Metering, what's gone wrong?

[Paul Dunning]

Hi All,

I’m wondering if someone would help me diagnose a metering problem.

Recently I seem to be underexposing and I’m trying to figure out what I’m doing wrong. I used to just take a stab at what the mid-tone should be but to take more control I decided to try averaging the darkest and lightest areas that I want to retain detail in.

The scene below is a case in point, the first thing I did was meter the highlights in the foreground and the sky, this indicated a .6 GND was needed and looking at the balance in the slide that part seems to have been about right.
Now the problem, for the foreground I metered a spot in the shadows and the brightest foreground highlight and took an average (circled areas in the image below), which worked out as 1 second @ F/29, unfortunately I didn’t make a note of the range but If I remember correctly it was approximately 1.5 stops above and below the average reading.

The resulting slide (raw scan from Silverfast) is shown below along with a snapshot of it's histogram. I would guess that it’s under exposed by at least 1.5 stops, maybe 2.

Film used was RVP50, expired 08/11 but kept in the fridge, could this be the culprit? I took A and B shots, one developed at home, the other (the one shown here) developed by Peak Imaging (this was part of a test to check my developing times are right, the one developed at home did turn out a bit brighter which I took to indicate my first dev time was a bit too long, but that's a discussion for another day

Do you think my metering technique is flawed or should I be looking elsewhere for the problem?




Thanks,
Paul.

P.S.

I wouldn't rule out user error, it's quite possible that I suffered some kind of brain freeze and made a dumb mistake but I took 6 shots on the evening in question and they all suffered from the same problem to varying degrees.

[Ian Johnson]

Hi Paul

I read your post today and felt the need to respond, since I have been having similar issues regarding exposure. I have just started in LF (end Sept 2011) and have shot about 100 sheets in total so far. So I am by no stretch of the imagination claiming to be anything other than a complete novice. The first batch I had processed (20) all came out reasonably well. However, a second batch of 40 showed what a novice I am, as the majority were underexposed.

I then had to re-evaluate and read as much as I could on spot metering (there's lot on this forum). As far as I can tell the seasoned, more knowledgeable practitioners don't look for mid-tones in a scene, but look at the scene and decide where they want to place certain tones (sunlit grass +0.3, yellow grass +1, white with detail +1.7 etc).

I have been using this method since then and my last two batches of 20 slides have been very good in terms of exposure. I often, for reassurance check exposure by measuring the back of my hand - this reading would be placed at +1.3. As a belt and braces approach I have also bought a small greycard from calumet which allows me to check where to place tones with which I'm not familiar. I also now make sure I take note of meter reading from back of hand, grey card etc for later comparison.

I found the best explanation of this method on Richard Childs' website, which includes how to decide on what strength grad to use.

Hope this helps in some way.

regards

Ian

[Paul Dunning]

Hi Ian,

Thanks for the reply, I was starting to think I wouldn't get one

I just had a look at Richard's blog entry on exposure, very interesting, thanks for pointing me towards it.

It's funny, when I started out I'd stand there for ages trying to find a mid tone to meter from, later I started doing something more similar to what Richard describes, i.e. pick a tone and decide where I wanted to place it, I'm sure I'll go back to this technique but I guess where I am confused is just in understanding how taking an average reading between the shadows and highlights can result in an underexposed slide where all the information is below the histograms mid-point when scanned. Maybe I'm just being dense, like my slides

Paul.

[dave_whatever]

Hi Paul

I won't comment on specifics of your shot as its hard to say without seeing the film on a lightbox how bad it is. Raw scans do always look dark, that par for the course.

However as far as metering goes I would bin the idea of finding the average reading between the lightest highlight and the darkest shadow. This method will work but only if the range of brightness is within what the film is capable of holding, and also it assumes that the base exposure is happy sitting halfway between the two, which might not often be the case. In a way by just following the mechanical method of determining the average readind you're kind of doing what 35mm SLR meters do, which will work a lot of the time but not others. If you've gone to the length of using a spotmeter then it really is worth taking the plunge and using it as suggested above and placing key values within the scene where you want them.

A few pitfalls on metering I have noted in my experience are:
- using a filter and forgetting to add exposure. If you forget to compensate for using a polariser for example you're likely to end up with an unscannable slide.
- shooting in very low light with Velvia, where during the exposure the light has faded even further compared to when you took the meter reading, hence needs even more exposure but its easily missed.

