SEI photometer

The SEI photometer is a comparison photometer. In other words, the brightness of the subject is assessed by comparing it with a spot of known brightness. This in turn is derived from a lamp of standardized brightness in conjunction with graduated filters: the brightness of the spot is varied by moving the the filters until it matches the brightness of the subject.

Because the human eye is an incredibly good comparator, the SEI photometer is inherently extremely accurate. Principally for this reason, it has retained a cult following to this day, especially (it is said) in Hollywood. Another reason is its enormous measuring range of a million to one, in three 'centuries' of 100:1, each of which is 100x more (or less) sensitive than the next. A third reason is the fact that it can be focused very close indeed, down to about 6 inches (15cm) from the front of the meter. Yet a fourth reason, it must be said, is that it is a sublime example of a beautifully made optical-mechanical device from the golden age of fine engineering: although just 7 inches (18cm) tall and about about 1-5/8 inches (4cm) in diameter, it weighs an impressive pound and a quarter (about 575 g).

It was made by Salford Electrical Instruments in the UK from about 1948 to the 1960s, and distributed in the UK by Ilford Ltd. and in the United States by The Zoomar Corporation. It was based on the Dunn and Plant Photometric Exposure Meter, first described to the Royal Photographic Society in the Photographic Journal 85B, 114-119, November-December 1945.

This in turn was based on a prototype built by Arthur Dalladay, editor of The British Journal of Photography, in about 1935; he describes it in a paper in the 1937 BJP Almanac on pages 127 to 138, under the heading 'Solving the Abnormal Exposure Problem'. Dalladay's meter was based in turn on the Bunsen grease-spot photometer, invented by Robert Bunsen in 1859. Bunsen (1811-1899) was Professor of Chemistry at Heidelberg University and is best known for the Bunsen burner though he was also responsible for the magnesium light and (with the British chemist Henry Roscoe, later Sir Henry) for the Bunsen-Roscoe law of reciprocity.

Only three comparison photometers appear ever to have been built for specifically photographic applications, the others being the Ainger Hall/Turl meter which also appeared in the late 1940s (we have never seen one) and the Papri, a Russian meter which used an orange LED and half a dozen unusually-sized Russian mercury cells: we bought one of these in the Soviet Union in 1990 but it was subsequently loaned out and not returned. Unfortunately, we never got inside it to see how it worked but we can confirm that the (unevenly illuminated) orange spot made it hard to match brightness.


SEI Exposure Photometer

The little meter on top is nothing to do with the exposure reading; or to be more accurate, it is used only for calibrating the bulb that is used for making the reading. The finger-wheel under the front of the telescope moves single-density neutral density filters in and out of the light path to give the impressive 1,000,000:1 reading range; the bottom of the meter, up to the third chrome ring (the one with the tapered top and the lines) rotates to vary the brightness of the comparison spot via sliding, graduated neutral density filters. The black button on the bottom turns the bulb on and off. This is the meter as we bought it, unmodified and (to be frank) working so poorly as to be unusable.


The principle of the meter is quite adequately explained by reference to the schematic diagram, reproduced from the original instruction book by kind permission of Ilford, though at first it it does take quite a bit of concentration to understand it. The Ainger Hall/Turl meter used a rather different principle, moving the light source up and down on a cam: the brightness was therefore varied by means of the inverse square law.

The light from the bulb (which is clearly drawn about half way up the meter, but not labelled) passes through the collecting lenses F and the neutral density wedges G before being reflected off the semi-silvered spot C in the middle of the optical block (two cemented 45 degree prisms -- the Ainger Hall/Turl meter used a pellicle with a mirror spot). The eye (on the right) can compare the brightness of the spot with the brightness of the subject as seen through the simple telescope made up of the lenses B. Both the eyepiece and the objective can be slid in and out along the optical axis, for focusing and to accommodate individual eyesight. The image is slightly magnified (about 1.3x) and is disconcertingly upside-down.

By way of further refinement, as well as the variable wedges G there are additional filters D and E.

D places one of three filters in the light path, 0x (clear), 100x (ND 2.0) and 10,000x (ND 4.0). In conjunction with the variable wedges which cover a range of 0 to 100:1 (ND 2.0 again) this gives the first century as 100:1, the second century as 10,000:1, and the third century as 1,000,000:1 (rarely much use except for highlight readings in bright sun). These filters are controlled by the black finger-wheel that you can see under the objective in the main picture. Depending on which of the three filters is in use, the exposure reading is taken from one of the three scales visible on both the unmodified meter (above) and the modified meter (below).

