Using Tiny Digital Camera as Light Meter - Question

[andrew vincent]

Apologies if this question has been asked a hundred times and I stupidly cannot find the thread (please direct me) but I have a Linhof master technika and have been using a Canon 350XT for exposure and histogram information and occasionally framing, since it's often more useful than my linhof viewer. It's great, but still big and bulky in the sack with everything else. Reading Ken Rockwell's review of the tiny Casio Z850, with its color histograms, I was thinking how wonderful it would be if I could just get some cheap, TINY pocket camera that I might even mount to the top of the technika for this purpose. Obviously the central and only main question is whether the exposure calculations and histograms would be accurate. Obviously, if I could carry a pentax spot meter as well, I'd do so, but like many, I've found the histograms on a digital readout- as well as immediate preview of multiple exposures- the greatest invention since the light meter itself, so I'm not about to give them up. I'd just like to get the size and weight down as low as humanly possible.

Does anyone have any experience they'd like to share in this regard, or have any suggestions for relatively inexpensive, tiny cameras that might do the trick? I'm not really interested in SLRs, since the XT is already the smallest SLR made, and is a bit too big.

Thanks!

[stompyq]

I use my D70 just like you use your XT. If you are going to use a small point and shoot to check the histogram i would pick one that has a RGB Histogram and go under the dark cloth to check. Just looking at the image would be really misleading b/c of reflections etc.

[roteague]

I can't see what value having a histogram would have on LF photography. The response of a digital camera is nowhere the same as film. You would be better off just using a good spot meter.

[Richard Årlin]

Few digital cameras have an f:stop and typical shutter speeds used for LF... mine, a Ricoh GR has, up to 3 minutes on manual and yes spot metering, (not apertures) but isn't a very complicated way to go... just asking ?

[Alan Davenport]

Thanks, roteague, for saying what I was thinking. Histograms are a unique artifact of digital photography. They have no impact on LF photography, at least not until after the latent image is on the film.

Note to self: Add to my list of mysteries of the universe --How did St. Ansel make a living without histograms...

[C. D. Keth]

Note to self: Add to my list of mysteries of the universe --How did St. Ansel make a living without histograms...

Happily ;)

[andrew vincent]

Thanks, roteague, for saying what I was thinking. Histograms are a unique artifact of digital photography. They have no impact on LF photography, at least not until after the latent image is on the film.

A three color histogram is simply a more accurate reading of the light reflecting off of a given scene, rather than the usual 16% gray reflection. This is all clearly presented on other fora in much greater detail, but I'm not going to bother trying to convince anyone - just wanted to hear from anyone who had any particular stories of something working out.

[Marko]

A three color histogram is simply a more accurate reading of the light reflecting off of a given scene, rather than the usual 16% gray reflection. This is all clearly presented on other fora in much greater detail, but I'm not going to bother trying to convince anyone - just wanted to hear from anyone who had any particular stories of something working out.

Hi Andrew,

Just ignore the usual APUG carry-overs :)

This topic has been discussed (and argued) many times here. A few quick takes:

http://www.largeformatphotography.info/forum/showthread.php?t=21811&highlight=digital+camera+light+meter

http://www.largeformatphotography.info/forum/showthread.php?t=18226&highlight=digital+camera+light+meter

http://www.largeformatphotography.info/forum/showthread.php?t=17467&highlight=digital+camera+light+meter

I hope you'll find your answers there.

[neil poulsen]

With the right f-stop correction, I wonder if it wouldn't be useful for color. It would take some practice.

For example, in shooting transparencies, one might better be able to predict the correct exposure from the histogram by making sure there's no truncation on the right. Some digital systems have about the same latitude as transparancies. Perhaps both sides of the histogram might be useful.

For shooting negatives, one would be better served to make sure that the histogram isn't clipped on the left. In both cases, one's only looking at one side of the histogram. Again, there might be a certain number of stops correction, which could be determined through practice or testing.

