The quest for a reliable spotmeter

[Stefan.Dalibor]

I'm shopping for a spotmeter (a real 1 degree meter, useable down to EV1 - so please, no suggestions for Sekonics [345]* or Gossen Starlite).

So far, I have tested the Gossen Ultraspot 2 and the Minolta Spotmeter F, but am not satisfied: Both are rather sensitive for infrared light (hot stove element +3EV, IR diodes like in TV remote controls or IRDA computer interfaces +5-6EV!); the Ultraspot is non-linear below EV 2.5 (checked with ND filters), while the Minolta (though being linear within 0.3EV over the complete sensivity range down to EV1) has internal flare problems (i.e. is sensitive to light sources quite a bit outside the marked 1 degree circle in the viewfinder).

Now it's clear to me that it's probably not a beginners exercise in opto-electronics to construct and produce a precise spotmeter for photographic purposes, but given the amount of money these devices cost and the experience that the manufacturers have, I had expected them to perform far better!

Is it really necessary to buy a Zone IV modified meter in order to get something that delivers reproducible measurements instead of approximations and reason for guesswork (what amount of IR radiation might this subject contain, am I measuring far enough from distinct bright spots, can I rely on the meters results if the light level is this low)?

What are the downsides of the Zone IV modified meters? I suppose you have to pay for better filtering and baffling by decreased sensivity... down to which EV value are they working reliably, and are they useable for slide film? Or is the spectral sensivity matched to BW film in a way that makes them less useful than an unmodified Pentax meter for shooting slides?

Has anybody checked the Seconics 778 regarding IR sensivity, baffling, linearity and spectral sensivity?

[Carl Weese]

I'm curious--how are you testing to show that the reading of the stove element is too high? What is the basis for saying it is +3 EV?

On a second note, I've found that while the Minolta Spotmeter F is a useful tool (it's my primary meter) it is distinctly non-linear in the top end of its range, overestimating the brightness of, say, a white object in full sunlight.

[Walter Glover]

Stefan,

I am surprised at your report of the Minolta Spotmeter F. I had one until recently (for the odd flash assignment) and checked it for linearity, flare, etc. against my Zone VI modeified Pentax Digital and it kept pace really well.

I use the Zone VI Pentax for both colour (reversal & negative) and Black & White. Its only foible - determined by testing and checking density with a MacBeth Densitometer - is that when reading through a 25a (Red) filter I have to give 1 more stop exposure.

I periodically shoot in very dim areas (black-painted rear projection booths for home theatres is an example) and I find that the low light accuracy is totally trustable.

My linearity test uses NDs to 3.0 (10 stops in a given light level) and the error is one LED blob ( - 1/3 stop )

Walter Glover

[bglick]

Stefan, quite impressed by your thorough research. I am curious also about how you are checking these meters for IR sensitivity? Please explain how you are doing this. Also, where do you see this being an issue in outdoor shooting using color film?

As for spot meters... I have tested several Sekonic spot meters and I have read the test results a few years back on the test results of other makers...and unfortuantely I do not have better news for you... the meters are never linear througout their entire working range. This can easily be tested by putting the the spot meter in the back of a view camera and stopping down one f stop at a time, then re check redings. Your ND test is also useful.

The compromise I have made is to use spot meters in the mid range of their working range where they are reasonably accurate and linear. However, when very low light levels are encountered, I rely on an ambient meter well suited up to -8 EV. It is made by Quantum. The good news is, once you are confident in your testing methods, you can record how far the meter is "off" at different EV levels and tape a cheat sheet to the meter! Just adjust in the field.

The other thing you need to be aware of.... not all spot meters are calibrated to the same reflectivity value. Some are set at 13% such as Sekonics while other are at 18%. So don't expect apples to apples readings between different makers of spot meters. Its not always easy tracking down this information. Also, not all grey cards are 18%. Hard to beleive there is so many issues with light meters, huh.... I agree...

[Lars Åke Vinberg]

Interesting... I'm sitting here with a Sekonic 778, always assumed that it was both accurate and linear.

Stefan, do you think that you could write a step-by-step testing instruction? I'd be happy to go through the steps and post the results.

