Anne Cook made a welcome return to Reflex with another illuminating talk that was, let’s face it, fun. We can’t always locate the terms guest speaker and fun in the same sentence and whether we learn anything apart from the hardness of chairs that children are subjected to in modern day education (chairs, damn it, the cold hard floor was what we got, etc etc) is not necessarily a topic we need to travel to club to speculate on. The feedback from members was particularly effusive which is always a good sign.
Ann showed, amongst many other fine examples some images taken at carnival, which is particularly apposite to this time of year as we are coming to the end of the Somerset season. The last one on the Somerset circuit is at Wells this coming Friday (18th November 19:00 get there early). Having attended the Burnham-On-Sea round this year I can definitely affirm that it is something well worth making the time and effort, if you can get there.
So we are going to look at taking photo’s of street processions at night. The most obvious thing is that we are talking of photographing very high contrast scenes, that are moving, admittedly slowly, that can look distinctly two dimensional as brilliant point of light swim through a sea of black, which generally require high ISO’s to get reasonable shutter speeds and some interesting metering challenges.
There will be the large floats, some a part of a two three or four articulated trailers, but there will also be smaller floats, hand carts and individuals decked out in lights, there will be groups and there will be all sorts of other wheeled and ambulatory traffic. Variety is not short at these affairs and the lights are the main attraction.
The basic answers to these conundrums is to crop tight, centre weight or spot meter and expose to the right.
Cropping tight means we are likely to even out the overall exposure in a reasonable amount of the frames we capture. We will still get areas of high contrast and some of them will still have totally blown out areas of white and areas of inky blackness, but there are a significant number where we can limit the dynamic range some. This can put more emphasis on the mid-tones but whether that is a good or a bad thing is a matter of taste and of colour.
Metering these days is a lot smarter than it was five years ago. Leaving our camera on matrix / evaluative (basically our cameras approximation of the whole frame) isn’t necessarily a disaster, modern algorithms are pretty good at drawing conclusions that prove to be reasonably “accurate” (they are accurate but that has to be measured against our expectations) but we are talking an average balanced somewhere between the highest and the lowest value. Most cameras have the opportunity to switch between two other modes that increasingly ignore larger parts of the scene.
This gives us more information about a smaller part of the scene that we have judged to be more important. How do we know that? Well we’ve pointed our camera at it and taken notice of what effect this has had on our suggested exposure triangle settings. This is either because the scales in our view finder for shutter/aperture or under or over exposure are indicating a problem or the scene in our EVF is too dark or too light. What the camera has done is taken all the zones programmed into it and set these against where we are focused to give the reading a final weighting. There are other considerations but these change form manufacturer to manufacturer and can incur consideration such as colour, highlights and so on – Nikon use a comparison data base of shots of similar light/shade characteristics. This is common to Evaluative and centre weighted modes.
More so than fancy names for “Taking-the-whole-frame-into-consideration”, centre weighted does what it says on the tin. It takes the information in the centre of the frame into account either severely downgrading or completely ignoring the information on the outside of the frame. It doesn’t, by and large take notice of the focal point, instead just giving it a general, centre balanced reading. By restricting the area taken into consideration, prominence is given to the area more likely to include the crux of the image. Good when the boarders contain light that is in strong contrast to the main subject and might otherwise have undue influence on the final decision.
Spot takes the central are only, around 2% of the frame, around the focus point. Everything else is ignored. Photographing wildlife or Super-moons (or just the plain old ordinary moon which in the carnival context is a lot more apposite) benefit from this. Taking the moon in particular, the difference is a blown out image using the evaluative or centre weighted options because the light reflected from the moon is many, many times stronger than that which might be bouncing around the rest of the atmosphere. The argument in the wildlife example stems from a similar situation, say in the question of a bird in flight where it will be darker than the sky it is set in, or for the variations between fur and the cover it is in or breaking from. A characters face in comparative dark to the lights of a float with, potentially, tens of light bulbs throwing a shadow, yields a similar situation. Our meter might read the scene but in ways only remotely related to the way we see it.
Exposing to the right in itself gives no clear clue as to what exactly it is we are talking about. It relates to the histogram (posh name for graph) that we find somewhere in our camera’s menus and which we can turn on so that it appears in our camera’s live view and or view finder (depending on make and model). Pretty much every camera has the facility including some, but certainly not all, compacts and is also referred to as ETTR, even in polite company. The way the histogram is set out the shadows are on the left and the highlights are to the right. We don’t want spikes at the extremes of the scale because that tells us we are producing images that are heavily under or over exposed. Fine, but only to a degree, if we are shooting high key or low key portraits or similar.
What we want, in an ideal world is for there to be a fairly even shaped curve in between the two extremes of the histogram. That gives us an image that is evenly exposed. Of course this depends on the lighting of the scene and in extremes we know that shadows are more recoverable than burnt out highlights (if it is truly burnt out there is no useable information there to recover). Therefore we expose deliberately for the highlights in the picture and recover the shadows in post using whatever editing programme we have access to.
Of course it is a matter of personal preference, some people are dead set against the idea of ETTR, personally I find it useful, especially when someone else is controlling the lighting. In the studio we have more options and it entirely depends upon the sort of image we want to craft. The argument evolves around the number of tones on show in the image. If we are shooting a 12 bit RAW image we have 4096 tones available across the range. Each f-stop in our range (remember our meters render an average for the scene) accounts for twice as many tones as the previous one as our f-number increases. The brightest part of the range in the image accounts for half of the available tones, the next stop about half of that again and so on. The whole calculation takes place within the dynamic range of the camera sensor and that again is dependent upon make and model.
What happens is that by exposing to the right stretches out the details to the left and the more that happens the less smooth the transition between the dark tones, and we risk compromising the quality of the picture. That is why we need to look at taking pictures of the carnival floats in the context of the other two considerations we talked about above because the more we operate towards the centre of the cameras dynamic abilities the more leeway we have to accommodate the extremes of light and dark.
Go and enjoy.
N E X T M E E T I N G
Kev Spiers and Rich Price on their return to Iceland.