One thing I will say on the above shot is that its worth being careful with using grads when you've got water in your ungradded area which reflects the sky. In your shot the river water is brighter than the sky, yet as a reflection of the sky it should be dimmer than the sky or maybe about the same at best if its to look "natural", or at least for the filter use not to draw attention to itself. I used to get tripped up by this a lot years ago on medium format, particularly worth remembering with rivers, lakes and the sea.

[Paul Dunning]

Hi Paul

I won't comment on specifics of your shot as its hard to say without seeing the film on a lightbox how bad it is. Raw scans do always look dark, that par for the course.

However as far as metering goes I would bin the idea of finding the average reading between the lightest highlight and the darkest shadow. This method will work but only if the range of brightness is within what the film is capable of holding, and also it assumes that the base exposure is happy sitting halfway between the two, which might not often be the case. In a way by just following the mechanical method of determining the average readind you're kind of doing what 35mm SLR meters do, which will work a lot of the time but not others. If you've gone to the length of using a spotmeter then it really is worth taking the plunge and using it as suggested above and placing key values within the scene where you want them.

A few pitfalls on metering I have noted in my experience are:
- using a filter and forgetting to add exposure. If you forget to compensate for using a polariser for example you're likely to end up with an unscannable slide.
- shooting in very low light with Velvia, where during the exposure the light has faded even further compared to when you took the meter reading, hence needs even more exposure but its easily missed.

One thing I will say on the above shot is that its worth being careful with using grads when you've got water in your ungradded area which reflects the sky. In your shot the river water is brighter than the sky, yet as a reflection of the sky it should be dimmer than the sky or maybe about the same at best if its to look "natural", or at least for the filter use not to draw attention to itself. I used to get tripped up by this a lot years ago on medium format, particularly worth remembering with rivers, lakes and the sea.

Hi Dave, and thanks for the reply.

I probably will ditch the averaging technique. It was something I wanted to try, the idea wasn't to so much to mimic what an SLR would do although of course in some respects that is what I am doing but taking more control by picking the points that I want to average. I figured that taking an average of two readings and then looking at the range indicated on the spot meter would tell me if I was within the latitude of the film, e.g. for the scene above I think the meter said my two readings were 1.5 stops above and below the averaged mid-tone which should have been fine, given that the range was good I went with the average reading for my exposure expecting all to be good. If the range had been greater than the films latitude then I would have adjusted up or down depending on whether I wanted to preserve shadow detail or highlights. It seemed like a good idea but I guess not

With the water being brighter than the sky, I was kind of aware of it at the time, my thinking was to make sure I didn't blow the sky and if the water was a bit bright I'd re-balance it after scanning, a bad habit I've picked up from years of shooting landscapes with digital

Thanks,
Paul.

[Julian Elliott]

Morning Mr Dunning!

Remember when you first started your venture into LF, I did say to you that Velvia was a pig if you fouled it up

When I'm out and about I use a variety of techniques. A couple I got from Charlie Waite and others are as previous posters have said.

The huge problem with exposure is we are always told to look for that "mid-tone". Well what the hell is a mid-tone! Placing those tones as advised is the more easier way of doing things. I try to find the place in my image that I want as black and then minus off 2 stops from the reading.

A technique I remember from Charlie was to use the incident reading. He took three readings:

1) At 90° to your left.
2) Facing you with half the ball in shade (so look for the line going vertically downwards).
3) At 90° to your right.

I do remember him muttering things under his breath about placing readings on zones or tones.

Confused or makes any sense?

Jools

PS I'm over in the UK at the end of March and can show you all I know. My only problem is I won't have access to a car

[Paul Dunning]

Morning Mr Dunning!

Remember when you first started your venture into LF, I did say to you that Velvia was a pig if you fouled it up

When I'm out and about I use a variety of techniques. A couple I got from Charlie Waite and others are as previous posters have said.

The huge problem with exposure is we are always told to look for that "mid-tone". Well what the hell is a mid-tone! Placing those tones as advised is the more easier way of doing things. I try to find the place in my image that I want as black and then minus off 2 stops from the reading.

A technique I remember from Charlie was to use the incident reading. He took three readings:

1) At 90° to your left.
2) Facing you with half the ball in shade (so look for the line going vertically downwards).
3) At 90° to your right.

I do remember him muttering things under his breath about placing readings on zones or tones.

Confused or makes any sense?

Jools

PS I'm over in the UK at the end of March and can show you all I know. My only problem is I won't have access to a car

Morning Mr Elliot

Makes perfect sense.

