Thinking of HDR with D7D - RAW for easy bracketing?

[[SiC]]

Okay, so I was riding my bike to work this morning - more time for useless thinking

The D7D can unfortunately only bracket ±1EV which supposedly isn't enough for "good" HDR images. (yes I would please like the same FW upgrade the A700 users got..)
But.. I've seen people doing HDR from a single RAW file, by "developing" three images at -1,0,+1EV for example.

I rarely use RAW, but maybe when I want to shoot HDR I can shoot three bracketed shots at -1,0,+1EV in RAW and then develop the -1EV image with another -1EV underexposure, and the +1EV image with another +1EV overexposure - thus creating my -2,0,+2EV HDR...

What do you think? Have someone tried it?

Thanks in advance!

BR,
/Zeb!

[Javelin]

I've been reading about this quite a bit. I think you can get some benefit from what your saying but I think the genreal idea is to get a wider DR than you can possibly get from a single shot. the problem with doing them handheld is they have to be tweaked like a panorama has to be to get the final image sharp. if your going to use a tripod to do it you can make the braket steps whatever you want but by manually adjusting the exposure to get the highs and lows you need.

[dewarp]

Hi BR

I use the D7D and have done a fair amount of HDR stuff. Please look at the three examples I put in the Photoclubalpha Gallery (.. > User galleries > dewarp > PD's HDR).

Obviously, I start with RAW files and convert these to jpg's. Both "Bones of a Fisherman" and "Wilgemond" were obtained from single exposures that I converted at three or more different exposure settings. "Panorama" was made from 3 or more separate exposures (from about 30) that were taken over a period of about 45 minutes as the sun went down (the foreground comes from when the sun was just below the horizon).

Unfortunately the HDR built into CS2 cannot handle multiple exposure jpg's originating from a single RAW file. So I use an old version of Photomatix. The output from this is somewhat unpredictable. You should try all of the different methods available in Photomatix and see which you prefer. Also, you may need to do some later manual masking in Photoshop to get the final acceptable image.

A typical problem - in "Bones of a Fisherman" there was a vertical lamppost on the dock in the background. The initial Photomatix runs produced a sudden increase in luminosity across the lamppost. What I had to do was use CS2 to remove the lamppost from each of the jpg's before I used them in Photomatix.

I have also done some HDR's the more conventional way, but tend then to switch the D7D to manual and take separate frames at -2, -1, 0, +1, -2. One serious problem with multiple frames, even when using bracketing at 3 frames per second, is movement of trees, people, etc. In "Panorama" the shots were selected more from the point of view of lack of displacement of the tree's leaves rather than the lighting of the specific exposure. Fortunately I had enough frames to be able to do this.

Hope this helps - regards Peter Dewar

[[SiC]]

Thanks for your input guys! Although I think you misunderstood me... I'm not gonna use one RAW exposure, I will take three exposures and then "develop" the -1EV as -2EV and so on...

Nice pictures dewarp, but I can see some strange shadowing effect... is this something Photomatrix did?
I have only used FDRTools so far.

/Zeb!

[01af]

... maybe [...] I can shoot three bracketed shots at -1, 0, +1 EV in raw and then develop the -1 EV image with another -1 EV underexposure, and the +1 EV image with another +1 EV overexposure---thus creating my -2, 0, +2 EV HDR ...

What do you think? Have someone tried it?

Never tried that but seems like a bad idea.

Developing a 'light-weight' HDR from one single raw file by exploiting the headroom will give less overall dynamic range than several shots but avoids any problems with moving people, leaves, clouds etc. When going through the hassle of actually doing several shots then you want the absolute best dynamic range and definition, so shooting at small EV steps and then expanding them artificially after the fact means combining the disadvantages of the 'light-weight' method and the real thing, just for a minor gain in camera-handling convenience.

-- Olaf

[[SiC]]

Never tried that but seems like a bad idea.

Developing a 'light-weight' HDR from one single raw file by exploiting the headroom will give less overall dynamic range than several shots but avoids any problems with moving people, leaves, clouds etc. When going through the hassle of actually doing several shots then you want the absolute best dynamic range and definition, so shooting at small EV steps and then expanding them artificially after the fact means combining the disadvantages of the 'light-weight' method and the real thing, just for a minor gain in camera-handling convenience.

-- Olaf

Thanks Olaf!
So, maybe it's not the preferred way to do it, but I will probably try it at some point anyway

/Zeb!

[jbtaylor]

Maybe someone can tell a newcomer what HDR is and how one might research how to use it or achieve it. Not sure if it is a noun or a verb.
Thanks.
JT

[[SiC]]

Maybe someone can tell a newcomer what HDR is and how one might research how to use it or achieve it. Not sure if it is a noun or a verb.
Thanks.
JT

Hi JT!

HDR means High Dynamic Range, and HDR imaging can be used for example when you have a scene with extreme dynamic range that the sensor cannot deal with (and may produce blown highlights for example).

Here's an example of such a scene:

( http://www.flickr.com/photos/thorstensson/2888955441/ ) <- for larger view

Just couldn't get the scene right with a normal exposure.

With HDR you combine several different exposures to get a good exposure of a high dynamic scene.

This was a rather fast attempt of answering, please feel free to ask more questions.

BR,
/Zeb!

[jbtaylor]

Thanks Zeb. Did I get the name right?
Does HRD refer to a camera function or a PP software function? Is this why I would use bracketing?
Please advise.
Thanks again.
JT

[[SiC]]

Thanks Zeb. Did I get the name right?
Does HRD refer to a camera function or a PP software function? Is this why I would use bracketing?
Please advise.
Thanks again.
JT

Hi again JT!
Sure did - My name is Sebastian but because of the length most people call me Zeb
AFAIK no camera can do HDR processing (does anyone else know?), but it probably will emerge at some point I think (at least the one exposure RAW derived version). So the answer is yes, that is why you use bracketing. The new v.4 FW for A700 will bracket ±2EV and so will the A900, I have the KM D7D and it will unfortunately only bracket ±1EV Depending on the scene I guess a 2EV span isn't enough, but 4EV will take you there.

/Zeb!

[jbtaylor]

Thanks again Zeb. I will check out Elements 7 for information about HDR when it arrives.

[01af]

Originally, HDR (which stands for High Dynamic Range) refers to image data formats where the bit depth per channel is more than 16 bits---typically 32 bits or even 64 bits. It also refers to the tools that can handle this kind of data. It's mostly used in the motion-picture industry, to create digital motion-picture sequences for special effects. HDR files can store images with an extremely wide dynamic range without the need of compressing it.

What photographers usually are referring to is dynamic compression, making a subject's wide dynamic range fit in the rather narrow dynamic range of our media, i. e. monitor screens or photo paper. This is not HDR; actually it's quite the contrary. For some time, the term DRI (Dynamic Range Increase) was the common, if slightly misleading, term for this kind of dynamic-range compression.

After Adobe Photoshop CS2 introduced the ability to handle HDR files, among photographers it has become common to confuse HDR data on one hand and the results of compressing HDR data in order to make them visible on the other hand. Photoshop's HDR feature can do three things: (1) it can acquire HDR data, either from other image-processing tools or from a set of digital captures taken at different exposure settings, (2) it can process and store the data uncompressed in a 32-bit format, and (3) it can compress the data, or the dynamic range thereof, in order to create a visible image from the HDR data.

Photographers often falsely believe that (3) is what HDR is all about. Actually, the most relevant points are (1) and (2). Of course, without (3) the other two would be pointless ... but it's important to understand that HDR, in the original sense, is about high-bit data and not so much about the compression thereof.

-- Olaf