Posted by Sean Bagshaw on February 11, 2009 at 4:39 PM
While the latest trend of surreal high dynamic range (HDR) images doesn't get much play on NPN, I have seen some discussion of HDR imaging here in the past. As outdoor photographers, we are often challenged with photographing scenes that span a very wide dynamic range. I recently gave a presentation on the topic to the Southern Oregon Photographic Association. Instead of just talking about the latest software and the "cool" HDR affects you can produce, I decided to focus on the idea that extending the dynamic range of what the camera can "see" has long been a goal of outdoor photographers and that there are many ways to effectively extend the dynamic range in an image. The presentation was well received, so I decided to adapt it in to an article on my blog. It is a bit long winded, but perhaps of interest to some. If you like, you can check it out HERE.
I would also be interested to hear of other techniques people use to extend the dynamic range in their images. I focused on the ones that I use regularly.
Posted by Robert Fisher on February 11, 2009 at 5:25 PM
In the digital realm, people were doing blended exposures before HDR became all the rage.
I know there are people who'll use more than one negative to create a print so as to get more than could be captured with a single frame of film.
Adams' Zone System is, in essence a method for getting more out of a single shot.
HDR has got a bad rap because of the whacked out, comic book stuff that people are doing and that is currently so popular. Used judiciously, it's a perfectly valid technique though.
Another member here on NPN commented a while ago about a Lightroom plugin called LR Enfuse that automatically blends different exposures together. It works well too.
Posted by Brad Hill on February 11, 2009 at 9:01 PM
Hi Robert,
I produce HDR images on a daily basis. My favourite technique is a manual one where I take multiple raw conversions from a single RAW capture (each using a different exposure setting) and sequentially combine them by layering and masking in Photoshop. I've tried various "automated" approaches (Merge to HDR, Photomatix, etc.) but prefer the control I have using the manual method.
I recently discussed the merits of this technique, including how it can be used in lieu of fill flash for sensitive species, on my website. I also included some tips for capturing images suitable for this technique. The discussion is found with this image (you have to click the "In the Field" tab below the image to see the discussion):
Posted by David Bostedo on February 11, 2009 at 9:46 PM
Sean - A bit less common in nature photography than other forms, but fill flash is often a way of squeezing a scene's dynamic range together in a single shot.
Posted by Robert Fisher on February 12, 2009 at 8:39 AM
Brad, nice stuff. But with respect, it's not HDR. It's more in the range of what John Paul Caponigro would call XDR or Extended Dynamic Range and even then I don't think it fits entirely.
By using a single file and reprocessing it, you're not gaining anything additional. You're merely redistributing what you already have. If you're opening up shadows you still have the issue of noise and if you've got highlights that are blown out, reprocessing the RAW file isn't going to bring them back.
Using your bear shot as an example (not to single you out, just to illustrate the point), it appears as though the brightness range of the shot fit pretty well into what the sensor could capture. It looks like you opened some shadows a bit and toned down some highlights a bit but generally, it looks like it fit pretty well. It had to, otherwise you wouldn't have been able to create the finished product you have. Given that, I'd suggest that careful work on the single RAW file could generate a result like the one you've got by multi-processing the file then using layers and masks. Don't get me wrong, I'm not for a second suggesting what you've done isn't good; it is. And I'm not suggesting you stop doing what you're doing. If it works for you, that's great. And you may get a different 'look' to the photo with your method that you like better than just working on a single RAW file. It's really not HDR though.
And to be clear, what we see in magazines and on the internet that people are calling HDR, really isn't HDR either. There aren't devices capable of reproducing HDR files. And by HDR I'm talking about the .hdr and .exr file types (the two most common). If you look at an HDR file on a normal computer screen it's a mess of blacks and whites. The HDR files we're looking at have gone through the tonemapping process to (a)convert them back to 16 (or 8) bit and (b) compress the high dynamic range back into a reproducible and visible 'low dynamic range' image. It's that converted, low dynamic range file that we see. After being tonemapped, they're no longer HDR files.
Oh yeah, I didn't say it earlier - nice essay, Sean (for what my opinion of it matters).
