Note– all this is far from scientific, but based on my observations and those of others I have spoken to or read of on this. In fact, I’d argue that such observations and gut feelings are all the more reliable in a time when statistics or technical specs used to promote new models are making capabilities sound more revolutionary than they may actually be.
Probably you have heard enough from me on this issue, but this is time to say more! The freedom from the incremental changes in digital photography is a great motivation for going ‘back’ to film. Film is still being sold and used. Its existence is contemporary with digital, even though it is getting harder to find and process. This is a key point and one, which the marketers of ‘the latest and greatest’ digital cameras would prefer you ignored. Any camera working now is potentially in use, whether it be film or digital. The very fact that film cameras are rendered redundant by their successor’s superior sensors is one that should make you pause to think. A film camera, loading film, is as current as the latest digital model. The fact that its features, such as AF, are good enough for its genre can make it a complete and hopefully lasting tool; a ‘keeper’.
Then there is the other issue with digital cameras that is rarely discussed- the variable performance at base ISO. We all know the newest CMOS sensors, generally made by Sony, but with some good ones by other makers, too, have terrific dynamic range and high ISO properties and ever-greater numbers of mega-pixels. Yet what about use at base ISO, by someone who is happy with a low megapixel count, so long as the image quality is up there? Well, it is far from certain that in this case they are any better than the CCDs they are replacing. In fact, in terms of acuity, clarity and general sharpness, the CCDs, despite their relatively harsh rendering, are actually often superior.
For many years, Leica stuck with Kodak CCDs in it’s M9 (and now Monochrom MM) and all medium format digital bodies seem to be using CCD. Where still image quality is paramount CCD is still widely used. Now this may be partly due to a lack of investment in replacement sensors (which the new Leica M240 has), but it at least shows they can have exemplary performance.
In my own case, the kind of photos I got with my D70 may have lacked some finesse, but they were immediate, clear and sharp in a way my newer cameras seem to lack. Newer CMOS files are open and bland, suitable for all kinds of complex manipulation. They are, in a sense, the negatives to the CCD slides. So lets’ make it clear- from a certain point of view an older sensor can get better pictures. In one case in particular- a shot of a temple hallway seen from outside, in fading autumn light that illuminates just a tiny portion perfectly with the falling rays- my attempts to reproduce the same effect with newer cameras have failed. Part of it is no doubt timing and mood, but that special clarity- a sensor-level clarity like I saw so often on slide film (clarity, not necessarily the same thing as contrast), seems to be a CCD attribute. More recently I’ve seen this in D3000 and D200 files, which both used the same 10MP CCD sensor. Terrible low-light performance (poor ISO boosting ability even in native settings), yet a crystal-clear clarity and in short a different image rendering than we see with the newer CMOS sensors.
If true, this is a kind of revelation, which again, manufacturers and marketers would perhaps like to obscure, unless of course they return to CCD technology for any reason other than lower price. It opens the way to reuse or feel comfortable continuing to use older, seemingly redundant bodies, even if they are a lot less flexible in lower light. In any case, it is more evidence that the current obsession with CMOS Bayer-patterned sensors is likely to be superseded at some point. Not so much by CCDs, as by something like the Foveon sensor, which also has remarkable clarity and also incredible colour, if handled right on the software side. The problem tight now is the terribly underperforming bodies Sigma offers, most of them with fixed lenses and slow AF.
Once the multi-layer sensors take over, CMOS will be just as much in the dustbin of history. Which makes it sensible to be wary of over-investing in a technology that won’t just be improved over the next few years, but possibly superseded in a dramatic fashion- dramatic at least for those who notice such things and as many professional camera and sensor designers are amongst such people, I think it’s impossible for them to ignore such destined improvements, once, at least the immense challenges of timely readout of huge megapixel counts (on a small APS-C variant Sigma is already working with 46MP!), video capability and higher ISO boosts (400 ISO is already pushing it), such sensors will be mainstream. For now we can be thankful for the amazing flexibility of current sensors, with their massive dynamic range. Tremendous changes are possible, even if the ‘best final image’ can be elusive.
Well, some may counter this by saying it is very speculative and even entering the realm of science fiction. When would such sensors come, being as the market demands flexible low-light sensors and not absolute image quality, specialised, medium-format related uses aside. I’d partly agree, though I think it is important to resist the urge to always upgrade even to a reasonable improvement by staying aware that these are still technologies in flux and will only get better. Also, at base ISO, which good light shooting and especially tripod-based work allows, you may be losing something very special in return for this flexibility, which I really do think is under-reported, much the same way as poor gradients of digital are tolerated when more could be done (increasing bit-depth to 16bit as an option, maybe?) to fix this. The latest may be the greatest, but it’s not necessarily the best.
As a side note, there is a digital fix of sorts for this issue- the Lightroom Clarity slider, something I never felt I needed before my D5100 (the D300 has less dynamic range, so perhaps more native ‘definition’). It may be digital, but you pretty much can’t live without it. I even invested in an even better, more customisable tool, namely the new Topaz ‘Clarity’ plug-in. It was just released at precipitous timing for me, coinciding with my return from Mt. Fuji with a collection of beautiful, but seemingly quite hazy photos. I just don’t think digital handles haze very well, as it reads it as lack of definition rather than as a part of the atmosphere. I’ve already found it invaluable, if a little easy to overdo.
My personal workflow here is to apply it and then look at the results on my iPad- which with it’s glossy, IPS screen offers it’s own free clarity boost and to see if things still look good there. Seeing as I expect most of my photos to be viewed on tablets and other mobile devices, it makes sense to optimise their publishing for that. After all, I still have the original RAW and TIFF files to edit again if I’d like to, or to change settings to optimise for printing, for example. In fact, I’d personally advise anyone to optimise for iPad (or Android/Windows/Linux tablets) in this day and age. One wonderful thing about them is their consistency, as opposed to computer monitors, which vary tremendously depending on type (IPS or not) and coating (glossy or matt). Thankfully most tablets and smart phones are glossy, with special anodizing processes recently used to reduce glare, often by bonding the LCD to the actual glass, making replacement much more of a pain, but reducing their horrible reflectivity outdoors.