JPEG Tone Curves / Dynamic Range

Our Dynamic Range measurement system involves shooting a calibrated Stouffer Step Wedge (13 stops total range) which is backlit using a daylight balanced lamp (98 CRI). A single shot of this produces a gray scale wedge from the camera's clipped white point down to black (example below). Each step of the scale is equivalent to 1/3 EV (a third of a stop), we select one step as 'middle gray' (defined as 50% luminance) and measure outwards to define the dynamic range. Hence there are 'two sides' to our results, the amount of shadow range (below middle gray) and the amount of highlight range (above middle gray).

To most people highlight range is the first thing they think about when talking about dynamic range, that is the amount of highlight detail above middle gray the camera can capture before it clips to white. Shadow range is more complicated; in our test the line on the graph stops as soon as the luminance value drops below our defined 'black point' (about 2% luminance) or the signal-to-noise ratio drops below a predefined value (where shadow detail would be swamped by noise), whichever comes first.

Note: this page features our new interactive dynamic range comparison widget. The wedges below the graph are created by our measurement system from the values read from the step wedge, the red lines indicate approximate shadow and highlight range (the dotted line indicating middle gray).

Cameras Compared

This chart shows the Nikon D4S' default dynamic range performance in JPEG against a range of competitors - the Nikon D4, the Canon 5D Mark III, and the Sony A7. Here you can get an idea of the highlight and shadow range above and below middle grey in a JPEG.

The Nikon D4S' default tone curve is remarkably similar to that of the Canon 5D Mark III, although in our tests it sacrifices 1/3 EV highlight detail for 1/3 EV extra shadow detail (mapped above 2% luminance in the JPEG file). In real-world usage, 1/3 EV differences are unlikely to be seen (and fall within the margin of error of our tests).

Compared to the Nikon D4, the D4S offers a whole stop more shadow detail, while not sacrificing any highlight detail. The D4S' tone curve offers higher contrast in the brighter midtones than its predecessor, along with a more subtle roll-off from near-white to completely 'clipped' regions. Compared to the Sony A7, the D4S also offers a whole stop more shadow detail, but clips to white 2/3 EV earlier.

In principle, despite the differences in allocation of highlight range vs, shadow range in these cameras, the range of usable tones in the JPEG - those above 2% luminance and below clipping - is quite similar between the Nikon D4S, the Canon 5D Mark III, and the Sony A7 (the Nikon D4, with its whole stop lower dynamic range, is a bit of an outlier here). Therefore, you could theoretically underexpose the D4S by 2/3 EV to capture the 2/3 EV of highlight detail in the extended highlight headroom of the Sony A7 - that is, if you're willing to process either your Raw or JPEG images to bring them back to the correct brightness. However, midtones and shadows would generally be rendered darker if you were to do so, and you'd be better off using Nikon's ADL (Active D-Lighting) modes if your intent is to capture extended highlight detail in your JPEGs (see below).

ISO Expansion

The Nikon D4S allows you to lower the ISO to L0.3, L0.7, and L1.0 (or ISO 80, 64, and 50 equivalents, respectively) for those who want to decrease sensitivity in bright light, to use wider apertures or longer shutter speeds. Unfortunately, since the underlying sensitivity of the sensor is not decreased in these modes (which are more digital 'pulls'), these modes will overexpose more readily - meaning highlights will have a tendency to clip. For example, at ISO 50 (L1.0), highlights abruptly clip to white around 1 stop earlier than ISO 100.

Active D-Lighting

The Nikon D4S makes use of the Active D-Lighting system that Nikon has used in its DSLRs since the D300. Depending on the setting applied, this will reduce the exposure to increase the amount of highlight range captured. The camera then analyses the image and selectively brightens it, so that the final output represents the expected overall brightness, but with more tonal detail in the highlight regions, balanced shadow regions, and a retention of local contrast. Local contrast is retained by performing local adjustments as opposed to blindly applying a global tone curve to pixels simply based on pixel intensity.

ADL Low doesn't yield any additional highlight detail (as it doesn't change the exposure), but reveals slightly more detail in shadows by gently pushing dark tones. You'll note the black clipping point pushed down at least 1/3 EV to a darker patch on our wedge, which remains the case for all the ADL modes. All ADL modes from Normal upwards actually decrease the exposure to capture additional highlight information. To be more precise, ADL Normal, High, Extra High, and Extra High 2 progressively decrease exposure to regain 1/3, 2/3, 1, and 4/3 EV extra highlight range above middle grey. This data is summarized in the table below.