It's worth giving it a try. Instead of 4x5, you could use 6x9 film for experimenting to save money.

[roteague]

A three color histogram is simply a more accurate reading of the light reflecting off of a given scene, rather than the usual 16% gray reflection. This is all clearly presented on other fora in much greater detail, but I'm not going to bother trying to convince anyone - just wanted to hear from anyone who had any particular stories of something working out.

The histogram is only indicative of what was captured by the sensor, not what was present in the scene. It would be valuable if sensors had the same response as film, but it doesn't.

[Marko]

The response of a digital camera is nowhere the same as film. You would be better off just using a good spot meter.


The histogram is only indicative of what was captured by the sensor, not what was present in the scene. It would be valuable if sensors had the same response as film, but it doesn't.

Following this logic, the same thing could be said about a spotmeter (or any other lightmeter).

A digital camera sensor and a traditional spotmeter both use photo-electric sensors with (very close to) linear response to light. Film, on the other hand, uses photo-chemical sensor with a response very different from linear. As a matter of fact, the response of any given film is very different from any other film, that's what makes them unique.

The histogram is a graphic representation of electrical signal as a function of the intensity of light that hit it. The same as needle sway on the old style light meters and the same as the led readout on the newer light meters. The ability of the photographer to understand and correctly interpret that data according to the film used is what makes the instruments useful.

If I am missing something about spectral response of a traditional lightmeter, its dynamic range or perhaps even its reciprocity awareness, I'd like to hear a little more about it than witty one-liner soundbytes.

[roteague]

Following this logic, the same thing could be said about a spotmeter (or any other lightmeter).

Perhaps, but my understanding is that the histogram is based upon the recorded pixels, not on the scene as a whole. That is why the point about it not being similar to film. A spotmeter doesn't have to be, since it is directly metering the subject and measuring the strength of the light being reflected.

[Marko]

Perhaps, but my understanding is that the histogram is based upon the recorded pixels, not on the scene as a whole. That is why the point about it not being similar to film. A spotmeter doesn't have to be, since it is directly metering the subject and measuring the strength of the light being reflected.

Well, that's where our understandings diverge, along with our opinions :)

My understanding is that both the digital camera sensor and the light sensor in the majority of contemporary spotmeters (or other lightmeters) are in essence electronic chips which translate the intensity of light that falls on them into the electrical signal of commensurate amplitude.

The histogram is simply a distribution curve as a function of the light intensity, the pixel-recording happens after the light is already measured. In fact, the pixel-recording also happens as a function of the light intensity levels captured, that is the essence of digital camera.

I also don't see how exactly is the response of a spotmeter similar to film? And if it is, then which film? Color or b&w? Negative or Transparency? How about reciprocity? And why exactly do you have to calibrate your entire system for each emulsion used?

But you don't have to take my word for it, there is a very simple experiment you can do to prove either your hypothesis or mine:

1. Simply take a spotmeter, take a few dozen close measurements of your subject, from shadows to highlights and plot a distribution curve.

2. Next, take a digital camera and frame your subject to contain the same area as measured with the spotmeter and take a shot, in RAW if possible, observing the histogram.

Chances are that both curves should be very similar in shape.

[Marko]

Perhaps, but my understanding is that the histogram is based upon the recorded pixels, not on the scene as a whole.

On further reflection, there are no pixels involved really, until the image is imported into an image-editing program. Until then, the entire image consists of either electrical (in camera) or magnetic signals (on CF card or disk).

The histogram cannot, therefore, be based on recorded pixels any more than the spotmeter's readout can be based on the film grain.

[roteague]

On further reflection, there are no pixels involved really, until the image is imported into an image-editing program. Until then, the entire image consists of either electrical (in camera) or magnetic signals (on CF card or disk).

Actually not, a digital camera is in essence a signal processing device. The signal, which is analog, goes through a series of filters and processors, to remove noise, determine color, etc, before going through an A/D converter - in camera. From that point it is digital - before it it written to either storage memory, or to the CF card.

