Why Dynamic Range is NOT Tonal Range

Why Dynamic Range is NOT Tonal Range

Now, it could be…but it’s not

It’s like a square is a rectangle but a rectangle is not a square…so let’s explore this

I recently was reading an article explaining dynamic range, in it, the author went on to explain when a camera has a limited dynamic range it will only show shades of gray not black and white. And I thought, no, that’s limited tonal range, not dynamic range.

Most everything we use in photography has a Full Tonal Range when lit with the same constant light source

  • Our Eyes; can see the full tonal range from Black to White
  • A High End Camera; can reproduce the full tonal range from Black to White
  • A Low End Consumer Camera; can reproduce the full tonal range from Black to White
  • Most decent LCD Monitors: Can produce a full range of tones from Black to White
  • Most Better Photo Printers: Can produce a NEAR full range of tones from black to white (Limited by Paper white {DMin} and Black Ink (DMax} )

It’s really not hard to  capture and reproduce everything from Black to White when under the same light. The number of tones in between can sometimes be limited by Dynamic range and also can be limited by bit depth but really reproducing black to white in a gradient is not hard.

So then when is Dynamic Range different than Tonal Range? When objects are not lit by the same light. Think of one of the most extreme example. A Cave lit with a 60 watt bulb and  a sunny day outside the cave. (As I showed in this article). This example is really beyond the dynamic range of the human eye in one look. Think of it, you’re in a cave or a movie theatre and you walk out to the sunlight and you have to blink and wait for your eyes to adjust and vice versa going into a dark space. So that is an Extreme Dynamic Range because the light source is not the same, yet we need or hope to see the full tonal range.

But besides this extreme example are many other scenarios of dynamic ranges everywhere in between. Some can be seen by our eyes but  not captured by our cameras, some captured by or camera and not reproducible on screen and some on screen cannot be faithfully printed. All because the scene dynamic range may exceed the device.

An example of Low Dynamic Range maybe a Sunny day, while it may appear contrasty, the dynamics are quite low because everything is lit by the same light and even shadows are only about 3 stops down from everything else

Bear something in mind; Just because a device (even our eyes) are capable of a wide dynamic range, does not mean that every scene is of a wide dynamic range, most are very ordinary. And even tonal range may not be that wide, think of a scene with just a blue sky and green grass.

Maybe it’s best to show this graphically because then we can also show what happens when the scene dynamics exceed the device…or even our eyes ( EV= Exposure Value, you can also think of it in Stops)

Dynamic Range

This  shows the tonal range across the dynamic range.

Note that in all, Tonal range runs from 0 (pure black) to 255 (pure white) despite the difference in dynamic range. The Print maybe somewhat limited on both ends of tonal range however

Next lets look at what happens when we shoot a High Dynamic Range Scene (20EV) with a Typical Camera. We can see that the camera still captures Black to White however it has truncated what was actually in the scene and naturally compresses the scene but not in a way that the full range of tones was captured in the camera or better put, it did not capture them as they actually appeared. If we look in this scenario, a highlight of 207 was recorded or “Mapped” as 255 white, anything above 207 was actually cut off or clipped. a shadow tone of 41 was recorded as 0 Black and anything below 41 was clipped. So yes the image has from black to white but did not record what actually was black and white in the scene and tones above or below a certain point.

Clipped

Because there is a compression of tones, often the Histogram of an image beyond the dynamic range of a device will look like this:

Wide Dynamic Range Histogram

Tones Bunched up on each end and clipped and a deep valley in the center mid-tones, where usually the majority of tones should be in a  normal scene/histogram

Ways to handle the extra dynamic range

Now as photographers we have some choices on how we want to handle this disparity.

The first thing we could do is to underexpose the image (Expose to the left) so that 255 white maps as 255 white in the camera.So we have the whites captured perfectly but the problem is there is a bit of tonal compression and the biggest problem is that a 105 Midtone now maps to 0 black.(These numbers are not absolute just picked for example) With software we can recover some of that by expanding shadows. However the problem with that is that the noise floor of the camera lies in the black, so when we raise the black/shadows we now make the inherent noise visible

Expose to the left

The other option is to overexpose the image or Expose to the Right. This will capture the Black point perfectly and map 0 to 0, but again we have a problem, mid-tone 147 has mapped to white   255 and there will be some tonal compression so we won’t have all the detail in highlight like we might in a properly exposed image or one of lower dynamic range. Think about clouds and seeing details in clouds and how that might be lost.

Expose to the right

*Please note that this is different than the typical “Expose to the Right” scenario many photographers use on an image that is within the dynamic range of the camera. In that case you push the exposure on capture (without blowout) and then pull the exposure in post to darken the shadows and make the noise less visible thus expanding the white and black point which HAVE been captured.

So either of the above two  scenarios are possible but have their limitation. Often the photographer has to choose what is more important to the image; highlight detail or shadow detail.

So that brings us to the final scenario where we shoot multiple exposures (typically 3 – 7) that change exposures between each exposure and therefore cover the entire dynamic range of the scene.

In this image we have 3 Overlapping Exposures, one shot at Average (0) one shot at +2EV and one shot at -2EV. The spacing between images will change depending on the total measured dynamic range (See this article) and the number of exposures used to cover the entire dynamic range of the scene.

HDR Exposure

Now you may say: Well the two images overlap, isn’t that enough? It may well be, but generally, since in every exposure there still is tonal compression, we may not get a smoothness to the gradient of black to white, so the 3rd exposure helps in  that matter and in general the more exposures (to a point) the smoother the gradation will be.

