Increase in Dynamic Range with H6D

[mike_minogue]

<p>I have a technical question in relation to part of what I read about Hasselblad's announcement of the new H6D.</p>

<p>From Hasselblad's website — “The 100MP sensor boasts an incredible dynamic range of 15 stops, allowing for unprecedented detail capture – from the deepest shadows to the brightest highlights.”</p>

<p>Prior to the H4D & H5D both had a dynamic range of 14 stops.</p>

<p>So here is my question, and I admit it will expose my ignorance of precisely how dynamic range works, but here goes:</p>

<p>Does the increase of the additional stop — the increase of the dynamic range — actually add an additional f-stop or does it simply increase the ability of the sensor to capture light as a portion of those individual stops?</p>

[alexo]

<p>The dynamic range of 15 stops means that the sensor is able to preserve detail within a 15 stop range. That means that if you expose for the gray tones, your highlight detail will be preserved within 7 stops of your exposure and your shadow exposure will be preserved within 7 stops of your exposure. </p>

<p>This is helpful when working in high contrast scenes where you have very bright highlights, such as the sun and very deep shadows. You don't have to make a decision as to whether expose for the highlights, thus preserving the detail and letting the shadows fall where they may or exposing for the shadows and blowing out your highlights.</p>

 

[mike_minogue]

<p>Thanks Alexander.</p>

<p>So just to confirm I have this correctly, lets say I puts Hasselblad's 24mm lens on an H5D with 14 stops of dynamic range. The 24mm has an aperture range from ƒ/4.8 - ƒ/32.</p>

<p>Lets then say I next use the same 24mm lens on the H6D with 15 stops of dynamic range. I'm assuming that each individual stop now has more dynamic range -- more of an ability to see light at each end of the spectrum? We haven't added an actual stop to the lens' capability, just more dynamic range within each individual stop - correct?</p>

[alexo]

<p>Hi Mike,</p>

<p>The dynamic range is a function of the sensor, not of the lens. Furthermore, a stop is a stop is a stop, they don't get better, they don't get worse. The dynamic range is not within an individual stop, it covers more stops.</p>

<p>When you're shooting a scene, and you make your exposure (regardless of the lens you use), you will have one part of the scene exposed properly and then you're going to have parts that are lighter than your "proper" exposure and some scenes that are going to be darker. So, let's say your exposure is 1/125 sec at f8. A sensor with 15 stop dynamic range will preserve details in places that are lighter than your perfect exposure and darker. So, in this instance, even though you're exposing at 1/125 and f8, your highlights will be preserved up to 1/16000 and f8 and your shadow detail will be preserved up to 1 sec and f8 FOR THE SAME SCENE. </p>

<p>In other words, you have a scene where there are some parts that are so bright, they need a proper exposure at 1/16,000 and f8 and some parts that are so dark that their proper exposure would be at 1 sec and f8. All of these parts are contained within that one picture. So, when you have low dynamic range, the parts that are in the deep, deep shadow will not contain any details. For example, if you have a car in the shadow, and you don't have a great enough dynamic range, you won't be able to see the car when you try to lighten that shadow in post processing. All you will get is black turning into gray. Whereas with greater dynamic range, when you lighten the shadow portion of your scene in Lightroom or some other program, you'll be able to see the car. The same concept comes into play with regard to the light areas. If you have a very light part of the scene that should have been exposed at 1/16,000 of a sec and f8, but you exposed at 1/125 and f8, that light area will also have details preserved when you try to darken it in post processing. Otherwise, it would just go from white to gray without any captured details.</p>

<p>So, the difference between a 14 stop dynamic range and 15 stop is that with a 15 stop, you will have detail preserved from 1/16,000 and f8 to 1 sec and f8 when you expose at 1/125 and f8, while the 14 stop chip will preserve detail from 1/8000 and f8 to 1 sec and f8 when you expose at 1/125 and f8.</p>

<p>I hope this clarifies things.</p>

[ondebanks]

<p>That's an excellent explanation from Alexander.</p>

<p>The only thing I can add is to address the question of "how is the extra dynamic range achieved?".</p>

