everything above assumed the product E × t is what counts, and Part XIII’s treatment of reciprocity failure is the reason to be careful with that assumption at the bottom of the scale. This cuts in the material’s favour — very dim light for a long time is generally less effective than the product predicts — but it is a reason not to trust an extrapolation across a factor of ten in either variable, and to measure at the geometry you actually use.

How fog builds with safelight time, on fresh paper and on paper that has been printed

• Unexposed paper
• Paper already given a mid-grey printing exposure

Show the numbers behind this plot

Two rising curves on axes of safelight time in minutes against density above base plus fog. The lower curve, labelled unexposed paper, stays almost flat for the first four minutes and only begins to lift after about six, reaching about 0.05 at ten minutes. The upper curve, labelled paper that has already received a mid-grey printing exposure, lifts much earlier: it is already detectable at about two minutes and reaches about 0.05 at four minutes and about 0.18 at ten. A horizontal guide line is drawn at 0.04 density, labelled as the change ILFORD name as condemning a safelight when it appears after one minute. A vertical guide is drawn where the pre-exposed curve crosses that line, at about three and a half minutes, labelled first detectable change; a second vertical at half that time, about one and three quarter minutes, is labelled the safe time, being half the time to a detectable change under Kodak's definition. The whole point of the drawing is the gap between the two curves: a test run on unexposed paper would have reported a working time several times longer than the one the pre-exposed paper actually allows.

Series	Safelight exposure time (minutes)	Density above base plus fog
Unexposed paper	0.00	0.00
Unexposed paper	1.00	0.00
Unexposed paper	2.00	0.00
Unexposed paper	3.00	0.00
Unexposed paper	4.00	0.01
Unexposed paper	5.00	0.01
Unexposed paper	6.00	0.02
Unexposed paper	7.00	0.02
Unexposed paper	8.00	0.03
Unexposed paper	9.00	0.04
Unexposed paper	10.00	0.05
Paper already given a mid-grey printing exposure	0.00	0.00
Paper already given a mid-grey printing exposure	1.00	0.01
Paper already given a mid-grey printing exposure	2.00	0.02
Paper already given a mid-grey printing exposure	3.00	0.03
Paper already given a mid-grey printing exposure	4.00	0.05
Paper already given a mid-grey printing exposure	5.00	0.07
Paper already given a mid-grey printing exposure	6.00	0.10
Paper already given a mid-grey printing exposure	7.00	0.12
Paper already given a mid-grey printing exposure	8.00	0.14
Paper already given a mid-grey printing exposure	9.00	0.16
Paper already given a mid-grey printing exposure	10.00	0.18

Shapes drawn to teach, not measured from any paper. What is taken from published sources is the ordering — ILFORD state that paper is more sensitive after the printing exposure than before — and the two horizontal and vertical criteria: ILFORD's 0.04 density change, and Kodak's definition of a safe time as half the time at which a detectable change appears. Your own curve comes from the fog test. 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.

Geometry in a real room

Everything above treated the lamp as a point. It is not one, and the deviation is in the direction that flatters the lamp: a diffused source of appreciable size falls off more slowly than inverse-square close in, and reflections off a light-coloured ceiling put light back into the room that the law knows nothing about.

Kodak’s placement rules are the practical distillation, and they are worth having as numbers. No direct illumination closer than 1.2 m to the working surface. For general ceiling safelighting in a large room, no more than one lamp per 6 m² of ceiling, with lamps at least 2.5 m apart. And a rule that is about the eye rather than the paper: avoid pools of relatively bright light against a dark, unilluminated background, because they are difficult to work in and fatiguing.

Two arrangements, and the choice between them is not obvious.

Direct, a small pool over the developer tray, is efficient — you get the light where you look. It also puts the highest illuminance in the room exactly where paper sits longest, which is precisely why both manufacturers’ test procedures tell you to run the test at the tray.

Indirect, bounced off a white ceiling, spreads the light and lowers the peak. It costs more lamp output for the same working illuminance, which is why Kodak’s table asks for 25 W indirect where it asks for 15 W direct, and it depends on a ceiling that reflects — which is the reason the room page tells you not to paint everything black.

