WEDDING SEASON SALE! 30% Off Training Systems!

Your content will be up shortly. Please allow up to 5 seconds
Gear & Apps

No Cartesian Duality Between Camera Software & Hardware | How Many Stops Have Cameras Improved In A Decade?

By Kishore Sawh on November 27th 2016

We’re rounding out the end of 2016, and that means it’s a time for a bit of reflection, and not on trivial, yawn-inducing, boring matters like social pro-or-digression, or on our behavioral growth, but on the things that truly matter, like camera development. I kid; that’s a bit tongue in cheek but it is the weekend of Black Friday, Cyber Monday, and a season loosely associated with, well, presents. So, as many of us will undoubtedly be looking into all that the camera tech world has released this year in order to make judgements and even purchases, it’s interesting to consider just how far camera tech has come in the last decade. 10 years is a long time, and with technology racing at the speed of thought, cameras of today should be tens of stops better that so many years ago, given Moore’s Law. Shouldn’t they? Well….

Sort of, but around 1.5 stops of sensor improvement is more precise.

That may be less than you’d have thought given the number of cameras released each year and the number of times you’ve felt it imperative to upgrade, but according to DxO, that’s the magic number. At the Automotice Sensor And Perception Conference (AutoSens) a presentation was given by DxO tited, ‘Dealing with the Complexities of Camera ISP Tuning’ this was revealed.


Those of a meticulous lab experiment persuasion may sit there thinking 1.5 stops is a lot, but for the rest of us mere mortals it doesn’t have the ring to it we’d imagine. I remember shooting the final F-14 Tomcat flights in 2006 with a Nikon D70 and I can tell you just shooting with a D3400 or D500 would’ve provided an exponentially better or easier experience, but the rest of what is shown on the slides helps to paint a clearer picture. While sensor tech seems to have improved by 1.5 stops, digital processing gain has been a significant 3 to 4, ergo what we have is a total improvement of 4.5 to 5.5 stops within a decade. If we use ISO to measure, a base ISO or 100 with 5.5 stops improvement would be like 100 up to and somewhat surpassing [100,200,400,800,1600,3200…] ISO 3200. That’s some perspective for you.


If you pay attention to the slides and do a little digging, you’ll see that there’s more to a good image or measure of a camera’s ability that that of the sensor; that the processor is critical, as is the physical build and how all the parts come together. It’s further proof that a Cartesian duality doesn’t exist in cameras really, that there is an inextricable link between the software and hardware. But 10 years and this is where we are, in another ten where will we be…


Source: Jeremy Gray of Imaging Resource

This site contains affiliate links to products. We may receive a commission for purchases made through these links, however, this does not impact accuracy or integrity of our content.

A photographer and writer based in Miami, he can often be found at dog parks, and airports in London and Toronto. He is also a tremendous fan of flossing and the happiest guy around when the company’s good.

Q&A Discussions

Please or register to post a comment.

  1. Ralph Hightower

    Okay, here’s my take on this topic. Back in the film days, the two majors, Canon and Nikon, had life cycles of 10 years between major revisions for their pro cameras.
    Now, with digital, it is an “Arms Race” to see who can out-megapixel and out-feature the other. A new revision comes out every 2-3 years with incremental changes (nothing spectacular).

    | |
  2. Luis Luna

    So, if we shot in raw we only have that 1.5 stops of improvement and not the 3-4 stops achieved by software

    | |
    • Stan Rogers

      No, not at all. Most of the signal processing chain they’re talking about happens between the photodiodes (the sensor proper) and the “raw” output. (You don’t want *real* raw output from the sensor. The individual sensel sensitivities are all over the map, with some being much “hotter” than others, both due to differences at the silicon level and differences in the colour filter array, etc., at each sensel location.) That includes lower-noise read circuitry, better amplifiers, more accurate analog-to-digital converters, better tracking and monitoring of sensel performance, and so on. It’s the “raw” file that sees the 3-4 stop improvement, not just the JPEG.

      | |
  3. Lee Hawkins

    Great article, thanks for sharing this! So I take this to mean that a new body with a two year old sensor will get much more because it has two years of improved processing?

    | |