Thursday, 14 June 2012

Anti-aliasing: Anti or Pro

A very useful and gentle debunking of the current fad for cameras with no anti-aliasing filter (AAF) over their sensor appeared in Amateur Photographer (9 June 2012 issue). The fad for no AAF seems to have arisen because some people seem to think they are getting a more accurate representation without an AAF and that the sole purpose of an AAF is to prevent moiré pattern formation. All my reading is that such reasoning is based on a false premise. The article in AP lays out why the arguments for having no AAF are flawed. However, the demand for customers has been such that several manufacturers now make cameras without an AAF, the Nikon D800E for example.

The extra ‘sharpness’ (‘crunchy’ as the article described it) from sensors lacking an AAF is is fact an artefact of the sampling systems inherent in digital imaging. The crunchiness may be attractive or judged desirable by some but it is no more a reflection of reality than an over-sharpened photoshop job. We are not being cheated into accepting less of the detail that is there by the manufacturers who fit an AAF; an AAF ensures we see the best representation of the image that is presented to the sensor by the lens. AAFs are expensive to manufacture and no designer is going to add such a costly item  to produce what some misguided individuals interpret as a degraded or second-rate image. As the author of the article explains, the fall-off in contrast between features brought about by the necessary but imperfect AAFs, is best corrected during processing by using a sharpening procedure, ‘unsharp mask’ for example.

I have seen the argument advanced (particularly in US-based internet-forums) that because medium-format digital cameras do not use AAFs (and the marketing bumph for such cameras make a virtue of that point) then the lack of an AAF is a desirable feature - the ‘what’s good enough for Hasselblad is good enough for me’ type of argument. Unfortunately, these declaimers fail to realise that medium-format cameras lack an AAF because it is too expensive for manufacturers to include one made to the required size, and that the marketeers have had to make a (false) virtue out of a necessity. Incidentally, Hasselblad must be feeling pretty sore after the comparisons showing their very expensive cameras are outperformed by the Nikon D800.

Nevertheless, some camera manufacturers listen to their customers and, right or wrong, have given the choice of a camera with or without an anti-aliasing filter.

I hesitate to recommend the article by Professor Bob Newman in AP wholeheartedly because while the conclusions he presents are clear, as usual in these articles the technical explanations are abstruse and assume a level of knowledge not possessed by probably 99.9% of readers. That 99.9% includes me and I know something of the work and importance of Shannon and Nyquist. The explanations of the diagrams in this and other articles are very poor. The Editor of AP either needs to allow more space for important articles so that concepts and explanations can be spelt out, or he needs to employ a science writer skilled in taking the bare bones of an article and ensuring that the intended readership can actually follow the argument. Apart from that criticism and suggestion for improvement, AP is to be congratulated for examining such technical matters in digital photography and informing the reader of what marketing material reflects reality and what is hype.

So, would I buy the 800E (lacking an AAF) rather than the 800 as a replacement for my D700. No!