Steve's Digicams Forums

Steve's Digicams Forums (
-   General Discussion (
-   -   What is the Percentage Range? (

TCav Mar 31, 2010 6:51 AM

I would like to think that we all know enough not to point our cameras directly into the sun, so I think that even the sun is more brightness than we need to account for in out photography.

There is no 100% bright; there's only the maximum amount of brightness that can be recorded by a sensor. That's why the EV scale is open-ended. What you're talking about is the dynamic range of an image sensor, and that's what HDR is for: to get around perceived shorcomings of the combined effects of image sensors and output devices in this area.

Mark1616 Mar 31, 2010 6:54 AM

I think he is saying, make the assumption the sun is 100%, not stating categorically that it is. So it is a theoretical range using things we can all appreciate.

Apart from that I can't 'shed any light' on this topic.

Mark R. Mar 31, 2010 10:00 AM


Originally Posted by Bynx (Post 1072980)
Well, two pretty smart guys duking it out over the sun and the sky. How about answering the question. I think its an important one and should be easy enough to answer. Its a technical question and should apply to all cameras.

OK, while I could go on about atmospheric refraction and how it has nothing to do with the whiteness of the light source, let me try to contribute to an answer for the original question.

I would suggest, for starters, to consult definitions of illuminance and EV.

Illuminance is measured in lux, defined as 1 lumen per square meter. It is therefore a measure of light flux per unit area.
See also:
From there, I quote:

The human eye is capable of seeing somewhat more than a 2 trillion-fold range: The presence of white objects is somewhat discernible under starlight, at 510^−5 lux, while at the bright end, it is possible to read large text at 10^8 lux, or about 1,000 times that of direct sunlight, although this can be very uncomfortable and cause long-lasting afterimages.
There is also a table of typical lux values at

On the other hand, EV is a scale based on factors of 2, so for every EV step, the amount of incident light doubles.

Comparing 10^−5 lux (starlight) to 10^5 lux (direct sunlight), this is a factor of 10^10. On a scale of powers of 2, this corresponds to about 30 EV steps.

Remember that the scale is open on both ends, so in terms of illuminance, a truly "black box" would read "ten to the power of negative infinity". So that's not really practical. You have to choose a lower and upper threshold, which is what I've tried to do with the examples from the lux table.

Caveat: illluminance is not the same as irradiance, which measures the physical power of the radiation. I'm not sure inasfar one should rather use irradiance to do the above calculation.

Also, none of us would use the same camera settings to shoot direct sunlight and a closed box. So I'm actually not quite sure I understand the "actual question", Bynx.

If there are photons, no matter how few, if you leave your aperture open for long enough, you'll get a picture. Conversely, if you make your aperture small enough and your shutter fast enough, direct sunlight won't over-saturate your sensor.

So what exactly is the question that we're supposed to answer?


mtngal Mar 31, 2010 3:29 PM

I can contribute nothing to the technical nature of your question, only a practical one from playing around with several different cameras. So, in the "for what it's worth" category, here are some thoughts about what I understand to be Bynx's question/concern/point, based on his second post, mentioning about HDR vs. raw vs. jpg.

Photography is more about reflected light (how light reflects off of objects) than it is about direct light (light coming from a source such as a lightbulb or the sun). So a lot of the discussion about light, which I personally find quite interesting from an intellectual point of view, seems to not quite answer Bynx's question, which is essentially, can a raw file practically capture enough information for all dynamic ranges in a possible picture. Since most people don't shoot light emitters as a subject/intentionally (well, there are a number of exceptions, such as pictures of cool neon signs), just light reflectors, I think his answer lies more with practical photography. Reflected light might be a smaller dynamic range than that of complete black to the brightest light possible (i.e., sunlight?), it's still a huge range.

I can't come up with figures or research - that's not my "thing". However, I have shot both jpg and raw and compared them. I've shot pictures in situations with a huge amount of dynamic range (I do that a lot as a matter of fact). I've played around a bit with HDR also.

My practical experience is this - raw will give you an edge over jpg. But it isn't all that much more - just as a guess, it gives maybe a half to one stop more on the bright side (at the most, and my current camera might not even give me that much) and maybe 1-2 stops more in the shadows (maybe a bit more). As soon as you start "pushing" the shadows, you introduce noise and the more you push, the more noise you get, and the harder it is to see details. And if the highlights are blown out, they are gone no matter what - raw will not give you something that's not there.

So your premise that in situations where there is a huge range of light/dark that HDR will do better than a single raw file is (in my experience) definitely true. I have a series of photos that I did as HDR that came out very nicely. When I decided that they would look even better in b&w, I tried to use the middle exposure only, since I didn't care if the sky was blown out/white. The only problem was that the middle exposure had some of the building blown out and even using the raw file didn't bring back detail in that spot. So I took the first underexposed file that had detail in those spots, and tried to lighten the shadows enough to bring out the detail in the other parts. The noise was so significant that running it through noise reduction software couldn't retain the fine details and the noise was so significant it was very distracting, even when converted to b&w. At that point I got frustrated and just did a b&w conversion of the HDR file, which contained information in both the brightest and darkest parts of the picture and had no noise.

My conclusion - jpg works well for many things. Raw works better than jpg but is still limited. HDR is the only way to go for certain scenes, and can be the closest to what my brain interprets a scene to be, and is the only way to capture detail everywhere.