ADL Mode
Highlight Range Gained
Shadow Range Gained
Low 0 EV 1/3 EV
Normal 1/3 EV 1/3 EV
High 2/3 EV 1/3 EV
Extra High 1 EV 1/3 EV
Extra High 2 4/3 EV ~3 EV

Since ADL modes are likely to be used in scenes with high dynamic range, we'd like to make a small point about how recoverable shadow details are with ADL modes engaged. Of particular interest is our observation that as soon as ADL is engaged, no pixels are clipped to absolute black (RGB = 0,0,0) - which is not the case when ADL is turned off. Furthermore, ADL Extra High 2 digs down to a Raw value of 1 and maps it to 2% luminance in the JPEG (meaning all the tonal range of the Raw file is in the JPEG).

You'll note the staggering increase in shadow range gained by engaging Extra High 2 (though that says nothing about whether or not there's any usable detail down there that isn't swamped by either sensor or shot noise). The implications of this are: if you wish to pull shadows out of your JPEG, you'll fare better with ADL on and, in particular, set to ADL Extra High 2 (just make sure the exposure reduction in the selected ADL mode doesn't drop your exposure down below what is necessary to preserve highlights!). ADL's Extra High modes will ensure that even the deepest detail recorded on your sensor are still represented in your JPEG and not, instead, clipped to black or below 2% luminance (RGB=5,5,5).

All that said, for maximum latitude in shadow recovery, we'd recommend you shoot Raw.

Active D-Lighting: In the Real World

We shot a scene (below) with a fair mix of both shadows and highlights. Here evaluative metering - with the focus point set to a subject in the shadows - chose an exposure that blew highlights when ADL was set to Off. You can see that with progressively higher ADL modes, more and more highlight detail is maintained (pay special attention to the bright upper left corner, as well as the reflection of the sky just above the door with the 'Dawson' sign). In this example, ADL does a nice job of retaining local contrast and shadow brightness while recovering some highlight detail in the higher ADL modes.

ADL Normal
ADL High
ADL Extra High
ADL Extra High 2
Show Clipped
Show Clipped
Show Clipped
Show Clipped
Show Clipped
Show Clipped
For easier visualization of the effects of ADL on the highlights in our scene above, hover over 'Show Clipped' to see which pixels clipped to white (indicated in red) in the various ADL modes. Fewer and fewer pixels are clipped to white as the ADL mode is increased. Note that the effectiveness of ADL will be highly scene-dependent. In this scene, for example, highlight exposure isn't sufficiently curbed until ADL Extra High 2, where the shot received 1 and 2/3 EV less exposure compared to ADL Off. This is still not enough to regain all the highlight detail in this particular scene, which would have likely benefitted from an even lower exposure* combined with a shadow push (ideally in Raw). For the curious of mind, the table below shows the actual exposure adjustments the various ADL modes perform and, in particular, why the above scene was hard to tame in the lower ADL modes that use rather conservative exposure adjustments.
ADL Mode
Exposure Adjustment
Low 0 EV
Normal -1/3 EV
High -2/3 EV
Extra High -1 EV
Extra High 2 -5/3 EV

We also note that ADL Auto chose ADL Normal in the above, rather high dynamic range scene. Therefore we feel ADL Auto to be a bit conservative and suggest that the photographer dial in an appropriate ADL mode depending on the dynamic range of the scene when using this feature.

*If anyone is left wondering why higher ADL doesn't offer more aggressive modes (that reduce exposure even further), you're probably not alone. Here it's important to keep in mind that the higher the reduction in exposure, the more underexposed shadows have to be raised to maintain proper brightness. Raising shadows while maintaining local contrast is tricky, and imposes a real limit on how aggressive automated algorithms can get (and that says nothing about the noise cost in shadows from the reduced exposure). Although we'd love to see more aggressive modes available for scenes that require it, for now we'd recommend shooting high dynamic range scenes in Raw. This way you can expose for highlights and then push shadows - which might have otherwise been clipped to black in the JPEG conversion - later in your favorite Raw converter.