I've often done similar to you, meter the shadows and then close down 2 stops, also sometimes meter the highlights and then open up two stops.

Other times I'll spot a recognisable mid tone and meter off that, generally with acceptable results, but as you said, what is a mid-tone, it's often difficult to spot one, we can pick likely suspects, for example, at the beach I've found metering sand at mid+1 seems to work quite well.

I was trying the averaging technique just to have another tool in the bag for scenes where it felt like it would be useful.

What I can't seem to get to grips with is why averaging results in an underexposed slide when the same technique works as expected if I use it with with my 5DMkII. I'm beginning to suspect bad film or some other problem.


Paul.

PS.

What dates are you in the UK for, end of March is quite busy for me but maybe we can arrange a dayout near you, I could chauffeur.

[jennym]

Have you checked the accuracy of your spotmeter in case the problem is with the meter? Do you have a way of checking it against another meter?

Just a thought,

Jenny

[Martin Jan Köhler]

Hello Paul,

I also had a lot of underexposed picture problems in the beginning,
also using RVP50 transparencies and a combined spot meter and incident meter.

I was only using the spot meter and tried to meter the brightest and darkest tones and place them.
Sometimes it worked well, sometimes not.
Then I went to a workshop with an austrian large format photographer,
and saw that he was mostly metering incident (by spotting metering a grey card).
We compared also our meters (he tested both, mine was 1/3 stop too dark) but that was not the main issue.

One thing that helped me overall is to rely more on incident metering for the foreground areas.
I then spot meter critical sections to make sure they are in the dynamic range of the film.

With sky included, the spot meter can really easily skrew up your measurings,
I measure the foreground for the desired exposure, then decide on a grad by comparing this foreground measuring with the sky.
That works quite well.

Also with RVP50 you have to compensate for the Schwarzschild-Effect (reciprocity failure), what I do is...
measured 1s -> do 1s
measured 2s -> do 4s
measured 4s -> do 8s
measured 8s -> do 16s
measured 30s -> do 1m
measured 1m -> do 4m
measured more -> go home crying

Then you have the problem with wide angles, f.ex. the center of my 80mm Schneider is 1,5 times brighter than the corners.
So I use a center filter (1,5 EV strength) and expose 1,5x (and then compensate for reciprocity).

And then if you do very near subjects with long bellow draws, you perhaps have to expose for bellows factor,
but that's more rare with 4x5 ... there's a QuickDisc PDF online which allows you to measure that in a simple way...
http://www.salzgeber.at/disc/

Regards,
Martin

[Paul Dunning]

Have you checked the accuracy of your spotmeter in case the problem is with the meter? Do you have a way of checking it against another meter?

Just a thought,

Jenny

Hi Jenny,

It's a good thought

I think it's accurate, I did a test a while back using the readings from the spot meter to set exposure on my digital camera and then comparing the results with the camera's meter and they were pretty similar, not very scientific but I figured it was a good indicator.

I'll try this again to be sure.

Thanks,
Paul.

EDIT: tested again and the spot meter and digital camera meter (in spot mode) agree to within 1/3 of a stop so I guess that can be ruled out as the problem

P.S.
Just in case it's relevant, the meter in question is a Sekonic 758D.

[Paul Dunning]

Hello Paul,

I also had a lot of underexposed picture problems in the beginning,
also using RVP50 transparencies and a combined spot meter and incident meter.

I was only using the spot meter and tried to meter the brightest and darkest tones and place them.
Sometimes it worked well, sometimes not.
Then I went to a workshop with an austrian large format photographer,
and saw that he was mostly metering incident (by spotting metering a grey card).
We compared also our meters (he tested both, mine was 1/3 stop too dark) but that was not the main issue.

One thing that helped me overall is to rely more on incident metering for the foreground areas.
I then spot meter critical sections to make sure they are in the dynamic range of the film.

With sky included, the spot meter can really easily skrew up your measurings,
I measure the foreground for the desired exposure, then decide on a grad by comparing this foreground measuring with the sky.
That works quite well.

Also with RVP50 you have to compensate for the Schwarzschild-Effect (reciprocity failure), what I do is...
measured 1s -> do 1s
measured 2s -> do 4s
measured 4s -> do 8s
measured 8s -> do 16s
measured 30s -> do 1m
measured 1m -> do 4m
measured more -> go home crying

Then you have the problem with wide angles, f.ex. the center of my 80mm Schneider is 1,5 times brighter than the corners.
So I use a center filter (1,5 EV strength) and expose 1,5x (and then compensate for reciprocity).