E places one of two filters in the light path, a weak neutral density (for matching the yellowish light of the unfiltered bulb to artificial light) and a blue (for filtering the bulb to match daylight better). This is controlled by a second black thumb-wheel under the eyepiece.

The switch M turns the bulb on and off -- that's the black button in the main picture -- and the variable ND wedges are driven by rotating the bottom of the meter, which rotates as a unit up to and including the third chrome ring (the one with the lines on it). Power is supplied by a 1.5V battery which has to be 'skinned' (have the outer insulating cover stripped off) in order to fit into the battery case.

In the LED version, this is all that matters, but in the incandescent bulb version there is also a rheostat N (at the bottom, operated by a sort of key in the bottom of the meter) to vary the brightness of the bulb. This in turn is read by an annular photo-electric selenium cell H which drives the microammeter A. The brightness of the bulb is set to a constant point on the microammeter.

The Ainger Hall/Turl meter again diverged here. There was no meter cell, microammeter or rheostat, but there was a single variable ND filter which allowed the meter to be calibrated to a candle flame. This is far more accurate than most people expect -- the light of a candle-flame is remarkably constant -- but it did mean that for acceptable reliability the meter had to be calibrated fairly often, preferably before each shooting session.

The bulb on the SEI is severely underrun and has a very long life, but eventually it becomes too murky to give out enough light to operate the meter and has to be replaced.

This is where life gets interesting. The bulb is frosted; has a black cap; and is soldered into a mount. As long as the meter was current, none of this mattered very much: you could simply buy a replacement unit. Today, however, you have to file away most of the metal; melt out the rest; then frost and paint a suitable bulb and solder it back in place. None of this is especially difficult but it is very time consuming.

Also, of course, you are dealing with a piece of optical-mechanical-electrical engineering that may be very nearly 60 years old, and other parts do fail. The meter cell may lose its sensitivity; the microammeter may cease to be wholly reliable; the rheostat can get dirty; and (as happened with ours) the 'finger' on the rheostat may break, and it is very hard to solder, though Roger succeeded.


Enter Huw Finney. He makes a small integrated circuit, the size of a large coin, with a white (and filtered) LED incorporated. You take out the old battery holder and lamp; drop in the integrated circuit; and insert the same U2 (C-cell) battery as before, but without the outer cover stripped off and without the battery holder. Because the LED is pre-calibrated and will give a constant light output for years, without aging and (within reason) even with a battery running at slightly reduced voltage, the rheostat, meter cell and meter cease to be relevant.

Better still, send the meter to Huw and he will give it a general overhaul -- he even recemented a slightly separated optical cube on ours -- and if you want he will also replace the original British Standard Logarithmic (BS log) scales with ISO, though he actually labels them ASA for the sake of a 'vintage' look.

It's not a cheap conversion -- it will cost a lot more than the non-working meter -- but after he has done it, you will have a 1950s legend brought up to 21st century technology, and far more reliable.



Amazingly, in the early 21st century this 55-year-old design was given a new lease of life when the incandescent bulb (no longer available) was replaced by a white light-emitting diode by Huw Finney: The only visible difference from the outside is the replacement of the old scales with new (and much more convenient) ISO arithmetic scales instead of the old B.S. (British Standard) log scales.

The meter on top of the LED version is redundant, as it was used for setting the brightness of the bulb, i.e. it is not an exposure read-out in any conventional sense. Huw has removed the photoelectric meter from some of his modified SEIs, but we preferred to have ours left as it was for the sake of appearances.

  < P>

Finney LED module

The LED module simply replaces the bulb of the original meter, and the C-cell is dropped in behind it.


using the SEI

First, you need to set the film speed on the bottom-most ring of all, the yellow/pink one in the above picture. On the unmodified version British Standard (BS) Log speeds are set. These are the same as ASA Log but as most people are familiar only with ASA Arithmetic this is of limited usefulness. To confuse matters further, almost all SEI meters (maybe all) were built using the old BS 1380 fractional gradient criterion instead of the newer fixed density criterion (0.10 above film base plus fog) -- and the safety factor (the inbuilt bias towards over-exposure of negative films) may have varied as well. In other words, the film speeds are only a rough guide and may need to be modified in the light of experience. The film speed comparison chart on the right is reproduced from the original handbook, again with kind permission of Ilford.

In fact, this table does not tell the whole story, as ASA and Weston are generally agreed to differ by 1/3 stop, so Weston 8 is ASA 12 and so forth. In practice, if you want to use an unmodified SEI, you will need to establish personal correction factors for ISO speeds, though these factors should be constant once they are established.