[Arne Croell]

Stefan,

I both have a regular Pentax V and a Zone VI modified Pentax digital. The latter is my regular meter. Years ago I had the chance to try out a new Gossen Ultraspot/Spotmaster for a day and make comparison tests with the Zone VI. Both meters agreed quite well with each other (within 1/3rd EV), except when measuring through a dark red filter (29) where the Gossen was 1-2EV too high. I also made an IR diode test with a TV remote control in a dark room (pointing the remote control directly into the meter). _Both_ meters showed a response, but the Gossen was several EV higher, i.e. more IR sensitive. That could be cured by using an IR filter (actually it was a combined UV/IR filter), the type that is sold for digital video cameras. With that filter on the meters they agreed quite well over the whole range - the filter took out 1/3rd EV over the range but that was the only side effect. As for flare, the modified Zone VI meter is better than the unmodified Pentax V (comparing apples and oranges here, but I never had an unmodified Pentax digital). However, it can still be improved by using a lens hood (Heliopan makes a cylindrical metal hood that fits the Pentax digital perfectly). So using a UV/IR filter and a lens hood on an umodified meter might get you close.

[Pete Andrews]

IR sensitivity is a BIG problem with most modern meters that use a Silicon cell, not just spotmeters. The manufacturers don't pay nearly enough attention to it.I bought an Elicar spotmeter that had exactly the same problem of excess IR sensitivity. Despite having a cell with a supposed IR filter over it, I could still get an appreciable reading off an IR LED, and could see a very noticeable (>2EV) difference in readings from various black coated surfaces that looked visually almost identical. (Epson inkjet ink appears to have a huge reflectivity in the near-infrared, by the way.)

In the end, I replaced the sensor cell with one that looked physically identical, but that showed 1/5th of the response to an IR LED than the one originally fitted to the meter. The response to visible light was practically the same for both cells.Whether this is a typical spread of IR response, or whether more recent manufacturing techniques or a different make of sensor was the reason for the difference, I don't know.Anyway, it was a worthwhile mod, IMHO, although recalibrating the meter afterwards wasn't easy.The sensor I used was an EG&G Vactec type VTB8440B, which only cost #2 UK, about the equivalent of $5 US.

If money is less of an object; I notice that Edmund Optics Ltd have a number of filters available for cutting IR. They sell B+W IR-cutoff filters in standard threaded mounts, in sizes from 46 to 58mm diameter, but the cost is upwards of $150 US.They also sell 1" and 2" square visible bandpass filters, as well as IR cutoff filters, hot mirrors, and 'tech spec' heat absorbing glass. These are all much cheaper, in the range of $30 to $60, and any of them would be effective at reducing the IR sensitivity of a spotmeter. They could be fitted either internally, or in front of the lens. However, they'd all require some compensation or recalibration of the meter, but only by about 10 to 20%.

[Andrey Vorobyov]

"What are the downsides of the Zone IV modified meters?"

-- the only downside I can tell is that if the brightness is below EV 2, the circle in the viewfinder is almost invisible.

"I suppose you have to pay for better filtering and baffling by decreased sensivity... "

-- No. I had to pay only in terms of money.

"down to which EV value are they working reliably, and are they useable for slide film?"

-- Down to 0 EV. In the letter supplied with the spotmeter (letter by person who makes the modification) there is a statement that the meter has been tested for linearity in the range 4-19 EV. I don't know what tolerances he used in the testing, but the meter seems linear even below 4 EV. (If you will test it with a film, dont forget about Schwarzshield effect)

"Or is the spectral sensivity matched to BW film in a way that makes them less useful than an unmodified Pentax meter for shooting slides?"

-- I heard the spectral sensitivity of the meter is made close to s.sensitivity of Tri-X. Since different films (even B/W films) have different spectral sensitivity, an orientation to a specific emulsion can make the work with other emulsion less preise, it is obvious. But I think it is still more precise than using the meter that is too sensitive to IR, UV light or just to saturated colors, in spite of that strange virtue that the latter meters are not claimed to be oriented to a particular emulsion...

Ok, it is a dream to have a meter with a customizable spectral sensitivity, but I'm sure the cost will be prohibitive.