Which of the two gives lower fog for the same working illuminance is a question this page does not answer. It is decided by measurement, on the fog test page, on your ceiling.

The practice that follows from the numbers

Every one of these is a consequence of something above rather than a piece of darkroom etiquette.

• A minimum distance and a maximum working time, written on the lamp. Not remembered: written, with the paper and the date, because it is a property of a geometry and geometries move.
• Paper stays in its packet until it is needed. It is the material with the longest exposure in the room and it is exposed to nothing while it is in the box.
• The face-down habit. Kodak’s instruction is to handle paper emulsion side down, to insert it into the developer emulsion side down, and to turn it over only when experience says the image has come up. That single habit removes most of the tray-side exposure, which is the largest part of the total.
• The easel is not lit. Kodak say directly: do not put a direct safelight where it will shine on an enlarging easel. Paper on the easel is exposed paper, and exposed paper is the susceptible kind.
• Retest whenever anything changes. The paper, the lamp, the bulb, the filter’s age, the dimmer setting, the distance, the ceiling colour. ILFORD list nine factors that determine whether safelighting is effective and safe, and wall and ceiling colour and filter age are two of them.
• Half the development in the dark, if the test is marginal. It is one of Kodak’s own remedies, and it costs nothing but the pleasure of watching the image come up.

Everything else in the room that emits

The lamp you tested is rarely the only source, and Kodak’s own list of what fogs paper in a room with a correct safelight in it reads like an inventory of a modern darkroom: light escaping from an enlarger head, lighted dials on equipment controls, pinholes between the darkroom and a lit space, and construction materials that are not as opaque as they look.

To that list a darkroom built this decade adds: the timer’s display; the indicator LED on a power supply, a charger, a router or a smoke alarm; a phone, which is both a light source and the thing that undoes your dark adaptation; the gap under the door; and the glowing hands of a clock.

Every one of those is a safelight that nobody tested. The fog test on the next-but-one page treats them exactly that way: the enlarger lamphouse with the lamp on and the lens capped, the timer display, the indicator lamps, and finally the whole room with everything switched on at once, each measured against the same control.

Is there a standard?

There is a reference, and the course names it without quoting it. Kodak state that their safelight filter recommendations for specific materials are based on test procedures similar to those described in ANSI Standard PH2.22-1998. Under the course’s standards convention, that standard is cited by number as the thing the published procedures resemble; it is not purchased, not quoted, and no threshold is presented as coming from it. No ISO standard for safelight testing was found in this course’s corpus, and the honest position is that the two procedures this part uses are the manufacturers’ own, with their own criteria, applied to your own material.

Recap

A safelight is a quantity, not a colour: the illuminance at the paper multiplied by the time the paper spends under it. Its failure shows up first as flat highlights and a soft print, not as a grey border, because a small uniform exposure does most damage in the toe of the curve.

Paper is blue and near-ultraviolet sensitive with a slight green tail; variable-contrast paper turns that tail into a working green sensitivity spread across three components, so green leakage lowers contrast as well as raising fog. Orthochromatic material takes a deep red; panchromatic film takes darkness.

A coloured LED emits its band from the semiconductor and is therefore the better source — but a white LED behind red acetate still carries its blue pump peak, and an RGB strip carries dies that are merely switched off. Distance is the strongest lever you have, because permitted time goes as the square of it, and the manufacturers’ own numbers are 15 W at not less than 1.2 m, for up to about four minutes.

None of that survives contact with a real room without a test, for one reason above all: paper that has already been printed on fogs far more readily than fresh paper, so any honest test pre-exposes its sheet — which is exactly what the fog test does.

Check your understanding

Question 1. Your prints have become dull: milky highlights and a maximum black that is slightly grey. The white borders look perfectly clean. What is the first thing you should suspect, and why do the borders look fine?