<p>Dynamic range (DR) is limited at the bright end by too much signal (saturation - the pixel is full to the brim), and limited at the faint end by noise (where a pixel has captured so little light, the meagre signal only equals the electronic noise added in reading out the pixel).</p>

<p>For many years, the trend in sensor development has been to reduce the readout noise per unit area of the sensor. Although pixels have been getting smaller and hence their "full to the brim" capacity has also been shrinking, their capacity has been largely stable on a per unit area basis. [Four 4.5 micron wide pixels with 25,000 electrons capacity equals one 9 micron wide pixel with 100,000 electrons capacity]. So the gain in dynamic range has mainly come at the faint end, by reducing the sensor noise.</p>

<p>This is particularly noticeable when you compare medium format CCD and CMOS sensors with a fairly similar pixel size. They give comparably rich highlight and mid-tones, but the low-noise CMOS runs away with it in the shadow details, where detail can still be perceived for another 3 stops or so of deep underexposure.</p>

[mike_minogue]

<p>Thanks Alexander! I appreciate your further explanation.<br>

By the way, what is the formula you utilized in your response? Is there a dynamic range calculator online?</p>

[mike_minogue]

<p>Thanks Ray!</p>

[alexo]

<p>Hi Mike,<br>

<br />I didn't use a formula or a calculator (although I'm sure there is one). I took a supposed proper exposure (in our example 1/125 sec at f 8) as 1 stop + 7 stops over exposure + 7 stops under exposure = 15 stops dynamic range. That's it.</p>

[rodeo_joe1]

<p>15 stops of dynamic range would require the shadow parts of the image to have a brightness of only 0.003% of the highlights. This is a ridiculous demand on the total flare (unfocussed stray light) from the lens and camera body combination. So, yes the lens does have a bearing on the dynamic range that can be captured. As does the flocking and baffling within the dark-chamber of the camera.</p>

<p>15 stops dynamic range is a stupid figure to quote, and quite unachievable in real photography. Yes, you can shine a tightly focused and controlled beam of light at a bare sensor while shielding the rest of it from any light at all, and come up with a >32,000:1 ratio. But that's completely meaningless outside of a test laboratory. 12 to 14 stops is more than adequate, and even that will be swamped by flare in most practical picture-taking situations. A dynamic range approaching 2^15:1 would only be of interest to astro-photographers using such a sensor within a large and custom-made dark chamber, and with the highest quality optics costing many times that of any photographic lens.</p>

[ondebanks]

<blockquote>

<p>15 stops dynamic range is a stupid figure to quote, and quite unachievable in real photography.</p>

</blockquote>

<p>Indeed. I think it is more a case of reassurance, in two ways:<br>

(1) that the deepest shadows have reasonable signal to noise, and are not junk. A given scene, properly exposed and accounting for flare, might not have more than 12 stops of recorded brightness range; but there is a case for capturing that with a camera that can in theory go to 14 or 15 stops, rather than one which maxes out at 12 stops where that darkest 12th stop has equal signal and noise.<br>

(2) that gross underexposure would not be a disaster and can be recovered from in post-processing.</p>

[rodeo_joe1]

<blockquote>

<p>"(1) that the deepest shadows have reasonable signal to noise, and are not junk. A given scene, properly exposed and accounting for flare, might not have more than 12 stops of recorded brightness range; but there is a case for capturing that with a camera that can in theory go to 14 or 15 stops, rather than one which maxes out at 12 stops where that darkest 12th stop has equal signal and noise."</p>

</blockquote>

<p>Relatively cheap DSLRs have a 14 bit A/D converter these days. And the extra 3 dB of S-N ratio that going to 15 bits provides is practically insignificant, since even achieving a 12 stop DR between highlights and flare-filled shadows is extremely difficult in real life and unlikely to actually add any pictorial value to the image.</p>

<blockquote>

<p>"(2) that gross underexposure would not be a disaster and can be recovered from in post-processing."</p>

</blockquote>

<p>Solution to this: Get a better lightmeter or learn how to use the one you have, bother to review the captured image in camera or simply sack the incompetent photographer!</p>