Mein kleines Mondlicht,

As a result of a question raised in an Internet photography forum regarding the usefulness of a spot meter for High Dynamic Range Imaging, I felt prompted to describe to You a procedure whose simplicity expedites and facilitates what can be one of the trickiest –yet extremely crucial– procedures in photography: metering.

Scenes whose dynamic range exceed that of the camera require the photographer to make a decision in terms of final rendering. It comes with the basic, primordial, and elemental question a photographer must always ask himself before bringing the eyepiece to his eye and pressing the shutter release button; namely, ‘what do I want?’ This question triggers a number of derivative interrogations: ‘What kind of photograph is this one going to be: an all-revealing documentary picture, or one that emphasises certain elements over others?’ ‘Thus, which tones do I want as my key tones?’ ‘Where do I want the viewer to look?’ And so forth.

The base exposure yields the building's facade too dark, even though the sky and the sun-litten wall are correctly exposed. Exposing for the facade, however, would blow out the sky and the wall: HDRI is an adequate solution to this problem. (24mm, 1/160 sec @ f/16, ISO 320)

Plate 1. The base exposure yields the building’s facade too dark, even though the sky and the sun-litten wall are correctly exposed. Exposing for the façade, however, would blow out the sky and the wall: HDRI is an adequate solution to this problem. (24mm, 1/160 sec @ f/16, ISO 320)

In this particular case, let me say I am after an-all revealing documentary shot. Thus, I want no blocked-up shadows and no blown out highlights. Evaluating my scene I find that I’m shooting at 2:26 P.M. in a hot summer day, under a bright blue sky partly populated by moving cumulus clouds. My subject, a building, faces to the east and its façade –which is at a right angle with a featureless wall– is in shade (now, the featureless wall is still mostly illuminated by the very bright sun; it helps, however, that the colour of the wall is grey, thus preventing the dynamic range from being even greater).  There is also a vehicle parked in front of the building, in whose chrome mouldings I can see extremely bright reflections of the daystar. Knowing what I want, I quickly make my creative exposure decisions:

  • I choose f/16, for I want everything to be in focus, from the foreground to the background.
  • I choose ISO 320, for I am using a circular polariser over my lens that cuts the light by roughly 2.2 stops, and I want to keep my shutter speeds as short as possible, both because there is a rather strong wind, and also because I don’t want to run the risk of people walking into the frame or cars suddenly parking in front of the building.
  • A quick snapshot (Plate 1) tells me that I will be wanting to use High Dynamic Range Imaging; ergo I shall need to take several exposures. I usually aim, in cases such as this one, for five frames spaced by two stops each. Since I am not including the light source in the frame, five exposures should suffice. The reflections of the sun in the car will blow to white, but I am not concerned about that, as they are very small and have no important detail in them; in fact, with my use of a small aperture these reflections will become ‘star bursts’, an effect I find particularly pleasing.

I have already set up my camera for my intended composition and, having made my creative decisions, am ready to shoot. Well, actually, not quite, for there is one more procedure to perform: metering.

I take out my light meter and take an ambient reading to gauge the overall exposure value. The reading (that has been automatically adjusted to compensate for the polariser [as is the case of all other light meter readings mentioned here]) is 1/160 sec @ f/16 with ISO 320.  Now, I could set this exposure as the base and calculate my bracketing from this starting point, but professionals are not allowed to guess: we are demanded to be precise. When precision is a must, spot metering becomes the only choice. And for this ‘choice’ I have two options: I can use the camera’s spot meter or the light meter’s.

The Sekonic L-758DR light meter in spot metering mode, waiting for me to look through the eyepiece and take the reading.

Plate 2. The Sekonic L-758DR light meter in spot metering mode, waiting for me to look through the eyepiece and take the reading.

Using the camera’s spot meter is the worst alternative. For starters, I would need to take the camera off the tripod in order to aim the centre circle in the viewfinder towards the areas I need to meter; now that I have my composition all carefully set up, I am perfectly unwilling to mess it up. Besides, the camera’s is a 3º spot meter and if the zones I need to meter are very small in the frame I would need to walk up to them so as to make sure the circle is contained within them that surrounding areas with differing brightness values are not throwing my readings off. Another hassle, eh? Alright, I could walk up to the windows and meter the dark patches, but can I walk up to the clouds? Not in this lifetime.

Now, my Sekonic L-758DR’s spot meter is 1º. As of now, one cannot get any more precision than that. Oh, and did I mention that it does not require me to touch my camera and mess up my composition? Again, only one ‘choice’.

Thus, I quickly look through the viewfinder and identify the brightest and the darkest zones in the frame. They’re the ones marked  in Plate 3 with red and green circles, respectively. Those are the points where I need to aim the meter at. I proceed to do so.

Plate 3. Spot-metering off the brightest and darkest areas of the image.