But that goes only for scenes where there's a very big dynamic range that you want to capture detail in. The vast majority of most people's subjects don't require it. And sometimes the smaller dynamic range capability of the camera sensor works in my favor, rendering an OOF background as black where it might have been ugly dark green blotches.

Going on to another part of your original question - I'm not so sure that overall sensor size affects how much dynamic range a camera is capable of capturing. I think the pixel rating has more of an influence - my old Sony F717 with it's 5 mp small sized sensor was quite capable of capturing a pretty wide dynamic range, more so than the 8 mp Panny FZ30 I briefly owned (the sensor sizes were, I think, similar). I've often wondered if a sensor's dynamic range capability were connected to its noise sensitivity - as sensors become noisier they can't capture detail in as big of a range (could that be manufacturers trying to deal with the increased noise, by cutting down the dynamic range so you don't see the noise in the shadows? I don't know).

And how much does a lens contribute to a picture's dynamic range? It can, but doesn't always, have an effect. I have two lenses, both high quality. One is a zoom and one a prime. The zoom is a much bigger lens, bigger front element (not surprising due to the complexity of making a constant aperture f2.8 zoom lens). The prime has a small front element and is (relatively) tiny, even though it is faster (f1.8). The prime has a greater dynamic range and renders blacks much better than the zoom with noticeably better contrast. Another prime lens I have is an elderly one with no coatings. It's definitely low-contrast and is hopeless when shooting sunsets and bright colors. But it doesn't over-saturate flowers, handles green very well and its really sharp so it can give details where other lenses make things look more smooth, due to over-saturation of bright colors. So yes, my experience is that lenses do make a difference - but not a huge amount, certainly not as much as the sensor does.

TCav Mar 31, 2010 4:18 PM

When you look at something, your iris adjusts the diameter of the pupil so that you receive enough light so you can see and recognize the detail on whatever it is you're looking at. Other nearby objects that may be brighter are darker are not as visible because you aren't trying to look at them. When you shift your gaze to those other objects, your iris readust so you can, but then what you looked at originallly is less visible.

When you take a photo, the camera's autoexposure system adjusts the exposure settings to capture an image with visible detail on whatever you pointed the camera at. When you look at other areas of the image that were not properly exposed, there is no way to go back an properly expose that part of the image, like there is with your eye.

HDR is a way to create an image that is more properly exposed from edge to edge, corner to corner, that is not possible any other way, and indeed, isn't possible in real life.Image sensors have a dynamic range of about 9 to 11 EV, which is plenty, but when 14 or 12 bit RAW data is converted to 10 or 8 bit images, and then displayed on 10 or 8 bit sRGB monitors, that dynamic range is lost. HDR is an attempt to compress the dynamic range of a scene so that it can survive all that. The dynamic range of any single component is irrelevant since there are so many bottlenecks to the process.

Bynx Apr 1, 2010 8:16 AM

The purpose of my question was to have the relevance of HDR brought out. I was getting a little pissed at those couple of guys who kept saying they could do it better this way or that way and HDR wasnt necessary. Raw was superior and better than going the HDR route. HDR is the only way to have a range which closely matches the range of the human eye. Tcav nicely points out the eye looking at one spot and adjusting, then moving to another darker or lighter spot and readjusting. Well in an HDR photo all the adjusting has been done so its just look around whether its light or dark. While I didnt get a precise answer to my question, its been answered in a practical way. Thanks to the three of you.

ac.smith Apr 1, 2010 2:21 PM


Originally Posted by Bynx (Post 1072980)
I think its an important one and should be easy enough to answer. Its a technical question and should apply to all cameras.

It is not anywhere near as simple as you make it out to be. The first problem is that you are asking for an answer on a linear scale (percentages) while we typically measure light photographically on a logarithmic scale and astronomically on an inverse logarithmic scale.

The visible (apparent) order of magnitude (inverse logarithmic) for the sun is -26.73 and the darkest sky on earth (which may not be true black) is about 9 ( A very coarse photographic approximation might be on the order of 36 EV.

Based loosely on DPReview's camera dynamic range tests a number of jpeg engines on DSLRs manage about an 8.5 EV dynamic range and at least some manage 10 EV in raw. That would seem to suggest that three RAW exposures spaced at 9 stop intervals might capture the dynamic range the OP referenced when processed as HDR. Fortunately, as stated above, most scenes do not have anywhere near that dynamic range.

A. C.

Bynx Apr 1, 2010 3:45 PM

Well ac, I sure didnt mean all that easy. But there are constants we are dealing with so there is an answer. A jpeg will cover 9 EV of the possible 36 EV which will actually be closer to 30. A Raw file would cover 12 EV?

iowa_jim Apr 1, 2010 6:28 PM

Don't look a gift horse in the mouth.

TCav Apr 1, 2010 8:26 PM

The Wikipedia article on HDR lists the dynamic range of a Canon EOS-1D Mark II. At ISO 50, it has a dynamic range of 11.3 EV, it peaks at ISO 100 with a dynamic range of 11.6 EV, and at ISO 3200, it has a dynamic range of only 8.7 EV. provides similar values for other cameras.

All times are GMT -5. The time now is 12:27 PM.