• Exhausted developer; borders are unaffected by developer state
• Safelight fog. A small uniform exposure does most damage in the toe of the curve, where highlight separation lives, and both manufacturers state that highlights degrade and contrast falls before fogging becomes visible in areas that received no enlarger exposure
• The paper is out of date; borders fog last
• The enlarger lens is flaring, which raises the highlights only

Show the answer and why

Answer: Safelight fog. A small uniform exposure does most damage in the toe of the curve, where highlight separation lives, and both manufacturers state that highlights degrade and contrast falls before fogging becomes visible in areas that received no enlarger exposure

This is the failure mode that keeps bad safelights in service. Kodak state that excessive exposure to a correct safelight, or normal exposure to an incorrect one, degrades highlights and lowers print contrast before fogging is visible in the borders; ILFORD describe the same thing as a general loss of quality, reduced contrast and a lack of clear highlights. The curve explains why: in the toe the paper is barely responding, so a small added exposure lifts it disproportionately and destroys the separation between the lightest tones, while the borders, which received nothing at all, need much more before they show anything. Exhausted developer and flare are both real and both worth checking, but neither is the first suspect for this exact triad.

Question 2. A printer swaps a filtered tungsten safelight for a warm-white LED strip behind the same red filter. Predict what changes on the print, and give the mechanism.

• Nothing changes; the filter is the same, so the transmitted spectrum is the same
• Fog rises, because a phosphor-converted white LED carries a blue pump peak inside the paper's strongest sensitivity, and a filter chosen to satisfy the eye against a smooth tungsten continuum is not chosen to block a narrow intense line
• Fog falls, because an LED is more efficient and therefore dimmer
• Contrast rises, because LED light is bluer and blue exposure gives high contrast

Show the answer and why

Answer: Fog rises, because a phosphor-converted white LED carries a blue pump peak inside the paper's strongest sensitivity, and a filter chosen to satisfy the eye against a smooth tungsten continuum is not chosen to block a narrow intense line

A white LED is a blue die under a phosphor, so its spectral power distribution has a peak in the blue that the phosphor did not convert. That peak sits where a chloro-bromide paper is most sensitive. The old filter was chosen against a tungsten source whose blue content is small and smooth; a narrow, intense line is a different thing to block, and the eye cannot report on it because the eye is exactly what the filter was designed to satisfy. Option three confuses efficiency with output — an efficient lamp at the same visual brightness emits more of whatever it emits outside the visible response, not less. Option four gets one true fact, that blue exposure gives high contrast on VC paper, and applies it to a uniform fog exposure, which lowers contrast rather than raising it.

Question 3. A lamp gives a measured safe working time of 45 s at 1.0 m. Estimate the permitted time at 1.8 m, and say why the estimate must still be tested.

• 81 s, and it must be tested because the lamp may have drifted
• About 146 s, from the square of the distance ratio — and it must be tested because a diffused lamp is not a point source at that range, because the ceiling returns light the law does not model, and because the safe time depends on the paper as well as on the lamp
• 81 s, because illuminance falls linearly with distance
• About 146 s, and no further test is needed because the law is exact

Show the answer and why

Answer: About 146 s, from the square of the distance ratio — and it must be tested because a diffused lamp is not a point source at that range, because the ceiling returns light the law does not model, and because the safe time depends on the paper as well as on the lamp

Permitted time is inversely proportional to illuminance and illuminance falls as the inverse square of distance, so time scales as the square of the distance ratio: 45 × (1.8 ÷ 1.0)² = 45 × 3.24 ≈ 146 s. Options one and three make the common error of scaling linearly. Option four is the more interesting mistake: the arithmetic is exact but its premises are not. A diffuser of appreciable size falls off more slowly than inverse-square close in, a light-coloured ceiling adds an indirect component the law does not describe, and the number you are scaling was measured on one paper — change the paper and the whole curve moves.

Question 4. Why does every honest safelight test give the paper a printing exposure before the safelight exposure?

• To make the fog easier to see against a grey background
• Because paper carrying a latent image fogs far more readily than fresh paper — the effect ILFORD name as latensification, and Kodak describe as a super-additive, cumulative exposure — so a test on unexposed paper reports a working time the printer will never actually have
• Because the developer needs something to work on
• Because it standardises the development time

Show the answer and why

Answer: Because paper carrying a latent image fogs far more readily than fresh paper — the effect ILFORD name as latensification, and Kodak describe as a super-additive, cumulative exposure — so a test on unexposed paper reports a working time the printer will never actually have

ILFORD state it plainly: paper is more sensitive to safelight fogging after it has been exposed in the enlarger than before, and their procedure treats the "after" strip as the more critical one. Physically, a grain that has already collected some photons is partway to a developable speck and needs few more to finish; a fresh grain has to be started from nothing and its sub-threshold events decay. Kodak call the combination a super-additive exposure and note that virtually all exposures are cumulative. Option one has a grain of truth — a mid-grey does make small changes easier to judge — but it is the consequence, not the reason. The reason is that the printer never handles unexposed paper at the tray.