On the modified version the lowest setting is approximately ISO 2 (the first marked figure is 3, but there are two 1/3 stop divisions below this) and the highest is ISO 10,000 (two graduations beyond 6400) -- though as already noted, the scales are labelled ASA not ISO.

In either case (modified or unmodified), you need to use different indices according to whether you are shooting negative film (colour or black and white) or slide films, digital or movies. For negatives (colour or black and white) set the BLACK index mark (an inlaid black dot and line on the first knurled/graduated ring) against the film speed. For slides, digital or movies, set the film speed against the WHITE index mark (inlaid white dot and by now rather faded white line on the same ring). Depending on which index you use, your exposures will vary by exactly 100:1, for very good reasons which will become clear in a moment. There is no 'mid-tone' or 'grey card' index, because the original designers knew that such indices are useless.

Next, but only with the standard version, check the lamp calibration. Press in the 'on' button and make sure that the needle in the meter on top (far left) lines up with the red 'Standard Brightness' mark. If it does not, use the rheostat on the bottom (left) to bring it into line. According to the original instruction book there is no need to take excessive care as 1/10 inch (2.5 mm) deviation corresponds only to 1/3 stop. With the LED version this step is omitted as it is unnecessary.

choose the area to meter

For negatives (film speed on black index), set the metering spot over the darkest area in which you want shadow and detail. This keys your exposures to the shadows, the very best way to ensure adequate shadow detail. This is, after all, the way ISO negative film speeds are defined, in terms of shadow detail.

For slides and movies and digital photography (film speed on white index), set the metering spot over the lightest area that you don't want to 'blow' to a featureless white -- NOT a specular highlight. This is why there are two indices, and why there is no 'grey card' index: just shadow and highlight. There is rarely much reason for highlight metering, as an incident light reading will give exactly the same result, much faster and more easily. The only reason we can see to meter this way is if you can't get close enough to take an incident light reading.

There is another way to use the meter, to which we shall return later under the heading 'keytone reading', but the simple rule is shadow readings for negatives (mono or colour) and highlight readings for slides, digital and movies.

Well, Folklore Museum, Gozo

The darkest area in which Frances wanted texture and detail was the shadowed area between the well and the wall, pretty much bang in the centre of the picture -- so that would be the area to meter. You can always use the true ISO with a spot meter, assuming the meter to be accurate and the film to be developed to ISO conditions, but reducing development time or using fine-grain developers will result in lower film speeds.

take your reading

Now you can take your reading. Align the spot over the area to be read. Press the 'on' button. Twist the bottom of the meter until the brightness of the spot matches the brightness of the area to be read. If the spot cannot be made to match, switch the ND filter (the finger-wheel under the objective) to a lighter or darker filter. If the colour of the spot is too obtrusive, try the other colour, using the finger wheel under the eyepiece.

Unlucky Pfennig

The three pictures above give you a rough idea of what you would see through the SEI. On the left, the spot is too bright -- brighter even than the reflected highlights; in the centre, it is a pretty good match; and on the right, it is too dark. The metering spot is slightly oval, much as it is represented here.

In practice, you wouldn't normally read a tone like this but it is much easier to represent the spot in a mid-tone than it is in the highlight of a real picture. Roger shot this on 4x5 inch Fujichrome with a Linhof Technikardan and (as far as he recalls) a 210/5.6 Rodenstock Apo-Sironar-N.

Take your actual reading from the aperture scale (the second scale up from the bottom of the meter, running from approximately f/0.8, or 2/3 stops faster than f/1, to f/100, 2/3 stops beyond f/64) and whichever of the three meter scales corresponds to the ND filter in use. The spots on the finger wheel are green (maximum ND), white (medium ND) and red/pink, corresponding to no ND. These are (weakly) colour-keyed on the original, and rather more brightly color-keyed but with a poor colour match on the Finney version. Either way, left is maximum ND, centre is middle ND. and right is zero ND, with the scales held the right way up to read (base to the right, metering head to the left). Obviously you can read either 'shutter priority' (choose your shutter speed and follow the line to the right to find the appropriate aperture) or 'aperture priority' (choose your aperture and follow the line to the left to find the appropriate shutter speed).

The reading on the left, at f/11+2/3, is 1/125; or 1+1/3 seconds; or 1+1/3 minutes, depending on which 'century' you are using. Note the black index dot (negatives -- shadow readings) against ISO 100 (marked as 1: ISO speeds beyond 75 are marked as 'x100'). The reading on the right, at 1 second (middle range) calls for f/1.4+1/2. Note the white index dot (slides, digital, movies -- highlight readings) against ISO 100.

the keytone method

Instead of reading either a shadow or a highlight, you can instead choose a constant 'keytone'. This is expecially handy in cinematography where you want (for example) the star's skin always to be represented as the same tone. The instruction book once again comes to our rescue, as seen on the right, but we have added the corrections in stops, which suit the ASA/ISO arithmetic system and will be more familiar to most people.