[Marko]

Actually not, a digital camera is in essence a signal processing device. The signal, which is analog, goes through a series of filters and processors, to remove noise, determine color, etc, before going through an A/D converter - in camera. From that point it is digital - before it it written to either storage memory, or to the CF card.

True. But it is still an electrical signal all the way until it gets written to the card. There it becomes a magnetic charge.

Analog in this case simply means that the signal has a continious amplitude as a function of its intensity, a quality it acquires before being processed in any way. Demosaicing does not change that, it only enables the three color channels to be formed, but that's in addition to the integral one. Digital means that the amplitude has been divided into discrete levels. But the distribution curve remains pretty much the same.

Again, the bottom line is that both digital camera and a spotmeter use the same basic principle for measuring light intensity, both have relatively linear characteristic throughout their range and both are oblivious to individual emulsion characteristic such as emulsion composition, spectral sensitivity or reciprocity failure, even sensitivity to light in general.

No matter how good spotmeter you use, you still have to calibrate it, factor-in the correct filter factor and calculate the reciprocity failue compensation. All of this for each film you are using.

What you are getting in a digital camera is basically a fixed combination of light measuring device and light capturing device already factory callibrated. I see no reason to not be able to callibrate it to any of the films if one so desires.

[Ole Tjugen]

The histogram is only indicative of what was captured by the sensor, not what was present in the scene. It would be valuable if sensors had the same response as film, but it doesn't.

Nor does a spotmeter, or even a different film. One film doesn't have the same response an another one. That's why there are so many of them and some prefer one, some another.

[buze]

I use my 350D almost all the time for it's histogram when I'm outside. The histogram represents the JPEG output histogram and the not the raw (even if you shoot raw); so it's quite conservative if you shoot film, even slides.

I looked into getting a smaller "histogram device" and I had found a P&S that had a "lightmeter" mode, but it was a lot of interpolation. The small P&S cameras have a tiny lens, and a max fstop of f4 or so. A DSLR can do f22 and still have an accurate measure. I'd rather extrapolate a couple of stops that 8 of them...
But still, I'd love to have a nice accurate enough histogram device. The 350D is cool to "preview" yellow/green/orange/red filter effects in black & white too...

But of course as with every tools, you need to be the master of it. In the same way as you do with a spotmeter, you have to factor "peaks" in and offset whatever the histogram might say.

[Graeme Hird]

The way I think of a digital camera's sensor when used in this manner is that there are 6 million (or more) digital spot meters arranged in a systematic manner over a small area. The histogram reveals the number of spot meter readings which show a particular EV.

I see no good reason not to experiment and calibrate the system to your own favourite film.

But returning to Andrew's question: no, I don't have any particular experience with these small cameras when used as light meters. As soon as a good one comes on the market, I'll be buying my first digital light meter with recording capability. Until then, I'll keep reading these threads to find out the latest news.

Cheers,
Graeme

[Frank Petronio]

I'm a terrible slob and I don't bracket, but after trying my Nikon d80 dSLR and my trusted 20+ year old Pentax Digital Spot Meter side by side for a few shoots... I sold the Pentax Digital Spot Meter. Haven't had a bad outcome yet, maybe 30-40 different set-ups and situations.

I use the smaller prime lenses that are almost matched to my LF focal lengths. The smaller primes make the d80 that much easier to carry.

I just make the ISO and aperture and fiddle until I get a pleasing histogram and a decent preview. Of course I trust the histogram more than the screen image, but both provide valuable feedback.

You know what could really blow (APUG) people's minds? Using the dSLR as a color meter. Remember just a decade ago when you had to cough up $800 for a Minolta Color Meter III to be considered an architectural photographer? ;-)

[andrew vincent]

Thanks Marko, for the links, and everyone else for the informative discussion. I borrowed a friend's digital spotmeter for a number of years and loved it, but eventually had to give it back ;) I was thinking of buying another one, but decided to just try using the Canon 350XT's internal metering and histogram and see how it worked out. I'm honestly not sure it's any better or worse - it's certainly different. And I think the most important thing is to calibrate everything precisely, whatever you're using.

