Use of cheap digital to act as both meter and viewscope

[Marko]

And to the histogram comment. Correct sir, and the reason I was displeased with trying to use a digital camera as a meter. The histogram, doesn't match up with any film I use when predicting shadows or highlights. A calibrated histogram to a film would be the bomb, but it doesn't exist, certainly not within any digital camera I've used.

A light meter inside a DSLR is very much the same as a stand-alone light meter. In both cases it is basically a photo-electric sensor that displays in one way or the other the intensity of the electrical current that the given amount of light generates. The manner in which that data is displayed is simply a matter of user interface. A histogram is, in a nutshell, a distribution curve of the signal strength across any given matrix of individual photocells that comprise the sensor. Any modern spotmeter could feature the same thing if equipped with an appropriate display and software/firmware, but all it would really accomplish would be to raise the price. DSLRs have it simply because they already have both the software and the display.

The question is not whether an electronic sensor could be calibrated to a film, the question is which film it should be calibrated to? And, at least in case of B&W films, which developer as well...

Just like with stand alone light meters, it is ultimately up to the photographer to interpret the data provided by a measuring device and translate that into desired tones and DR. With that in mind, which measuring device one uses is pretty much a matter of personal preference.

[Bob McCarthy]

I agree with the comment it is a meter, but it is not linear in the same way. The range left to right is not a standard quantity but an output that is matched up to the sensor, whatever the range of the sensor is. I suppose one could create a zone (ish) scale overlay over the rear lcd, by testing and trial and error. The middle of the range is not necessarily zone V, as most camera have bias in recommended exposure settings to protect highlight blowout.

If one didn't use the matrix meter and computer in the camera to determine exposure, and switched to the spot meter and calibrated the histogram, I suppose it would be usable.

BTW, the histogram is not raw data delivered from the meter, but a distribution of values from an internally generated JPEG processed in the camera with the thoughts of the manufacture built in to give the average user a high percentage of usable shots. That is far from just a light meter.

I tried to make it work, but settled on the reliability and sensibility of my Zone VI modified meter.

I might mention that I do believe in using a film camera as a meter. My F5 meters for slide film pretty darn well, but then the matrix meter was designed for the same film we use in LF.

Bob

[timparkin]

Although I don't use my digital for a light meter, I did play around with changing the jpg settings as they alter the histogram (good point for digital folk, always set your jpg settings to default or you'll get an incorrect raw histogram). I managed to get a good velvia replica by upping the contrast to maximim (on a 5D) and I also changed the preview to black and white to show intensity levels... It just didn't work for me as a way of understanding exposure though.. I bought a Sekonic 558 and didn't get on with this particularly either and then bought a Pentax Digital Spotmeter which is the bees knees :-)

Tim

[Bob McCarthy]

Let me suggest a simple test.

Put your digital on a tripod.

take 3 shots, one raw, one low iso jpeg, one high iso jpeg.

Look at the histograms. They will all be different. Sometimes a little, sometimes a lot depending on scene contrast range.

Same scene, same lighting, different histogram. The manufacturer is modifying the data in Jpeg processing.

[roteague]

I tried to make it work, but settled on the reliability and sensibility of my Zone VI modified meter.

I might mention that I do believe in using a film camera as a meter. My F5 meters for slide film pretty darn well, but then the matrix meter was designed for the same film we use in LF.

Bob

That is pretty well my conclusion as well. I sometimes use my Nikon F6 as a light meter; I've found my D200 has a tendency to underexpose everything too much. I will normally use the same film in both my LF and F6, and the matrix meter on the F6 is just fantastic.

I recently bought a Sekonic L-758D Digital Master light meter, however I'm having a bit of trouble getting used to using it - the view finder doesn't have the field of view my old Sekonic had.

[Bob McCarthy]

I managed to get a good velvia replica by upping the contrast to maximim (on a 5D) and I also changed the preview to black and white to show intensity levels...

Tim

Interesting approach. The downside of the histogram on many cameras, is its measuring luminance with the green channel only. For outdoor landscapes, you can imagine the bias. Not all cameras do this now, but it was the standard way. Pro cameras will output all three channels. It your histogram will output an rgb histogram, you' are seeing all three channels.

[joolsb]

Any modern spotmeter could feature the same thing if equipped with an appropriate display and software/firmware, but all it would really accomplish would be to raise the price.

Hardly. A spotmeter only takes readings of what it happens to be pointed at and an incident reading just reads an overall light level. What a histogram does is to represent the output of the many thousands of photosites of a digital sensor in a graph. It's an entirely different approach. Unless it was to work just like a digital camera, a handheld meter cannot produce a workable histogram, no matter how clever the software might be.

[Marko]

BTW, the histogram is not raw data delivered from the meter, but a distribution of values from an internally generated JPEG processed in the camera with the thoughts of the manufacture built in to give the average user a high percentage of usable shots. That is far from just a light meter.