We have now covered and captured the entire dynamic range of the scene  20EV

The problem

We now still have a problem. We are back to our original Problem. We have a wide dynamic range and now must again translate that to a device (monitor) or a print that doesn’t HAVE that dynamic range. Since we have it all captured we now turn to software to do that translation. And just like naturally a tone will “Map” to another tone than it actually is, now we will take control over that “Tone-Mapping” and compress the wide range of tones into a lower dynamic range. We will make sure what is white is white, what is black is black and compress all that is in between to fit between our black and white points smoothly.

By Tone-Mapping we are making a Lower Dynamic range image from the High.

We can use various software from strictly HDR programs to Lightroom and Photoshop to carryout that Tone Mapping, Setting White and Black Points and then using HighlIghts and Shadows to adjust what is between and hopefully place them correctly. From that we get a usable and “Viewable” Master image.

From that master image we can go one step further using Soft Proofing to see the “Gamut” (Tonal Range) of our print on-screen and make further adjustments to make sure that all the tones fit within the confines of that print and make a Print Proof File.

Tone mapping To Compress DR

Tone mapping To Compress DR

That’s what it all looks like in graphs and theories but lets see what it looks like with real life images.

Real Life

The images I shot are of a very tough scene with high dynamic range. It is half of an outdoor exposure including a bright white fence in full sunlight and then the other half is the interior of the house lit only with incident window light that has some elements of pure black.

The scene was shot with a camera that has  a dynamic range of about 12.6EV. The entire scene measured from mid-tone point to mid tone point is about 17EV, measured from black point to white point is about 15.5EV (or 15 1/2 stops if you prefer)

The first image is the averaged exposure (as measured with a Gossen hand held incident light meter) f/8 1/20 ISO 100 I’ve included the histogram (notice that tell tale histogram)

Averaged Exposure (0)

Averaged Exposure (0)

We can see that a good amount of the highlight tones have been mapped up into the white range loosing detail. The blue sky, the shadow white of the trellis and parts of the other house have ventured up to the 240-255 range and not where they belong. On the other end we can see that the lower shadows have all gone to black and all the mid tones are very subdued which is confirmed by the histogram.

The next image is  a -2EV image (f/8 1/80 ISO 100) as you can see the white point and highlights have all come into line and are proper,however this has now plunged practically all of the mid tones – lit by different light – down into the shadow to black region. This image gets the mid-tone to highlight (lit by daylight) to white correct

Exposed to the left -2EV

Exposed to the left -2EV

Next we have the +2EV image.( f/8 1/5 ISO 100), in this exposure we can see that the shadow area has perfect detail but for all practical purposes there is no highlight detail (outdoors) at all and everything has mapped to white (Blowout)

Expose to the Right +2EV

Expose to the Right +2EV

This final image is the processed 3 exposures, tone mapped to compress and contain the entire dynamic range and the resulting full tonal range. You can see from the histogram that the image runs from pure black to pure white but now we also see the the energy in all the mid-tones is now correct and without that huge valley. This image best represents what my eyes saw at the time.

HDR Tone Mapped

HDR Tone Mapped

Please bear in mind as I talked about before. The interior of the house has a full tonal range but it is lit with filtered window light. The exterior has a full tonal range but is lit with a different light source, full daylight there is about a 7 Stop difference in Luminance between the two light sources. Those two tonal ranges give us our total dynamic range.

From this we can make a better statement about what Dynamic Range is. It’s often stated as the difference between the Brightest tone and the darkest tone. But that’s only partially true. It is the difference between the brightest tone lit with the brightest light, and the darkest tone lit with the lowest light

To show you what I mean, Here are 3 images of Black and White printed on some Luster Paper (Tonal Range limited)

This first one is shot under full daylight f/10 1/125 ISO 100 It is Black to White

Black to White Daylight

Black to White Daylight

The next one was shot in full shade which is 3 stops less than the full daylight and was shot f/9 1/30 ISO 100

Black to White Shade

Black to White Shade

Both of them are full tonal range (limited by the paper) and look pretty much the same because they were both lit by a single source even if the source was different and of different magnitude

Black and White with Two Light Source

Black and White with Two Light Source

The third image, was shot both in the sun and in the shade (lower half) so we now have the two same tones but lit by different light and how that affects the  dynamic range of the image and even the tonal range because the shaded white is 3 stops below the daylight lit white. But  they both are as they are supposed to appear when you are shooting under multiple light sources. The shaded white should not be 255 white.

Which brings me to the other part of my article, which is knowing how to place tones.  Typically White needs to be placed +3EV, Black needs to be placed -3EV.  not at 0 (when measured reflectively). I just think that is something that needs to be understood as I often see people exposing things oddly…such as “Getting detail in White and Black” There isn’t supposed to be. It’s the stuff in between that’s supposed to have detail.

One final note

Going back to out test image of the indorrs/outdoors. One thing needs to be noted about some of the newer cameras that have Dynamic ranges above 12EV. They actually capture more DR than is displayable. so the middle exposure we see is not all the information that is available. What we see in that 0EV exposure is the preview that is made of that exposure by the software taking into account the 8 bit monitor profile and rendered and mapped. But  we can extract more information and Tone map that information to viewable.

This image is the 0EV Image Tone-mapped. It looks very similar to the HDR image with the exception of more noise in the shadows and less detail in higlkight. but far better than the un-mapped (Edited)image is

How dynamic Range works

I hope this article was helpful. Sometimes I have ideas in my head and can’t fully put them to words (articulate), this is one of those articles

Thanks
PT

Post a Comment

Your email is never published nor shared. Required fields are marked *

*
*