Posted by Darwin Wiggett on December 21, 2007 at 3:08 PM
Thought there might be some interest in this article - I use both techniques all the time but favor grads when possible - this article explains the reasons why: http://singhray.blogspot.com/ Darwin
I have always found that my HDR images have a washed out and well fake appearance to them. I always attributed it to my lack of skill with the product, now I see it isn't solely my abilities though I am sure that still plays a major part. Anyway excellent article Darwin.
Posted by Bret Edge on December 21, 2007 at 3:52 PM
Darwin: I'm with you. I use GND filters in the field and will occasionally do a digital grad if needed. There are times when a GND filter won't work and for that I'll create an HDR exposure.
As usual, a great article!
Posted by Bill Chambers on December 21, 2007 at 5:24 PM
Thanks Darwin!
Posted by Jesse Speer on December 21, 2007 at 5:56 PM
Excellent blog, and comparison. You don't see too many of these "the old ways are still good" articles in today's techno-frenzied world. I agree that grads generally trump HDR - even though they both are great tools. It's easy to misuse both, but far more so with the latter - not too mention the uncontrollable effects (movement-->ghosting, etc.). But you covered all that nicely. With processing software offering more options and getting more complicated each year, it makes grads that much more appealing. Grads have a pretty easy learning curve!! And I'm a huge believer in minimizing processing time.
I actually don't use either, but that's another story not for this thread. Thanks for sharing.
Posted by Greg Cope on December 21, 2007 at 6:04 PM
Excellent article on the advantages/disadvantages of both techniques Darwin!
Since I shoot 4x5 and occasssionally 645, I'm stuck with using ND Grads. Although they are easier to use on my 645; on the 4x5, with the rather dim gournd glass, ND-Grads are evil. Personally, I go to lengths to not have to use them at all.
Nice article, Darwin. Thanks for sharing it with us.
-PB
Posted by Tom Nevesely on December 21, 2007 at 6:42 PM
Darwin, great article! I love using my ND grad filters but have wanted to try out HDR. Somehow though, I always end up using a filtered shot even though I've also shot the scene for potentially using HDR. Filters just seem simpler to me. Your comparison is great and seeing both versions side by side I prefer the one shot with filters. I still want to learn HDR even if only to know more about it.
Posted by David Bostedo on December 21, 2007 at 6:53 PM
Interesting article Darwin. To be upfront, I've never used GNDs before, just polarizers and NDs; I always do anything I would have done with a GND instead with multiple exposures and a digital GND. (Maybe that invalidates this comment for some?) I've also never used real HDR software - just masks and gradients to simulate a GND.
Anyway, the article seems to mostly focus on Photomatix versus a GND filter. Is it possible that one of the reasons for your filter preference is that Photomatix is overkill? Seems to me that a simple simulation of a GND using gradients and masking would be much simpler and not suffer from the problems you mentioned (with contrast mainly). A digital simulated GND should produce identical results in the area where the filter is clear, so I don't really understand the problems you're seeing with Photomatix.
In my opinion, the only real advantage to using a real GND vs. a digital one is that multiple exposures are often needed for the digital one (just like HDR). This can of course lead to registration errors and loss of some image quality. (And makes it pretty impractical to do with film cameras). But of course there are big advantages - like no fixed number of stops or fixed grad line - it's totally adjustable. (These particular issues have of course been discussed on here before).
Posted by Darwin Wiggett on December 21, 2007 at 8:13 PM
Hi all, thanks for the comments. Let me just say there are as many ways to approach and solve the contrast problem as there are photographers - you can use HDR software, you can do manual exposures blends using layers and masks in Photoshop, you can use grad filters... I use them all (sometimes all three techniques on one image). But when I know a grad filter will get me what I want I will turn to it first, save myself tons of work in post-processing and get myself really solid quality images right in-camera. I say use whatever works best for you - there is no right answer, my blog entry was just about what works well for me most of the time and why - your mileage may vary. For me, grad filters still have an important place in my camera bag. Darwin
I say use whatever works best for you - there is no right answer, my blog entry was just about what works well for me most of the time and why - your mileage may vary.
That's eminently reasonable, and is completely in line with my own thinking. Because of where I shoot, I usually find myself in a situation where, if there's a dynamic range issue at hand, using a GND falls somewhere between problematic and utterly nonfunctional. I'm talking about settings where sky/land lines are not only entirely uneven, but are normally punctuated by foliage-bearing trees, and the inherent gaps within. I started experimenting seriously with digital blending more than four years ago and with HDR more than a year ago and I use both. I own a set of Lee GNDs, from back in my film-shooting days, but almost never carry them around anymore for the above cited reasons. (One of the shortcomings of GNDs that I rarely see mentioned is when they're used on scenes with glass-like reflections and render the reflections lighter than the thing that's being reflected. That seems to get a critical pass most of the time; perhaps we're so used to seeing the effect in photographs that it scarcely seems worthy of mention.)
FWIW, when it comes to HDR work with Photomatix, I almost always use more than three exposures (usually five, sometimes, four, occasionally more than five), separated by one stop. I've found this to produce far more vibrant results than going with the recommended three-exposure/1.6 or 2-stop option. I also nearly always end up running two tonemapping approaches on the exposure set and then overlay and blend these images in PS. I find this approach leads to the elimination of the so-called "HDR effect" when it comes to toning and a richer array of colors.
I am thinking of trying to see if I can squeeze some value out of those virtually moth-balled GNDs, however, for the rare occasion when I do find myself shooting a scene where grad-lines and reflection issues aren't problems.
