I just encountered a line at Rob Galbraith, in a news item about the upcoming-in-2012 Canon 1D X: "...a standard ISO range of 100–51,200 (which can be expanded to as high as 204,800...."
It made me realize that I habitually let go with a silent internal existential guffaw every time I read one of those outlandishly high ISO numbers.
It's not ridiculous to use, or learn, the meaning of ISO 400, 640, and 1000 and so on. You can memorize such numbers easily and learn to translate them into stops easily.
(Fairly easily, for the innumerate amongst us. Including moi.)
But who mentally works with fluency and fluidity with numbers like "51,200" and "204,800"?
The ISO standard actually consists of both the arithmetic and logarithmic numbers; that is, it's not "ISO 200," it's actually "ISO 200/24°." But when talking about it we ignore the 24 part (which came from the old DIN standard) and just use the 200 part (which came from the old ASA standard).
Seems to me the conceptual appeal of the old ASA scale is blasted to smithereens by the new, or recent, necessity of using such extended values. The jump from 51,200 to 204,800 (in our example) is also perceptually wopplejawed—it seems like a whole great whacking lot; it's actually only two stops.
Time to switch conventional usage from the now-standard arithmetic scale back to a logarithmic one?
Note here that there are no official ISO speeds over 10000. They're extrapolations that fall outside the official standard; so far they derive only from manufacturer specifications. Changes have to be made to the standard one way or the other to talk about speeds like the ones above.
Here's an example section of the old DIN scale and the ASA equivalents:
The difference between 51,200 and 204,800 expressed in the DIN-derived part of a postulated ISO standard would be from 48 to 54. That seems more proportionate to me than using huge numbers in the tens and hundreds of thousands—especially since we're now just identifying film speeds and seldom if ever converting the numbers to stops in our heads anyway. Who needs to translate 80000 to 100000 in stops, these days?
Not a super-important point either way. But I'm just sayin'.
Note: Links in this post may be to our affiliates; sales through affiliate links may benefit this site. More...
Original contents copyright 2011 by Michael C. Johnston and/or the bylined author. All Rights Reserved.
Featured Comment by psu: "My vote would be to get rid of the ISO number altogether. The ISO/ASA ratings made sense for film where you want a single calibrated number to base your workflow around. Except that even in the film days it was at best a guideline, especially with black and white.
"In digital capture ISO is just a shorthand for setting the gain in the sensor and adjusting the internal image processing to compensate. This value is expressed in terms of the old ISO value, but it's not clear to me why. Generally I shoot in two modes: Either...
1. I want the camera to use the lowest setting it can to hold my desired aperture/shutter speed combination, or
2. I want the camera to use its base setting to get the highest possible image quality.
"It might be useful, in addition to mode (2) to have fixed "low, medium, and high" speed settings. But in general I don't see any reason for me to be hitting a button called ISO and twirling through 15 or 20 settings 1/3rd of a stop apart. The only time I ever set ISO manually now is either to obtain mode (2) above or to get a fixed ISO when I'm using flash because the Nikon auto-ISO system is too stupid to do the right thing with the flash.
"Since buying the D700 I never look at the ISO setting in the EXIF at all except to marvel at how good ISO 1600 looks."
Mike replies: That would make altogether too much sense. A future frontier in camera design will be (I hope) the intelligent simplification of camera design centering around, as my friend Nick puts is, functionality rather than features; Apple has been very good at this in the computer/music player/smartphone/personal notepad realm, but few cameramakers so far even bring it into their thinking. "Featurism" intrudes into, and interrupts, functionality to a significant degree with most current cameras, especially for people who aren't photographers or photography mavens—even cameras meant to be simple aren't. In fact it sometimes seems to me that recently, what a sophisticated knowledge of camera-tech is most good for is for designing the setup of a digital camera and creating standard procedures (some of which are workarounds) so that the device becomes effectively functional instead of just confusing and frustrating.
There's no reason at all for amateur cameras not to have a 1-2-3 scheme for setting sensitivity: 1 being the setting for optimal quality, 3 being the setting for acceptable quality in the lowest light, and 2 being the best balance between the two. The manual could tell you what the measured values of amplification are; the camera user doesn't absolutely need to know. Pro cameras could be a little more sophisticated than that, but, as you say, they could still be usefully a lot simpler than they are, and, as you say, referenced to signal amplification rather than to an analogy of inherent light sensitivity.
Featured Comment by Andreas Weber: "At least make it 'ISO 50000, 100000, 200000;' there's a reason why nobody talks about 1/128th or 1/256th sec., either! Please, in the name of all engineers...."
Featured Comment by Eduardo Cervantes: "The beauty of the American standard is that doubling the number doubles the sensitivity. That's why it remains relatively easy to apply even at such ridiculous ASA's. The European standard just adds three values for each doubling in sensitivity. I don't know if that is easier to apply when shooting.
"Many years ago, after my country's currency suffered numerous devaluations against others, buying a camera was in the millions. A meal in the thousands. Digital calculators were of no use anymore as they didn't have enough space in the display for so many digits. After things settled down, the government took three zeros out of the nominations and things went back to 'normal.' I wonder if the same thing could be apply to the American standard. Two zeros out; 100 ASA becomes 1 ASA. 6400 becomes 64. Lower than 1 ASA become fractions. The same for some others like 640. It becomes 6.4. Designers have done it in tachometers for years and it worked. Just commenting."
Mike replies: You're on to something. I'd say we could just jettison the intermediate stops, because, really, in the digital age, thirds of a stops are like Andreas Weber's 1/256th sec.—needlessly fine distinctions.
Why not just round the numbers like the shutter speeds are rounded, then use "k" once past 1000? So the progression would go 1, 2, 4, 8, 15, 30, 60, 125, 250, 500, 1k, 2k, 4k, 8k, 15k, 30k, 60k, 125k, 250k, 500k. That takes us all the way to ISO52428800. Seems high enough. And writing "51,200 and 204,800" becomes "500 and 2k." Much simpler.
Of course, we realize that the further our inventions get from standard practice, the less likely they are to ever be implemented.