Exposure Latitude

In this test we look to see how tolerant of pushing exposure the a9's Raw files are. We've done this by exposing our scene with increasingly lower exposures, then pushed them back to the correct brightness using Adobe Camera Raw. Examining what happens in the shadows allows you to assess the exposure latitude (essentially the dynamic range) of the Raw files.

Because the changes in this test noise are primarily caused by shot noise and this is mainly determined by the amount of light the camera has had access to, the results are only directly comparable between cameras of the same sensor size. However, this will also be the case in real-world shooting if you're limited by what shutter speed you can keep steady, so this test gives you an idea of the amount of processing latitude different formats give.

Compared to its peers, the a9 fares better than the Nikon D5, but falls slightly behind the Canon 1D X II and more so compared to the a7R II. There's an odd, almost horizontal pattern to the noise in the deepest shadows, while in other, brighter areas of our scene this is less evident. In fact, in these areas, the a9 and 1D X II appear neck-to-neck, which Bill Claff corroborates in quantitative studies.

At smaller +1EV pushes, there's barely a difference between the cameras, but by +2EV you begin to notice a tiny bit of noise creeping into all cameras save for the a7R II. Progressively higher pushes of 4EV, 5EV, and 6EV really start to separate the cameras from one another, with the a7R II well ahead of the pack, with the D5 performing the worst.

The a9 still provides quite usable images at up to +3EV pushes, but +4EV or higher pushes from base ISO reveal a sort of patterned noise in shadows that may prove unacceptable to demanding users. Notably, this is an issue for all the high-end sports cameras presented here, likely due to constraints placed on sensors read out at high speeds. Our tests indicate the 1D X II fares the best amongst the current high-end sports cameras due its lack of patterned noise in the shadows. The (non-sports) a7R II, optimized for image quality, provides usable images even after a +6EV push at base ISO (at equivalent viewing size).

ISO Invariance

A camera with a very low noise floor is able to capture a large amount of dynamic range, since it adds very little noise to the detail captured in the shadow regions of the image. This has an interesting implication: it minimizes the need to amplify the sensor's signal in order to keep it above that noise floor (which is what ISO amplification conventionally does). This provides an alternate way of working in situations that would traditionally demand higher ISO settings.

Here we've done something that may seem counter-intuitive: we've used the same aperture and shutter speed at different ISO settings to see how much difference there is between shooting at a particular ISO setting (and using hardware amplification) vs. digitally correcting the brightness, later. This has the advantage that all the shots should exhibit the same shot noise and any differences must have been contributed by the camera's circuitry.

It's immediately obvious the a9 is not ISO-invariant (what is 'ISO-invariance'?). This means the camera is adding a fair amount of read noise that results in noisy shadows, limiting dynamic range at base ISO. That's why, for the same focal plane exposure, performing analog amplification by increasing ISO in-camera gets you a cleaner image than performing that amplification (or brightening) in post-processing.

It's not the typical performance we've come to expect from Sony cameras, but it's similar to most sports oriented cameras (save for the a99 II). Compared to its peers the 1D X II fares a bit better, the D5 somewhat worse. Meanwhile, the a7R II pulls well ahead of the pack. The a9's advantage over the D5 drops away at a push less than ISO 400 + 4EV, and with as little as ISO 800 +3EV all the speed demons are performing similarly. Shooting natively at ISO 6400, all cameras even out in performance.

Ultimately, like its sporting peers, the a9 sensor is optimized for speed and the high readout speeds of its sensor leads to higher read noise levels. This limits low ISO dynamic range by adding noise to the lowest signals the sensor captures.

Effect of Drive mode

There is little to no difference in base ISO dynamic range in different drive modes. So the good news is that the drop to 12-bit in continuous drive comes at no cost. The not-so-good news? 14-bit Raws aren't any better than the 12-bit ones, they're just bigger.

Click here to load the above as an interactive widget.

As we published earlier, it's interesting to note that there's no visual change in shadow noise levels as you switch drive modes: single mechanical to continuous electronic all look the same. This is particularly interesting because all Single drive modes, including fully electronic, support full 14-bit Raw. The Continuous drive modes, however, switch the image pipeline into a 12-bit* mode which, by definition, means files with no more than ~12 stops of dynamic range.