Well, yes, a histogram is a distribution curve, I believe I said that much. It is not a light meter. But the light meter built into the camera is a light meter. My point was exactly that the two are not the same and that a correct interpretation is up to the user.

And I do agree that the linearity (whatever that really meant) is definitely not the same with digital sensors as it is with film.

I might mention that I do believe in using a film camera as a meter. My F5 meters for slide film pretty darn well, but then the matrix meter was designed for the same film we use in LF.

And again, the question is: which film?

I don't use slide film at all. I use negative film and then mostly B&W. Each of those has its own, very different response curve, which brings another question: in which developer?

Just to make myself clear, I am not suggesting that a DSLR is as good as a dedicated spot meter for this purpose, quite to the contrary. I use a Seconic 558 myself for LF, after all. What I am suggesting is that it is very much possible to use it that way, provided that one takes all the trouble of calibrating it to the film and developer one wants to use, just like it is normally done with regular spot meters.

Let me suggest a simple test.

Put your digital on a tripod.

take 3 shots, one raw, one low iso jpeg, one high iso jpeg.

Look at the histograms. They will all be different. Sometimes a little, sometimes a lot depending on scene contrast range.

Same scene, same lighting, different histogram. The manufacturer is modifying the data in Jpeg processing.

Yes, they do, they have to. JPEG is a format which utilizes variable compression, a lossy one at that. A file has to be processed in order to be compressed. That's one of the reasons RAW is so much better.

But again, a histogram is just a tool, the same as the camera that displays it. It does not do the thinking for us any more than a stand alone light meter does. It is the user who is supposed to interpret the data.

[Bob McCarthy]

Well, yes, a histogram is a distribution curve, I believe I said that much. It is not a light meter. But the light meter built into the camera is a light meter. My point was exactly that the two are not the same and that a correct interpretation is up to the user.

And I do agree that the linearity (whatever that really meant) is definitely not the same with digital sensors as it is with film.




And again, the question is: which film?

I don't use slide film at all. I use negative film and then mostly B&W. Each of those has its own, very different response curve, which brings another question: in which developer?

Just to make myself clear, I am not suggesting that a DSLR is as good as a dedicated spot meter for this purpose, quite to the contrary. I use a Seconic 558 myself for LF, after all. What I am suggesting is that it is very much possible to use it that way, provided that one takes all the trouble of calibrating it to the film and developer one wants to use, just like it is normally done with regular spot meters.




Yes, they do, they have to. JPEG is a format which utilizes variable compression, a lossy one at that. A file has to be processed in order to be compressed. That's one of the reasons RAW is so much better.

But again, a histogram is just a tool, the same as the camera that displays it. It does not do the thinking for us any more than a stand alone light meter does. It is the user who is supposed to interpret the data.

The problem is the manufacturers modify tones, not just compression. It has been found with digital, that over exposing reduces noise, so the whole exposure is shifted to the right as long as the highlights are not negatively impacted(blown out). So the scene determines the exposure not just the light. This is absolutely true with raw setting, lesser so with jpeg. There is a disconnect between the meter and the histogram displayed, modified by the manufacture to help the "great unwashed" get a high percentage of acceptable shots.

If you go to my first post, i mention that for B&W the digital sensor does fit better, but is not reliably accurate in my opinion. And as you point out, film type and pushing /pulling does modify the DR of the negative. Obviously the histogram could never accommodate this. I thought the poster who modified his jpeg setting to fit his slide film had a clever approach for that singular situation.

Your also right that I was only thinking in terms of transparency film, not color negative, in my post. Sorry for any confusion.

Bob

[Joseph O'Neil]

The problem is the manufacturers modify tones, not just compression. It has been found with digital, that over exposing reduces noise, so the whole exposure is shifted to the right as long as the highlights are not negatively impacted(blown out). So the scene determines the exposure not just the light. This is absolutely true with raw setting, lesser so with jpeg. There is a disconnect between the meter and the histogram displayed, modified by the manufacture to help the "great unwashed" get a high percentage of acceptable shots.
Bob

-snip-

I agree completely with you on that point. Just yesterday I was down to the local river (flooding - great shots ), and I had both my Tachi and my Nikon D40. How my D40 "saw" the light compared to my Pentax spot meter was two different things altogether.

So yes, while you can take time to calibrate your digital SLR to match your film, there are always occasions it will be out. Low light for example, as I find that even if you set the ISO on your DSLR the same as your film, at least in my D40, the CCD sensor is a lot more sensitive to low light than film, especially when you start getting into exposures of 1/2 or less. CCD chips do NOT suffer reciprocity failure.

I'm of the opinion if you forget your light meter, hey, use your DSLR for sure, better than nothing, but even thought I drag my DSLR with me when backpacking with my 4x5, I find from first hand experience, in may ways, pulling out the ole spot meter just works better in many ways

joe