It may seem a little odd to modern eyes to vary the film speed like this, but in truth, it is the quickest and easiest way of doing it. J.F. Dunn, in his 'Exposure Meters and Practical Exposure Control' (The Fountain Press, London 1952), devotes an entire chapter to spot meters and adds 'Unshaded average grass and red-brick walls, +11 degrees' which translates to 3+2/3 stops.

In other words, an ISO 100 reversal film would be re-rated at EI 125 (+1/3 stop), EI 160 (+2/3), EI 200 (+1) and so forth, right up to 500 ('bronzed face tone') or even 1250 ('Unshaded average grass and red-brick walls') -- though we have to confess ourselves puzzled by the footnote about sunsets with reversal colour films.


extending the range

By means of additional ND filters over the objective, the range can be extended indefinitely, e.g. for reading the exposure inside a furnace. There is also the possibility of using a 10x ND (ND3)  for normal readings, which often allows the use of the brightest (unfiltered) scale instead of the middle scale (100x, ND6). This is convenient because it gives you a brighter telescope image.

Tibetan Freedom Fighter

Something that is not immediately obvious is that the key-tone approach is as applicable to black and white as to colour. Here, a 'bronzed face tone (diffuse highlight)' would probably be appropriate for a reading off the cheek highlight, so for a series of portraits you could set up the lighting as you wished and keep to a constant tone, regardless of background, clothing or anything else. Given the general nature of Tibetan complexions, you might however want to give more or less exposure according to skin tone. Or of course you could just use an incident light reading... Roger used a Leica MP and 90/2 Summicron for this, shooting on Ilford XP2 Super.

the bottom line

Does the SEI deserve its awesome reputation? And is it -- especially in its LED form -- a usable meter today? The answer to both questions, as so often, is a firm, unequivocal "It depends."

Certainly, as the first ever top-flight professional spot meter, an unmodified SEI is an incredible bit of kit, and as an example of the utmost quality in electro-mechanical-optical design, it is up there with the best. For a collector, a plain SEI is indeed a prize worth having.

The story with usability is a bit more complicated. Although it is possible to use an unmodified meter, we would suggest that the LED conversion is all but essential if you really want to use the SEI. Then, even with the modification, it is distinctly slow and cumbersome next to just about any other spot meter on the market. The upside-down image is also, as already noted, distinctly disconcerting, and with the first ND in place (which is the normall setting for outdoor photography, shadow readings) the image is somewhat murky.

On the other hand, it is one of the most precise meters ever made, and the half-degree spot can even be used (if your hands are steady enough) to meter areas that are smaller than half a degree: the spot has only to overlap with the subject, after all.

In the studio, where you are in full control of the lighting, the SEI is arguably the most convenient of all meters if you need to match tones, as you simply meter the first lighting set-up and then, in subsequent shots, adjust the lights to match the spot rather than adjusting the spot to match the lighting.

Perhaps the strongest reason for actually using an SEI is because it is such a beautifully made piece of kit. If you use top-flight cameras -- Gandolfi, say, or Alpa -- then the SEI is the only meter that even begins to have the same feel of quality. Also, there is no question, it becomes a lot faster to use when you have had a reasonable amount of practice in using it.

If you want to eat your cake and have it, perhaps the best bet is an externally unmodified meter (with the original scales) and the LED conversion as well: the conversion is, after all, completely reversible, so you will have both a meter to use and a collectors' piece. The only difficulty lies in establishing personal conversion factors for ISO to BS Log, but this should only take a roll or so of each type of film (negative and colour slide). Once you have a conversion factor for each type you should be able to make direct conversions for other types: double the ISO, add 3 degrees to the BS Log, etc. One degree is 1/3 stop, which is all you need to know.

We'll be honest, though. We don't use ours anything like as much as we expected. We have three other spot meters (Pentax analogue and digital, and Gossen Spot Master 2) and the two Pentaxes see most of the use, with the Gossen reserved mostly for spot flash metering. But for quiet, reflective photography, especially with a 5x7 inch camera, the SEI is the meter we automatically think of.

Go to the unillustrated list of modules (in either alphabetical or date order)

or go to the illustrated list of modules

or go to the home page

last updated: 15/02/05

© 2005 Roger W. Hicks