Posted by Brad Hill on February 12, 2009 at 9:17 AM
Hi Robert,
You're probably right, but I think we're playing a bit of a semantic game here. The process I use extends the visible tonal range of the initial image dramatically and beyond what could be done working on a single raw conversion (this last point could easily turn into long debate itself, but I won't belabour the point). Use the technique on a sensor with a wide dynamic range (like a D3 or D3x - see dxomark.com) and the dynamic range of the final image parallels that which the human eye can perceive in the scene, which is all I really care about anyway (I'm wondering what you'll think of THAT statement!). I'm comfortable with you calling it XDR or whatever you want to call it. Perhaps once we start using HDR to extend the dynamic range of an image BEYOND that perceived by the human eye we should start calling is Super Dynamic Range (SDR) processing? Or Super Sensory Dynamic Range Processing (SSDR)?
Cheers...
Brad
Posted by David Bostedo on February 12, 2009 at 9:37 AM
And to be clear, what we see in magazines and on the internet that people are calling HDR, really isn't HDR either.
In my opinion, magazines and common usage have already made HDR into a non-technical term. I've seen it used vastly more times to describe any image or technique with any increased dynamic range rather than the "real" meaning. Unfortunate, maybe, but I'd guess it's going to stay that way.
I do like "XDR" to mean all the other techniques, then letting "HDR" go back to it's more correct usage...but I don't think it's going to catch on.
Posted by Robert Fisher on February 12, 2009 at 10:03 AM
Caponigro tried, David. He's written an entire series of articles on Digital Photo Pro magazine on the various techniques, including HDR within that.
Brad, I see you took some offense despite my disclaimer. One of these days maybe I'll just learn to keep my mouth shut. Actually, nah that ain't gonna happen either. Sometimes semantics are important. And I, personally, think this is one of those cases. That said, if you go back and read my comments carefully, I said if your technique works for you that's great. So there's no need to get snippy and ask "I'm wondering what you'll think of THAT statement!" because I already agreed with it.
Posted by Brad Hill on February 12, 2009 at 11:03 AM
Robert,
Don't read too much into my reply - you have totally misinterpreted my comment. Yes, I said:
I'm wondering what you'll think of THAT statement!
But I said it simply because MY statement was highly debatable and controversial - you'll see so many estimates of the dynamic range of the human eye AND you'll see so many conflicting "measures" of the dynamic range of camera image sensors that my statement was as loaded as they come! It had nothing to do with snippiness. I have no problem with what you said at all! Did you miss the humour that I ended my comment with?
Ultimately I do think this discussion starts to become a mis-directed semantic game: ultimately we're just all trying to find a way to produce images that are visually appealing and playing with (extending, re-mapping, re-distributing, HDRing, XDRing, etc) the tonal range of the scene is just one way to do it. I look at the argument about HDR vs. XDR as one where we're trying to force fixed points onto a phenomenon that is actually continuous in nature. As technologies improve (and the dynamic range captured by image sensors improve) this will become a historic argument that served little purpose (and the terms HDR and XDR will fade in our distant memory). I completely accept your right to think otherwise, i.e, that it's more than semantics.
You'd have to try a whole lot harder than this to offend and/or insult me - I have real thick skin.
Cheers...take care...
Brad
Posted by Howard Cheek on February 12, 2009 at 12:17 PM
Thanks for this great link Sean, I've saved it to my favorites.
Posted by Sean Bagshaw on February 12, 2009 at 12:34 PM
You are welcome Howard! And thanks to the rest for some great thoughts and debate.
I think the distinction between extended dynamic range imaging (XDR, EDR...?) and high dynamic range imaging (HDR) is a good one and one worth defining. I use the term extended dynamic range in my article frequently, but I think I use it a bit interchangeably with HDR. With the added input from this thread, I think it is helpful to organize the terminology so that all of the techniques I describe in the article, as well as others that have been proposed in this thread, whether they be darkroom, filter, fill flash, RAW processing, blending or HDR software, are different ways to extend the dynamic range beyond what you would get with a simple click of the shutter. HDR imaging software takes a spot in the XDR line up, but has probably claimed the HDR title for itself at this point.
Posted by Brad Hill on February 12, 2009 at 12:55 PM
Hey Sean,
A very nice summary. I see value in using the term Extended Dynamic Range (EDR or XDR) to describe the process by which we extend the existing dynamic range of our camera's image sensor and produce an image that has the characteristic of having a wide dynamic range ("high" is such a bad word to use in describing the process in that the word is inherently relative and arbitrary). Thanks very much for starting this thread. I quite enjoyed it!
