One photo for HDR?

The usual process is to take at least 3 photos at different exposures and then combine in e.g. Photoshop or Photomatix. Unless the subject is absolutely static, I always get ghosting, even when using Canon's auto 3 shot with one click process.

Why not just take one photo in RAW, make 2 copies and adjust the exposure in a RAW processor to +1 and -! and then combine the 3? One pic, no ghosting. I've used this on a couple of occasions but I suppose it can't be that simple. I'm guessing exposure correction in RAW is not quite the same as "when taken". I suppose a single shot can have 0 in the shade and 255 in the light and so no correction would be possible in PP. I think it would work however if the single shot retained information at the extremes.
--
Descartes to his wife - "You annoy me, therefore I exist."

Your single RAW image contains no more dynamic range than any other single RAW photograph. Good RAW conversion to, say, 16 bit tif files will bring out as much of that range as it contains or as much as the user requires.

To gain on the above, multiple images at, say, 2 stops apart each will give a dynamic range extended by 4 stops. That's real extra range.

However, you still output to the same media, screen or print, so your new data will have to be squeezed into the same old same old. Thus, not every HDR will result in a gain in the end product. Generally the subject needs to suit the technique and some overwork HDR to the point of creating quite unnatural, exaggerated results.

It's satisfying to do when it works well but can easily be overworked.

Cheers, Tony.

Dr. Leonard said:

The usual process is to take at least 3 photos at different exposures and then combine in e.g. Photoshop or Photomatix. Unless the subject is absolutely static, I always get ghosting, even when using Canon's auto 3 shot with one click process.

Why not just take one photo in RAW, make 2 copies and adjust the exposure in a RAW processor to +1 and -! and then combine the 3? One pic, no ghosting. I've used this on a couple of occasions but I suppose it can't be that simple. I'm guessing exposure correction in RAW is not quite the same as "when taken". I suppose a single shot can have 0 in the shade and 255 in the light and so no correction would be possible in PP. I think it would work however if the single shot retained information at the extremes.

Click to expand...

I use the one-photo approach a lot with Photomatix Pro for recovering highlights and shadows, for slight tonemapping, when the subject moves or changes too fast, or simply when I didn't bracket. It works surprisingly well in a lot of situations. I usually go up to + - 2. Lately I have used it more with my compact than with my DSLR because of the smaller dynamic range and the resulting blown out skies.

--

D5000 - Nikkor AF-S DX 18-105mm F/3.5-5.6G ED VR - Nikkor AF-S 50mm f/1.4 G - Nikkor AF-S 70-300mm VR - Canon PowerShot S95 - Canon PowerShot S3

Dr. Leonard said:

The usual process is to take at least 3 photos at different exposures and then combine in e.g. Photoshop or Photomatix. Unless the subject is absolutely static, I always get ghosting, even when using Canon's auto 3 shot with one click process.

Why not just take one photo in RAW, make 2 copies and adjust the exposure in a RAW processor to +1 and -! and then combine the 3? One pic, no ghosting. I've used this on a couple of occasions but I suppose it can't be that simple. I'm guessing exposure correction in RAW is not quite the same as "when taken". I suppose a single shot can have 0 in the shade and 255 in the light and so no correction would be possible in PP. I think it would work however if the single shot retained information at the extremes.

Click to expand...

I suppose working with Photomatix this technique makes sense, however there are different HDR processors that may give you equivalent effect from a single image - i.e. Oloneo. Even in LR4 you can influence how the captured dynamic range is represented in the final image, making this technique pretty much redundant.

If you chose to do it, depending on your camera and your shooting style, you may get better effect by adjusting exposure by 2 stops rather than one.
--
Peter

Dr. Leonard wrote:
cut here>

Dr. Leonard said:

Why not just take one photo in RAW, make 2 copies and adjust the exposure in a RAW processor to +1 and -! and then combine the 3? One pic, no ghosting. I've used this on a couple of occasions but I suppose it can't be that simple. I'm guessing exposure correction in RAW is not quite the same as "when taken".

Click to expand...

cut here>

'Exposure' adjustment should have been titled 'Rendering' adjustment or something similar. There is no possibility of changing the exposure after the event and it leads to the misconception above. There are, as mentioned before, methods to get the most out of one RAW captured data but that is only tone mapping and can be applied to any RAW image, under any circumstances, to get the most dynamic range available therein.

Most RAW converters (like ACR, etc.) allow modification of Fill light, Black levels, Contrast, Highlight recovery among others. All those latter adjustments ARE (somewhat crude) tone mapping. HDR software can give more precise regulation of the same process but is not any further towards HDR with a single image than RAW developer software.

