With the previous article on HDR, the zone system and dynamic range as background, I can now explain exactly what my B&W discovery was: it’s mostly to do with the highlights, but only in certain areas. And to make things more confusing, creating a natural-looking – perhaps even filmic image – required me to take processing steps that were both highly counterintuitive, but also go against everything else I’ve done and used successfully in the past. Read on if you dare; I can’t promise enlightenment, but I can certainly try for insight.
The story begins once again by understanding the way our eyes work: a scene is seldom so bright that we don’t see any detail – i.e. luminance variation, but we’re used to portions of our view being almost completely black and devoid of detail. This is a fairly nonlinear recording medium. B&W film is much the same: because of the nature of the emulsion chemistry and grain structure, it’s very difficult for a negative to go completely black – i.e. be overexposed with irrecoverably blown highlights. It requires a lot of light to completely turn all of the silver halide crystals black. In real terms, this means that most of the extremely large latitude of film resides in the highlight region; I find that in practice, changes in exposure make more difference to the shadows – zones IV and below – than to the highlights. You can see this behaviour quite clearly when you push a film with low reciprocity error like Fuji Acros – the shadows become more dense and block up a little, but the highlights remain much the same. A one-stop push results in losing a stop of dynamic range (i.e. tonal differentiation) in the shadows. This matches the luminance map our eyes see quite well, and thus appears ‘natural’.
On top of that, production of the final positive image – be it digital or optical wet printing – inadvertently introduces further nonlinearity: when you scan/ copy with a DSLR, the film is transmissive: light has to shine through it to reach the final positive medium. Since the film is never fully transparent and never fully opaque, it has the effect of appearing to never clip – especially with digital copies. In fact, you always have to move the lower range of the curve or levels slider after inversion to set your black point to be truly black. Practically, this means you have a lot of headroom in the highlights; because the chemical reaction is analog, there are nearly infinite shades of light grey that can be reproduced in the highlight zones. Overexposure or very bright areas therefore have gradual transitions, which contribute to the overall impression of there being detail/ dynamic range/ tonal discrimination (pick one, they’re all related) in the highest zones. The film effectively acts as a neat downsampling mechanism for luminance information.
Here’s what happens in the hybrid film/digital process: the top wedge represents the dynamic range of the scene; film is a bit less, and compresses the top and bottom ends in a neat manner. However, when the film is transmissively illuminated/ scanned, then it appears to be the third wedge; i.e. within the dynamic range of the digital capture device, which is the final wedge. We don’t have clipping even though the original scene exceeded the dynamic range of direct digital capture because the film has done the condensing of upper/ lower luminance ranges for us.
On the other hand, a direct digital capture of the scene is constrained by sensor physics: if there isn’t enough light to trigger a signal voltage, then it’s recorded as black. If there’s too much, then you have overexposure and pure white. That clipping point on either end is very abrupt and discrete because digital is binary: there’s either luminance information, or there isn’t. It’s this abrupt clipping that makes digital images – especially B&W – look, well, digital.* You can of course avoid this by carefully ensuring that the input scene has less dynamic range than the capture latitude of your sensor; however, this isn’t always possible, and even if the majority of the scene lies within the tonal range of the sensor, you will almost always have single specular highlight points that fall outside of it.
*It isn’t so much of an issue with colour, because all three channels have to saturate before you go to pure white. If only one channel saturates, there are still the other two remaining to provide colour information and tonal discrimination.
Raw digital files are very linear – especially from CMOS cameras – which means that you have to increase the contrast in post processing to make the scene appear natural and to preserve relative differences in luminance between adjacent areas in the final image. It is possible to do this locally – that’s what dodging and burning accomplishes – but if you push it too far, you land up with areas in the image that look unnatural because their relative contrast is too high or too low compared with the remainder of the frame.
It seems obvious that to avoid this, we could just pull the top and bottom ends of the curve so they don’t clip fully to black or white; however, this results in very flat, unsatisfying looking images that somehow lack punch, and worse still, look odd because our eyes need an anchor in zones 0 and X so we can calibrate the relative luminance of the overall scene. And you of course still want to preserve the gradual transition to over/underexposure, whilst maintaining tonal separation in those zones…not so easy, is it?
I’ve always said that one should never burn highlights or dodge shadows: this is because it can very quickly result in overlapping tones, i.e. where areas that are zone X in reality become the same luminance as zones IX or VII; our brains are very quick to recognise this and interpret it as looking unnatural or odd. However, through forensic examination of both film and ‘scanned’ negatives, I think I’ve found a solution: one that does in fact require burning highlights and dodging shadows, albeit very lightly. However, to avoid tonal overlaps, you only do it when there are two or more zones of difference between the specular highlight and the adjacent areas (or the same for shadows).
Effectively, what this does is lower the contrast in that area of that image only, so it doesn’t appear too ‘sharp’ or ‘aggressive’, whilst still preserving the perceived contrast and tonal separation within that localised area. There are no more hard and fast rules than that; to decide where to apply it and exactly how strongly requires experience and practice. However: do not do this for colour images, it still results in unnatural-looking highlights and washed out shadows. This technique is especially effective when you have areas with high frequency and high contrast detail – e.g. blades of grass in direct sun – because it preserves the luminance detail in those areas, but avoids them feeling harsh and halftoned or binary black/white. Note that it’s also very important to take care when sharpening, too: sharpening increases contrast at edges; if you increase it too much, especially with high frequency detail, it can cause the impression of clipping when you have abrupt transitions between high and low exposure zones in a physically small space.
Perhaps the best tests of effectiveness are first blind discrimination, then print: the images in this article are a mix of scanned film, which still looks distinctively ‘smooth’ despite passing through the usual digital workflow, and direct digital captures. I have done a number of test prints, and there’s a definite difference visible: direct digital capture images prepared by this workflow are much smoother and more natural looking, and not easy to distinguish from scanned film. What’s more interesting is that the film images in this article were shot under what I think of as moderately contrasty conditions for the tropics – late evening and partially cloudy – the digital images were shot at high noon with a clear sky, and extremely high contrast. Let me finish with one clarification, though: the aim here is not to make images that could pass as film captures: it’s to take one specifically desired property of film capture – i.e. smoothness of tonal transitions at the extreme shadow/highlight zones – and try to achieve the same result with a direct digital capture, in order to get a better, richer final image that blends the advantages of film (tonality) and digital (workflow speed, resolution, shooting envelope/ high ISO performance etc.). MT
There will be updated Photoshop Workflow videos available in the near future, including one specifically covering black and white conversions and preparation for print in much more detail with examples – I have made a lot of improvements and refinements to the process in the two years since the first video…
Places left for 2014 Making Outstanding Images Workshops: Havana and London – click here for more information and to book!
Visit the Teaching Store to up your photographic game – including workshop and Photoshop Workflow videos and the customized Email School of Photography; or go mobile with the Photography Compendium for iPad. You can also get your gear from B&H and Amazon. Prices are the same as normal, however a small portion of your purchase value is referred back to me. Thanks!
Images and content copyright Ming Thein | mingthein.com 2012 onwards. All rights reserved