As for the tiny camera idea, after poking around the local mall, it seems that they all just have a fixed lens with a neutral density filter swinging in the way to attain a marginally higher effective f-stop, but even that is only f8 tops. With the Canon I have, I can just stop down to f22. Plus, with the (admittedly junky) zoom lens I picked up for $50 (which is so soft I don't even bother with a lens cap) I can have a fairly good zoom range (28mm-90mm in 35mm terms) to both preview and take quick visual notes for future exhibitions/trials. I found that the only cameras at the mall to have enough manual settings to be worthwhile started to get above $500 and were almost as big as the Canon 350, so that wasn't worthwhile.

I'll stick with this for now. Maybe pick up a used digital spot and go back to that for a while just for kicks...

[George Kara]

I find digital cameras to be much more useful for lighting setups with people. Moving all that equipment around without shooting anything seems to make the non-pro nervous and begin to stiffen up.

Taking a number of shots and showing the person images really seems to help with their anxiety. For using flash, I do check the histogram and shoot at the same asa as the film. I have always found the histogram to be accurate when compared to the film. Of course Im using a meter for flash however. I imagine the p&s digital may have a hot shoe but have no direct experience with one.

[ppisczak]

Andrew,

Like another poster, I use my Nikon D70s as previewer, just like our forebearers used poloroid film. I often keep the digital snaps I take so that I can send to friends or keep on my Ipod to show friends or Customers; this eliminates the need to scan later on. Makes digital cataloging a bit easier too.

It is my thought (prejudice) that the small digitals cannot compete with a decent digital SLR in terms of lens quality or raw files. The histogram is of course interesting and useful, and a good record for later on. But no matter what, I always use my spot meter to take and execute my light meter readings. My brain, though some may argue this ;) is better wired to interpret any scene than the best matrix meter. Nonetheless, it's all good. Best of luck


pj

[joolsb]

It is my thought (prejudice) that the small digitals cannot compete with a decent digital SLR in terms of lens quality or raw files.

The better compacts certainly can compete on lens quality and on RAW files. My Panasonic LX2 with Leica lens can hold its own up to A4 enlargements with all but the best DSLRs, given a subject without large areas of a single colour (where sensor noise becomes an issue).

[neil poulsen]

The connection between spot meter readings and histograms is interesting.

It looks like the histogram can be thought of as the accumulation of all possible spot meter readings (one per pixel) that could be taken in the frame of the photograph. An 8-bit spot meter reading would be a number between 0 and 255, depending on the brightness of the given pixel. (The pixel represents a specific location in the frame of the photograph.)

What I don't understand, is the relationship between that 8-bit number and the number that one would read on an analog meter reading of that given pixel, or more precisely the analog meter readings of that pixel's location in the scene. For example, how does that 8-bit number relate to the 1 to 19 reading I would get on my Pentax V Analog 1 degree spot reading? Don't know. Also, what would be the quality and accuracy of each pixel's reading, when compared the reading obtained from an analog spot meter.

When I spot read a scene, I take a number of readings in different locations to get an idea of how the light varies in that scene. The histogram represents the extreme case of that methodology. It's an accumulation of readings from "all possible locations" in that framed scene.

The other question that arises is how one would read the histogram to decide on an exposure and development. For black and white, I rely on my Zone III placement to determine my exposure and rely on my Zone VIII placement to determine my development. I wonder how one could extract the same information from a histogram? How would I determine both my development and exposure, based on the histogram? I suppose one could base exposure on the left end of the histogram, and base development on where the right end of the histogram falls in the 0-255 range. That is, base exposure on the EV that just falls short of truncating the histogram on the left. Correspondingly, base development on whether the histogram falls on the right end of the histogram. A narrow histogram would imply the need for expansion, and a wide histogram would imply the need for contraction.