Question 5. You are given the spectral data for three lamps and a variable-contrast paper, and must choose. Lamp A is a deep-red LED. Lamp B is an amber LED. Lamp C is an RGB strip driven red. What is the correct choice and what is the specific objection to each of the others?

• A. B sits nearer the paper's green tail so it buys less margin, and C carries blue and green dies that are off only as a matter of software
• B, because amber is what the manufacturers recommend for paper, and both A and C are unnecessary
• C, because an RGB strip can be tuned to any wavelength you like
• Any of them, provided the fog test passes

Show the answer and why

Answer: A. B sits nearer the paper's green tail so it buys less margin, and C carries blue and green dies that are off only as a matter of software

Deep red is furthest from a paper whose sensitivity is blue, near-ultraviolet and a slight green tail, so it buys the most margin per unit of visible brightness. Amber is brighter to work by and is what Kodak recommend for paper — they are open that this trades a little protection for a better working condition — but it sits closer to the tail, and on a variable-contrast paper the green end of that tail lowers contrast as well as raising fog. The RGB strip is the one with a mechanism rather than a margin problem: the blue and green dies exist, and "off" is a controller state one reset away from being wrong. Option four sounds like this course and is not: the fog test tells you the working time you have, not whether you chose a lamp whose failure modes you can live with.

Question 6. ILFORD recommend an SL1 or 902 filter with a 15 W bulb at not less than 1.2 m, safe for up to 4 minutes. Your session has paper out for eleven minutes. What follows?

• Nothing; the recommendation is conservative and eleven minutes will be fine
• The recommendation does not cover your session, so either the geometry changes — most cheaply by moving the lamp back, since permitted time goes as the square of distance — or the handling changes, or you test and find your own number
• Buy a dimmer bulb and the four minutes becomes eleven
• Split the session into three parts of under four minutes each

Show the answer and why

Answer: The recommendation does not cover your session, so either the geometry changes — most cheaply by moving the lamp back, since permitted time goes as the square of distance — or the handling changes, or you test and find your own number

A published recommendation is a working limit for a stated geometry and material, and it is not a budget you can overspend. Three honest responses exist. Move the lamp: going from 1.2 m to 2.0 m multiplies the predicted permitted time by (2.0 ÷ 1.2)² ≈ 2.8. Reduce the exposure rather than the illuminance — paper in the packet, emulsion side down, the developing dish out of the direct pool, half the development in the dark. Or measure your own safe time, which is what the fog test exists for. Option four is the seductive one and is wrong for a reason worth remembering: Kodak state that virtually all exposures are cumulative, so three four-minute exposures are one twelve-minute exposure as far as the paper is concerned.

Check my answersClear

Alternative route: without a meter, a densitometer or a darkroom

Nothing on this page needs a facility. What the next two pages need is a way of measuring illuminance, and a reader who has neither a lux meter nor Part XV’s photodiode head is not shut out.

Without a lux meter, you can still get everything this page is for, because the fog test measures the thing that matters — time — and not the thing that causes it. The lux reading is a convenience that lets you predict what a change of distance will do; the test tells you what it actually did. Work in ratios instead: run the test at one distance, move the lamp, run it again.

A photopic lux meter is in any case a poor instrument for deep red, because lux is weighted by the eye’s response and the eye’s response is collapsing where a deep-red LED emits. Part XIV’s calibration page sets out why, and the practical consequence here is that a low lux reading from a deep-red lamp is not evidence of a low photographic effect. Where you have the photodiode head, use it, and read a relative number rather than a photometric one.

Without a darkroom at all, the whole of this page is still the argument you need for choosing a bought lamp and for deciding where to stand it in whatever space you have. The measurement can wait; the reasoning cannot, because it is what stops you buying an RGB strip.