Brad
Posted by Arnab Banerjee on February 12, 2009 at 3:56 PM
Sean
Great article.
I have found Photomatix is very quick and give you an idea of what the final image will look like - or capable of. For fine art print, it's hard to depend on it as you rightly pointed out - too much noise, weird color shift and those moving elements! But I have seen that you can actually blend some part of the Photomatix image into your manual blending process i.e. just using Photomatix as another exposure or layer.
I have also heard (never tried this) the usage of Photomatix using different settings (like detail enhancer vs. tone compressor) and then blend the different outputs together.
All in all, Photomatix is a great tool and gets better with each version - but needs to be used carefully, as you mentioned.
Good article. I just want to let you know about a technique I recently discovered and shared with NP members, about using Photomatix for recovering clipped hightlights in the histogram. Sadly, the picture with the explanation was removed from my gallery but you can find it HERE in Photo.net.
For example you could apply this technique HERE to recover the yellow part of the sky that is only present in the Photomatix image version and not in the hand blend version due to the clipping in red channel (I guess). Combining both versions (Photomatix and hand blending) will yield a better rendition of the picture and I think is the best way to go.
Posted by Rod Barbee on February 19, 2009 at 11:08 AM
Discussion
14 repliesI would also be interested to hear of other techniques people use to extend the dynamic range in their images. I focused on the ones that I use regularly.
I know there are people who'll use more than one negative to create a print so as to get more than could be captured with a single frame of film.
Adams' Zone System is, in essence a method for getting more out of a single shot.
HDR has got a bad rap because of the whacked out, comic book stuff that people are doing and that is currently so popular. Used judiciously, it's a perfectly valid technique though.
Another member here on NPN commented a while ago about a Lightroom plugin called LR Enfuse that automatically blends different exposures together. It works well too.
I produce HDR images on a daily basis. My favourite technique is a manual one where I take multiple raw conversions from a single RAW capture (each using a different exposure setting) and sequentially combine them by layering and masking in Photoshop. I've tried various "automated" approaches (Merge to HDR, Photomatix, etc.) but prefer the control I have using the manual method.
I recently discussed the merits of this technique, including how it can be used in lieu of fill flash for sensitive species, on my website. I also included some tips for capturing images suitable for this technique. The discussion is found with this image (you have to click the "In the Field" tab below the image to see the discussion):
http://www.naturalart.ca/galleries/latest/detail_2.html
Cheers...
Brad
By using a single file and reprocessing it, you're not gaining anything additional. You're merely redistributing what you already have. If you're opening up shadows you still have the issue of noise and if you've got highlights that are blown out, reprocessing the RAW file isn't going to bring them back.
Using your bear shot as an example (not to single you out, just to illustrate the point), it appears as though the brightness range of the shot fit pretty well into what the sensor could capture. It looks like you opened some shadows a bit and toned down some highlights a bit but generally, it looks like it fit pretty well. It had to, otherwise you wouldn't have been able to create the finished product you have. Given that, I'd suggest that careful work on the single RAW file could generate a result like the one you've got by multi-processing the file then using layers and masks. Don't get me wrong, I'm not for a second suggesting what you've done isn't good; it is. And I'm not suggesting you stop doing what you're doing. If it works for you, that's great. And you may get a different 'look' to the photo with your method that you like better than just working on a single RAW file. It's really not HDR though.
And to be clear, what we see in magazines and on the internet that people are calling HDR, really isn't HDR either. There aren't devices capable of reproducing HDR files. And by HDR I'm talking about the .hdr and .exr file types (the two most common). If you look at an HDR file on a normal computer screen it's a mess of blacks and whites. The HDR files we're looking at have gone through the tonemapping process to (a)convert them back to 16 (or 8) bit and (b) compress the high dynamic range back into a reproducible and visible 'low dynamic range' image. It's that converted, low dynamic range file that we see. After being tonemapped, they're no longer HDR files.
Oh yeah, I didn't say it earlier - nice essay, Sean (for what my opinion of it matters).