Like other posters I hope you will share your methids of testing for linearity and IR-sensitivity

Cheers Andrey

[Stefan.Dalibor]

O.K., here are my &raquo;spotmeter tests&laquo; - I don't claim them to be of scientific quality, but think they do have real-world relevance... after ruining 8 Velvia Quickloads (hopelessly under-exposed) and a probably unique opportunity because I trusted the Ultraspot's results in a brickworks (with a rather high level of IR radiation due to large heaps of red-glowing bricks), I repeated Walter's tests (see ) with the Ultraspot, and switched it for a Minolta Spotmeter F after recalibration by Gossen didn't improve the meter's deficiencies. For all experiments, I put the meter on a tripod and repeated them in in daylight and artificial light. <ul> <li>Low-light linearity - you need ND filters (I used two Hoya NDx4 and a no-name NDx2): <ol> <li>Measure your grey card in bright light (EV 10 or similar). </li>Under the same lighting conditions, measure the grey card through your ND filter. Memorize the difference (e.g. EV -1). <li>Look for a place with dim lighting, just enough to make your grey card readout add up the low light sensivity level as spec'ed by the meter maker plus the ND filter factor determined above (e.g. EV 1.3 + EV 1 = EV 2.3). </li> <li>Repeat the measurement from above: do you get a reliable readout (the Ultraspot tended to show an underflow)? How much light does the ND filter hold back now, according to the meter? </li> </li> </ol> </li> <li>Infrared sensivity - you need a kitchen stove (and permission by your wife to operate it until an element starts to glow red w/o having a pot on it - anyway, keep a pot full of water at hand to cool down the stove element if you are done with the experiments , and a TV remote control (the latter preferredly loaded with a fresh set of batteries): <ol> <li>Measure the cold stove element in dim lighting (should read something like EV2-3). </li> <li>Turn on the stove, repeat measurements every minute. </li> <li>With my stove (30 years old Brown Boveri, element approx. 1.5kW), both the Ultraspot and Spotmeter F started to read contiously increasing EVs after ca. 3-4 minutes, up to EV+3 (Spotmeter F) / EV+4.5 (Ultraspot) before the stove element even started glowing (you should check this by switching off the lights regularly and stop the experiment once the red glow become noticeable - don't ruin your stove!). </li> <li>Fix the TV remote control in front of the spotmeter, so that the 1 degree circle points to the IR diode in the remote control's tip - if you don't want to dismantle the remote control's housing (most have an opaque plastic cap over the IR diode), you might have to try for a while to find out the exact location of the diode. The distance between the diode and the spotmeter should be chosen so that the diode fills the 1 degree circle in the spotmeter's viewfinder. </li> <li>Make measurement with the remote control switched off, memorize the spotmeter readout. </li> <li>Repeat measurement while operating the remote control - you'll probably have to make several (like 5 to 10) measurements here, because the remote controls I know send an oscillating signal. But if measured while on, my remote control made the Ultraspot and Spotmeter F really freak out (EV+5-7!!). If you think this has no practical relevance, try to get reliable measurements in a room where computers communicate via IRDA interfaces - I tried, and the Ultraspot's results where so ridiculous that even I noticed before burning film and switched to an incident meter. </li> </ol> <li>Flare - you need a piece of cardboard (grey, about 40x40cm) and a strong light source in an otherwise evenly lit room. I measured <ol> <li>outside the window on an overcast day for daylight </li> <li>a strong electric bulb (100W, opaque glass) inside for artificial light </li> </ol> <ul> <li>Point the spotmeter to the light source so that it fills the 1 degree circle indicator, make measurement and memorize readout. </li> <li>Now put the cardboard between the spotmeter and the light source; minimum distance to the meter should be approx. 1m (to exclude the possibility of vignetting the meter's optical system). </li> <li>Cut a hole in the cardboard where the spotmeter's 1 degree circle indicator points to - the hole should be just a bit larger than the circle in the viewfinder (i.e. you should see the light source in the viewfinder only through the middle of the circle indicator). <li>Repeat measurement - do you get the same reading as the first time? IMHO you should... for me, the ability to meter a small spot - and only a small spot - is the point about using a <it>spot</it> meter in the first place. </li> <li>If you don't get the same readout through the cardboard, and are patient, you can try to enlarge the hole incrementally until you get the same readout as w/o cardboard. This way, I found out that the Ultraspot works really &raquo;on the spot&laquo; (i.e. same readout within EV0.2 with a hole just large enough for the circle indicator), while the Spotmeter F should really be flagged as &raquo;2-3 degree meter&laquo; - the sensitive area in the viewfinder seems to be roughly the double size (area-wise) of the circle indicator. </li> </ul> </li> </ul>

[Stefan.Dalibor]

Oops, forgot to insert URL to Walter's original tests (sorry, Walter in the lenghty treatise above here it is.

Stefan