The reading off the white area of the cloud (we’ll call this the Highlights Value, for it determines the highest shutter speed value) is 1/2000 sec at my predetermined aperture of f/16 and ISO of 320. I could just proceed to make my bracketing with this value and forgo the reading of the shadow area (as a matter of fact, if I remember correctly, that is exactly what I did, having already determined that I wanted five exposures, and since I was using the PhotoBuddy app on my smart phone to calculate my bracketing), but for the illustrative purposes of this epistle, let us say that I take a reading of the very dark area in the window (which we’ll call the Shadows Value, as it determines the lowest shutter speed) and it registers a shutter speed of 1/8 sec.

I now have the high and low limits for my range of exposures. All I have to make sure now is that my brackets fall within that range. I’m ready to start making pictures now.

Again, I proceed to do so (Plates 4-8).

Image 2.

Plate 4. Highlights Value.

Image 3-B.

Plate 4-B.

Image 3.

Plate 5. Highlights Value + 2 stops.

Image 3-B.

Plate 5-B.

Image 5.

Plate 6. Highlights Value + 4 stops.

Image 5-B.

Plate 6-B.

Image 6.

Plate 7. Highlights Value + 6 stops.

Image 6-B.

Plate 7-B.

Image 7.

Plate 8. Highlights Value + 8 stops; alternately, Shadows Value.

Image 7-B.

Plate 8-B.

 

And after merging the exposures in Photomatix and processing the resulting image in Lightroom and Photoshop, just like I would do with any other, non-HDRI picture, I obtain my final photograph (Plate 9), the one I send to the client.

Image 8. The final image.

Plate 9. The final image.

 

Discussion

Now, before I wrap this epistle up, Liebchen, and let You go back to your quotidian tasks, I want to elaborate briefly on the principles behind the technique.

Lightmeters that sport both incident (the white dome) and reflective (spot) modes operate differently under such modes. With incident metering, what one sees is what one gets, essentially. You press the button, the meter takes a reading of the light hitting the dome, You read the values and plug them into the camera, and presto!, perfect exposure.

Image 11. A light meter that has both incident and reflective metering is a boon for the photographer doing architectural and landscape work.

Plate 10. HDRI or no HDRI, a light meter that has both incident and reflective metering is a boon for the photographer doing architectural and landscape work.

Reflective metering, on the other hand, report a value that is that of middle grey (or mid tone, or 18% grey, or Photoshop’s 50% HSB— whatever You want to call it).Thus, when I took a reading off the white patch in the cloud –the Highlights Value– the meter did not see that area as white, like my eye did, but as middle grey; and when that value was plugged into the camera and the exposure made, the camera arranged all other tones around that middle grey: that is why the histogram (Plate 4-B) has over 90% of the tonal values hugging the left edge and the image looks grossly underexposed, with many areas being rendered as full black, or 0,0,0 in RGB values. You can see this clearly in Plate 11, which is the same as Plate 4 but with the Raw processor’s clipping warning shewing (all the bright blue patches would print or display on screen as pure black: all detail in those areas has been lost).

The opposite is true with the Shadows Value: the meter decided the black patch in the window was middle grey and the camera, following its lead, re-arranged all tones around it: now we have a histogram pushed towards the right with no contact with its left edge (Plate 8-B), and an image completely nuked into overexposure but that has detail in that one formerly black area where the spot meter was pointed at. Plate 12, illustrates this: the clipping warning tells You that all the areas displayed in bright red have absolutely no detail and would print or display as pure white.

Image 10. All detail in the areas highlighted in blue has been obliterated and replaced by pure black.

Plate 11. All detail in the areas highlighted in blue has been obliterated and replaced by pure black.

Image 11. The opposite of Image 10: all areas that were previously black have now detail, but areas that were formerly white (and even many that were not) have been blasted to full white and no longer have any detail.

Plate 12. The opposite of Plate 11: all areas that were previously black have now detail, but areas that were formerly white (and even many that were not) have been blasted to full white and no longer have any detail.

All of this has one very important advantage: setting the brightest zone –even if it’s already pure white– as the mid-tone means the highlights will not be clipped, i.e. they shall all have detail. Likewise for the shadows: You cannot possibly have any pure black areas, for now even pure black has become 50% grey (just take a look at Plate 12: there are no bright blue clipping warnings). This, is the whole purpose of High Dynamic Range Imaging: no detail lost to clipping.

To recap, we can refer to the final image again, with its near-perfect rendition of tones and its nice bell-shaped histogram. Any blocked up shadows and/or fully white hightlights are intentional, rather than accidental. Yea, we might still want pure black and pure white areas, for the purpose of HDRI is not low-contrast, flat, lifeless photos, but giving the photographer the ability to set the black and white points himself.

Plate 12. The final image again…

Image 10-B. And its respective histogram, shewing an even distribution of tones.

Plate 12-B. …and its respective histogram, shewing an even distribution of tones.

And that is, Little One, how a hand-held light meter with a spot function is used to aid in the creation of high dynamic range images.

Ever Yours,

HVG.