Posted by Darwin Wiggett on December 21, 2007 at 8:49 PM
Kerry,
Your entry is so great - there are types of environments where a NDG won't work and then you need to pull more solutions out of your arsenal. Photomatix is a fantastic tool but it needs tweaked to get the most from it. I too often blend two tonemapped versions of HDR into one file and then often take the original RAWs and blend in parts of these exposures into the final image as well. It sounds like you do the same kinds of things and your results show it! Often the results of mixed HDR blends are amazing - but they take tons of time and lots of knowledge and experimentation. If I can get it with a grad, I will... and where I shoot (prairies and mountains) grads are easy to use. Darwin
Posted by Doug Dolde on December 21, 2007 at 9:06 PM
I use SIngh Ray grads but often there is still a lot of shadow detail that needs to be pulled out. Also I find the soft grads hard to use because it's hard to see the effect compared to a hard edge grad and forces more bracketing.
I use Photokit's great dodge (and burn) tools. They apply a gradient variably opacity layer very smoothly and easier than I can do on my own. I am amazed what I can pull out of the shadows with these.
Photomatix is a fantastic tool but it needs tweaked to get the most from it. I
Absolutely! I think one of the reasons some people sour on it is that they expect great results "out of the box." That's a recipe for disappointment. I remember being appalled at the first image I tonemapped (Detail Enhancer, wouldn't you know) with Photomatix but concluded very quickly that the best approach was to view the product as something akin to a RAW capture out of the camera; work, and a significant amount, might well be needed to attain something desirable. That mind shift seemed to work wonders.
I agree completely with what you said about the amount of time and experience needed to make HDR blending work pay off...and, no question...if you can obtain equal or superior results with a filter, it certainly makes sense to do so. You mentioned one of the certain downsides to HDR work in your blog piece--issues with moving foliage, etc. That's a real problem, and even though Photomatix has an optional algorithm to try to deal with it (it's pretty robust), it's imperfect and this issue all but renders HDR--and even manual blending on occasion--impossible in certain situations.
Discussion
13 repliesAs usual, a great article!
I actually don't use either, but that's another story not for this thread. Thanks for sharing.
Nice article, Darwin. Thanks for sharing it with us.
-PB
Anyway, the article seems to mostly focus on Photomatix versus a GND filter. Is it possible that one of the reasons for your filter preference is that Photomatix is overkill? Seems to me that a simple simulation of a GND using gradients and masking would be much simpler and not suffer from the problems you mentioned (with contrast mainly). A digital simulated GND should produce identical results in the area where the filter is clear, so I don't really understand the problems you're seeing with Photomatix.
In my opinion, the only real advantage to using a real GND vs. a digital one is that multiple exposures are often needed for the digital one (just like HDR). This can of course lead to registration errors and loss of some image quality. (And makes it pretty impractical to do with film cameras). But of course there are big advantages - like no fixed number of stops or fixed grad line - it's totally adjustable. (These particular issues have of course been discussed on here before).
That's eminently reasonable, and is completely in line with my own thinking. Because of where I shoot, I usually find myself in a situation where, if there's a dynamic range issue at hand, using a GND falls somewhere between problematic and utterly nonfunctional. I'm talking about settings where sky/land lines are not only entirely uneven, but are normally punctuated by foliage-bearing trees, and the inherent gaps within. I started experimenting seriously with digital blending more than four years ago and with HDR more than a year ago and I use both. I own a set of Lee GNDs, from back in my film-shooting days, but almost never carry them around anymore for the above cited reasons. (One of the shortcomings of GNDs that I rarely see mentioned is when they're used on scenes with glass-like reflections and render the reflections lighter than the thing that's being reflected. That seems to get a critical pass most of the time; perhaps we're so used to seeing the effect in photographs that it scarcely seems worthy of mention.)
FWIW, when it comes to HDR work with Photomatix, I almost always use more than three exposures (usually five, sometimes, four, occasionally more than five), separated by one stop. I've found this to produce far more vibrant results than going with the recommended three-exposure/1.6 or 2-stop option. I also nearly always end up running two tonemapping approaches on the exposure set and then overlay and blend these images in PS. I find this approach leads to the elimination of the so-called "HDR effect" when it comes to toning and a richer array of colors.
I am thinking of trying to see if I can squeeze some value out of those virtually moth-balled GNDs, however, for the rare occasion when I do find myself shooting a scene where grad-lines and reflection issues aren't problems.
Your entry is so great - there are types of environments where a NDG won't work and then you need to pull more solutions out of your arsenal. Photomatix is a fantastic tool but it needs tweaked to get the most from it. I too often blend two tonemapped versions of HDR into one file and then often take the original RAWs and blend in parts of these exposures into the final image as well. It sounds like you do the same kinds of things and your results show it! Often the results of mixed HDR blends are amazing - but they take tons of time and lots of knowledge and experimentation. If I can get it with a grad, I will... and where I shoot (prairies and mountains) grads are easy to use. Darwin
I use Photokit's great dodge (and burn) tools. They apply a gradient variably opacity layer very smoothly and easier than I can do on my own. I am amazed what I can pull out of the shadows with these.
Absolutely! I think one of the reasons some people sour on it is that they expect great results "out of the box." That's a recipe for disappointment. I remember being appalled at the first image I tonemapped (Detail Enhancer, wouldn't you know) with Photomatix but concluded very quickly that the best approach was to view the product as something akin to a RAW capture out of the camera; work, and a significant amount, might well be needed to attain something desirable. That mind shift seemed to work wonders.
I agree completely with what you said about the amount of time and experience needed to make HDR blending work pay off...and, no question...if you can obtain equal or superior results with a filter, it certainly makes sense to do so. You mentioned one of the certain downsides to HDR work in your blog piece--issues with moving foliage, etc. That's a real problem, and even though Photomatix has an optional algorithm to try to deal with it (it's pretty robust), it's imperfect and this issue all but renders HDR--and even manual blending on occasion--impossible in certain situations.