As for color negative film, we usually expose for the shadows. So, base color exposure on what's happening at the left end of the histogram, provided the right end doesn't blow out. Correspondingly, we usually base color slide exposure on highlights, which would tend to focus our attention on the right end of the histogram.

For my purposes, I would have to base decisions for exposure (and development) on a lot of study of histograms in different lighting situations. Also, we spend a lot of money on a light meter to measure just one spot and get an accurate result. We don't spend nearly the same amount of money on some cameras that might be used for this purpose, and which take measurements on all spots within the scene. That raises the question on whether the digital point and shoot camera is even up to the task?

[Michael Graves]

Thanks, roteague, for saying what I was thinking. Histograms are a unique artifact of digital photography. They have no impact on LF photography, at least not until after the latent image is on the film.

Note to self: Add to my list of mysteries of the universe --How did St. Ansel make a living without histograms...

Don't know...but my mother seemed much happier after she had her histogram done.

[Marko]

The connection between spot meter readings and histograms is interesting.

It looks like the histogram can be thought of as the accumulation of all possible spot meter readings (one per pixel) that could be taken in the frame of the photograph. An 8-bit spot meter reading would be a number between 0 and 255, depending on the brightness of the given pixel. (The pixel represents a specific location in the frame of the photograph.)

What I don't understand, is the relationship between that 8-bit number and the number that one would read on an analog meter reading of that given pixel, or more precisely the analog meter readings of that pixel's location in the scene. For example, how does that 8-bit number relate to the 1 to 19 reading I would get on my Pentax V Analog 1 degree spot reading? Don't know. Also, what would be the quality and accuracy of each pixel's reading, when compared the reading obtained from an analog spot meter.

When I spot read a scene, I take a number of readings in different locations to get an idea of how the light varies in that scene. The histogram represents the extreme case of that methodology. It's an accumulation of readings from "all possible locations" in that framed scene.


Neil, histogram is actually a distribution curve of light intensities in an already captured image. Some cameras produce per channel histograms (three of them), while others show only a compound one. That is very different from spotmeter readings you take in the scene. In order to apply the histogram reading, you need to understand that difference, especially how it relates to your film.


The other question that arises is how one would read the histogram to decide on an exposure and development. For black and white, I rely on my Zone III placement to determine my exposure and rely on my Zone VIII placement to determine my development. I wonder how one could extract the same information from a histogram? How would I determine both my development and exposure, based on the histogram? I suppose one could base exposure on the left end of the histogram, and base development on where the right end of the histogram falls in the 0-255 range. That is, base exposure on the EV that just falls short of truncating the histogram on the left. Correspondingly, base development on whether the histogram falls on the right end of the histogram. A narrow histogram would imply the need for expansion, and a wide histogram would imply the need for contraction.

As for color negative film, we usually expose for the shadows. So, base color exposure on what's happening at the left end of the histogram, provided the right end doesn't blow out. Correspondingly, we usually base color slide exposure on highlights, which would tend to focus our attention on the right end of the histogram.

Maybe the best way to think of it would be to imagine reflective densities measured on a polaroid taken at the scene. That being said, I don't think you can use a digital camera as a spotmeter, much less instead of one. I would rather compare it to incident light meter with a distribution curve readout. It can show you where the most tonalities lie for a give exposure/f-stop set and it can show you the dynamic range spread, but only for that particular sensor - your film will behave differently and you need to know what the difference is in order to use the tool effectively. No different, really, than any other tool you use.


For my purposes, I would have to base decisions for exposure (and development) on a lot of study of histograms in different lighting situations. Also, we spend a lot of money on a light meter to measure just one spot and get an accurate result. We don't spend nearly the same amount of money on some cameras that might be used for this purpose, and which take measurements on all spots within the scene. That raises the question on whether the digital point and shoot camera is even up to the task?

It could be if one knows exactly what he is doing with it. It all depends on your usage and your expectations.