You're probably right, but I think we're playing a bit of a semantic game here. The process I use extends the visible tonal range of the initial image dramatically and beyond what could be done working on a single raw conversion (this last point could easily turn into long debate itself, but I won't belabour the point). Use the technique on a sensor with a wide dynamic range (like a D3 or D3x - see dxomark.com) and the dynamic range of the final image parallels that which the human eye can perceive in the scene, which is all I really care about anyway (I'm wondering what you'll think of THAT statement!). I'm comfortable with you calling it XDR or whatever you want to call it. Perhaps once we start using HDR to extend the dynamic range of an image BEYOND that perceived by the human eye we should start calling is Super Dynamic Range (SDR) processing? Or Super Sensory Dynamic Range Processing (SSDR)?
Cheers...
Brad
In my opinion, magazines and common usage have already made HDR into a non-technical term. I've seen it used vastly more times to describe any image or technique with any increased dynamic range rather than the "real" meaning. Unfortunate, maybe, but I'd guess it's going to stay that way.
I do like "XDR" to mean all the other techniques, then letting "HDR" go back to it's more correct usage...but I don't think it's going to catch on.
Brad, I see you took some offense despite my disclaimer. One of these days maybe I'll just learn to keep my mouth shut.
Don't read too much into my reply - you have totally misinterpreted my comment. Yes, I said:
But I said it simply because MY statement was highly debatable and controversial - you'll see so many estimates of the dynamic range of the human eye AND you'll see so many conflicting "measures" of the dynamic range of camera image sensors that my statement was as loaded as they come! It had nothing to do with snippiness. I have no problem with what you said at all! Did you miss the humour that I ended my comment with?
Ultimately I do think this discussion starts to become a mis-directed semantic game: ultimately we're just all trying to find a way to produce images that are visually appealing and playing with (extending, re-mapping, re-distributing, HDRing, XDRing, etc) the tonal range of the scene is just one way to do it. I look at the argument about HDR vs. XDR as one where we're trying to force fixed points onto a phenomenon that is actually continuous in nature. As technologies improve (and the dynamic range captured by image sensors improve) this will become a historic argument that served little purpose (and the terms HDR and XDR will fade in our distant memory). I completely accept your right to think otherwise, i.e, that it's more than semantics.
You'd have to try a whole lot harder than this to offend and/or insult me - I have real thick skin.
Cheers...take care...
Brad
I think the distinction between extended dynamic range imaging (XDR, EDR...?) and high dynamic range imaging (HDR) is a good one and one worth defining. I use the term extended dynamic range in my article frequently, but I think I use it a bit interchangeably with HDR. With the added input from this thread, I think it is helpful to organize the terminology so that all of the techniques I describe in the article, as well as others that have been proposed in this thread, whether they be darkroom, filter, fill flash, RAW processing, blending or HDR software, are different ways to extend the dynamic range beyond what you would get with a simple click of the shutter. HDR imaging software takes a spot in the XDR line up, but has probably claimed the HDR title for itself at this point.
A very nice summary. I see value in using the term Extended Dynamic Range (EDR or XDR) to describe the process by which we extend the existing dynamic range of our camera's image sensor and produce an image that has the characteristic of having a wide dynamic range ("high" is such a bad word to use in describing the process in that the word is inherently relative and arbitrary). Thanks very much for starting this thread. I quite enjoyed it!
Brad
Great article.
I have found Photomatix is very quick and give you an idea of what the final image will look like - or capable of. For fine art print, it's hard to depend on it as you rightly pointed out - too much noise, weird color shift and those moving elements! But I have seen that you can actually blend some part of the Photomatix image into your manual blending process i.e. just using Photomatix as another exposure or layer.
I have also heard (never tried this) the usage of Photomatix using different settings (like detail enhancer vs. tone compressor) and then blend the different outputs together.
All in all, Photomatix is a great tool and gets better with each version - but needs to be used carefully, as you mentioned.
Good article. I just want to let you know about a technique I recently discovered and shared with NP members, about using Photomatix for recovering clipped hightlights in the histogram. Sadly, the picture with the explanation was removed from my gallery but you can find it HERE in Photo.net.
For example you could apply this technique HERE to recover the yellow part of the sky that is only present in the Photomatix image version and not in the hand blend version due to the clipping in red channel (I guess). Combining both versions (Photomatix and hand blending) will yield a better rendition of the picture and I think is the best way to go.