This indicates that even the 14-bit Raws have at most ~12 EV of dynamic range at the pixel level, placing base ISO dynamic range nearly a full stop behind the a7R II at similar (normalized) viewing size. Indeed, this is what Bill Claff found when we sent him a9 files for analysis, with little to no difference in dynamic range across drive modes.

Notably, the a7R II's 14-bit Raws have significantly less noise compared to its 12-bit Raws. Given the striking similarity between 14-bit and 12-bit Raws from the a9, we wonder if the analog-to-digital converter (ADC) is always being run in 12-bit mode regardless of drive mode. If the ADC truly is being driven in 14-bit mode, there are enough other sources of read noise to make sampling at anything above 12 bits unnecessary, according to our (and Bill's) results.


'Dual Gain' helps improve high ISO dynamic range

In our ISO-invariance widget above, you may have noticed that noise dramatically increases as you go from ISO 800 to ISO 400 (how's that for sounding completely back-to-front?). Below you'll see this more clearly as we 'zoom in' to this ISO range: shadow noise dramatically clears up as you go from an ISO 500 image (with 3.7 EV push) to an ISO 640 image (with 3.3 EV push):

Things clean up at ISO 640 (as with the a7R II) because of the sensor's 'dual gain' architecture, where the camera increases the conversion gain (effectively in-pixel amplification) during readout, helping overcome the camera's relatively high (for a Sony design) read noise.

Above ISO 640, the camera is fairly ISO-invariant, since it's overcome most of its downstream read noise, but there's still some benefit to increasing ISO to keep noise levels low if your scene demands it. Below ISO 640, the lower conversion gain means that you'll start to see more (read) noise if you push shadows as opposed to having them pushed in-camera by increasing ISO gain.

Take home

The good news is that those worried about the camera dropping to 12-bit readout in continuous shooting needn't worry: there's no decrease in quality, since a 12 bit file can contain all its dynamic range.** The not-so-good news is that this is because the a9 doesn't appear to have more around 12 EV pixel-level dynamic range to begin with, putting its normalized base ISO dynamic range at least a stop behind that of the a7R II.

A more relevant comparison might be to the a9's direct peers: it performs neck-and-neck with the Canon 1D X II (and perhaps a bit better at ISOs above 640 thanks to its dual-gain design), and a full stop better than the Nikon D5 at base ISO. Note though that extremely pushed shadows from the a9 exhibit a visually-distracting horizontal pattern in dark tones that the other cameras compared here don't.

By high ISO, general image quality catches up across all cameras, as downstream read noise becomes less important and the a9's dual gain architecture gives its performance a boost.

Ultimately, the a9's lower ISO dynamic range limits the exposure latitude of its Raws, so you'll have some limited ability to expose high contrast scenes for the highlights, then tonemap*** (raise) shadows in post. For most sports photographers this won't matter much, but those shooting high contrast scenes may require workarounds other cameras - such as the a7R II - don't.


Footnotes:

* We confirmed that continuous modes were 12-bit, while single shot modes were 14-bit, by comparing histograms of respective Raw files. The 14-bit single drive files have 14-bits of data compared to the 12-bit files (the histogram shows Raw values 1, 2, and 3 are vacant while the 14-bit files do have pixels with these values).

** Interestingly, this means there's little advantage to those large (47MB) uncompressed 14-bit Raw files, save for the lack of compression artifacts. In a perfect world, Sony would have offered a 12-bit Raw mode with an effectively lossless compression curve (without that second stage of localized compression that leads to edge artifacts) for smaller file sizes with no visual loss in quality.

*** There's a very specific reason I like to use the word 'tonemap' instead of 'raise the shadows'. We're forced to raise shadows of high contrast Raw files exposed for the highlights today because of the limited brightness of most current displays. Future displays capable of far higher brightnesses (perhaps even ten-fold) will need less shadow pushing, or tone-mapping, to make visible what you currently see as 'shadows' in such traditionally underexposed Raw files. For example, shadows you currently push +4 EV will likely be visible without any pushing at all on a 4,000 nit-capable display.