And then if you do very near subjects with long bellow draws, you perhaps have to expose for bellows factor,
but that's more rare with 4x5 ... there's a QuickDisc PDF online which allows you to measure that in a simple way...
http://www.salzgeber.at/disc/

Regards,
Martin

Hi Martin,

Thanks for the the comprehensive reply, very useful.

I'll give the grey card method a go this weekend.

Paul.

[Paul Dunning]

Thanks everyone for your thoughts...

I've now ruled out the spot meter as a problem, it seems to be calibrated fine, at least it's within 1/3 of a stop of the spot meter in my 5DMkII.

So I'm think I'm left with the following...

Problem with film, seems unlikely but I'll try shooting some new in-date film alongside the expired (08/2011) stuff which I've been using.

Sticky or inaccurate shutters, also I would think unlikely but I guess I should try and test this somehow, 1 second sounds like 1 second.

And the most probable, me being a complete numpty and just getting it wrong in the field despite thinking I know how to do it

Paul.

[Charles Twist]

Hello Paul,

I've been quite busy and have only just looked at this thread on my calibrated monitor. Are you sure you are that far out? OK the sky is definitely dark and that could be a mistake: you say

the first thing I did was meter the highlights in the foreground and the sky, this indicated a .6 GND was needed

Well clearly the brighter part of the river is brighter than the sky...

Now looking at the two circled areas: wouldn't you say that the average of those two is near enough neutral? V50 goes dark very quickly and 1.5 stops under scanned on a V700/V750 is going to give a fair bit of darkness. Add to that the oft-quoted optimism of V50's speed: between my meter (out by 0.2 stops) and the emulsion, I systematically add 0.5 stop of exposure. Add to that any reciprocity failure or light fall-off, and your result might actually be 1 stop darker than you expect. It took me ages to overcome the darkness of V50, so don't give up just yet.

It's also worth looking at the aperture labels: are they in the right place? Two of my lenses have an offset, by up to a stop.

Does that help?

Charles

[dave_whatever]

Problem with film, seems unlikely but I'll try shooting some new in-date film alongside the expired (08/2011) stuff which I've been using.

I'd be amazed if the film was the problem unless its been stored in an over or something. I've shot slide film over a decade old, uncertain storage, with no noticeable change in film speed. 2011 is practically brand new!

[Paul Dunning]

Hello Paul,

I've been quite busy and have only just looked at this thread on my calibrated monitor. Are you sure you are that far out? OK the sky is definitely dark and that could be a mistake: you say

the first thing I did was meter the highlights in the foreground and the sky, this indicated a .6 GND was needed

Well clearly the brighter part of the river is brighter than the sky...

Now looking at the two circled areas: wouldn't you say that the average of those two is near enough neutral? V50 goes dark very quickly and 1.5 stops under scanned on a V700/V750 is going to give a fair bit of darkness. Add to that the oft-quoted optimism of V50's speed: between my meter (out by 0.2 stops) and the emulsion, I systematically add 0.5 stop of exposure. Add to that any reciprocity failure or light fall-off, and your result might actually be 1 stop darker than you expect. It took me ages to overcome the darkness of V50, so don't give up just yet.

It's also worth looking at the aperture labels: are they in the right place? Two of my lenses have an offset, by up to a stop.

Does that help?

Charles

Hi Charles.

Yes, that does help.

I guess I knew that some exposure was lost through scanning but I hadn’t thought of it being as much 1.5 stops, this explains a lot.

The .6GND was clearly a mistake; I’ll be more careful next time

I just had a new look at the slide for the image in my original post on the light box, it’s a bit darker than I’d like so it was underexposed but not by as much as I was originally thinking based on the scan. I'm now thinking that the result is correct for the metered exposure and that the average of those two points has indeed come out around neutral, so I think what I should have done is opened up by about 0.5 stops to get the highlights a bit brighter, and probably used a .3 GND for the sky instead of a .6

I think my aperture labels are correct though I'm not certain how I'd know if they weren't.

I’ll admit my confidence has taken a hit after the last few outings; I seemed to be getting much better exposures 3 months ago although they were a bit hit and miss, recently it seems to be consistently bad so I need to get to the bottom of what I’m doing differently. This thread is proving very helpful in that respect and I now feel that I’m close, I just need to tweak the way I’m doing things a bit. Don't worry about me giving up any time soon, I didn't get into this for an easy life that's for sure

Thanks,
Paul.