Hopefully, someone will produce a sensor that can be read out at about 12% of the way through an exposure (or preferably adjustable) so you get two readings, about 3 stops apart, the first of which is suitable for highlight detail readings and the full length reading for the rest of brightness levels. Then we could all stop messing with multiple images and HDR software and get on with our photography. There would still be movement in a subject with fast motion (like waving leaves) but no need for a second separate shot to be taken - no fresh mirror-lift nor shutter closure and reopening. Also, the longer of the two exposures would have blur and the short exposure would be inside that blur.

Oh, well! Just a dream for the time being, I suppose.

Cheers, Tony.

And best done from a base raw file, but yes it can be done, yes it produces pleasing results if the base file is properly exposed but no, it does not actually extend the dynamic range of the image or camera.

See examples in the Flickr group dedicated to it "HDR from a single RAW "
http://www.flickr.com/groups/raw2hdr/

Most HDR/tonemapping software packages can save you a few steps and import the raw directly.

--
http://www.flickriver.com/photos/newmikey/

I guess we need to consider a couple of points re this subject.

First there is the 'technically' correct HDR methods. Second there is the 'artistic' methods.

Now what I mean is that to get the most of dynamic range then multiple shots is the way to go. But if you are after 'artistic' results then how you get there is irrelevant.

I guess 'true' HDR and 'realistic' results are the realm of those who promote the multiple shot approach. Those who like to bend the rules, simply don't care, and do it their own way.

But one key thing that must be kept in mind. Not all shots are suitable to either method. Often simply PP will get excellent results.

regards

brucet said:

I guess we need to consider a couple of points re this subject.

First there is the 'technically' correct HDR methods. Second there is the 'artistic' methods.

Now what I mean is that to get the most of dynamic range then multiple shots is the way to go. But if you are after 'artistic' results then how you get there is irrelevant.

I guess 'true' HDR and 'realistic' results are the realm of those who promote the multiple shot approach. Those who like to bend the rules, simply don't care, and do it their own way.

But one key thing that must be kept in mind. Not all shots are suitable to either method. Often simply PP will get excellent results.

regards

Click to expand...

What PP method do you recommend over pseudo-hdr for recovering highlights and lifting shadows and why? How much time does one need to get the job done with your method?
--

D5000 - Nikkor AF-S DX 18-105mm F/3.5-5.6G ED VR - Nikkor AF-S 50mm f/1.4 G - Nikkor AF-S 70-300mm VR - Canon PowerShot S95 - Canon PowerShot S3

First not all photos are the same so there is no one solution.

I use AfterShot Pro to convert NEFs to 16 bit tiffs. This gives the best, for me, files to work on. I then open those 16 bit tiffs in PSPx4. Then the following workflow.
Levels.
Fill Light/Clarity.
Curves.

Then, subject to each photo, I'll attend to any noise issues. Then if needed I'll go back a make selections on any area that needs specific attention.
Finally USM.

BUT. And it's a big BUT. Every photo needs a different approach. I'm not a big fan of following a set routine for every photo. Only occasionally will I batch a group of photos.

I'm not adverse to dropping a 16 bit tiff into Photomatix and seeing what the out come may be. Works sometimes and not other times. Only one way to find out. I have tried dropping an NEF straight into Photomatix but haven't been happy with the results.

HDR, Tonemapping, Pseudo, call it what you like. I've seen great paintings done with a 6 inch brush. Doesn't mean it can't be done. Only means that you like the end result regardless of how you got it.

If I had one bit of advice to anyone taking on HDR/Tonemapping in any form. That would be to learn to recognize what photos can or need to be HDR'ed or Tonemapped.

With my work I utilize both bracketed HDR shots and tonemapping extensively.

http://www.cre8ivephotography.com

good luck.

I agree with brucet -- there are many routes to enhancing photos. For HDR, current best practice is, from a tripod, three or more exposures in raw format at different shutter speeds. The exposures are blended and tonemapped to taste. Doing other than best practice can also yield satisfying results.

For example, from a single JPG exposure -- tonemapping (freeware FusionHDR http://fusion.ns-point.com/ ) followed by segmentation simplification (Buzz Pro 3) for a artistic effect.

IMHO, satisfying results are more important than procedure, best practice or otherwise.

David Dollevoet

djddpr said:

satisfying results are more important than procedure, best practice or otherwise.

David Dollevoet

Click to expand...

I agree. Although in this case, I personally prefer your first image, as the processed one shows a lot of loss of detail that I think adds to the image.
But, each of our eyes are different.

--
Jim

Jim B (MSP) said:

But, each of our eyes are different.

Click to expand...

As an Old English Sheepdog, your statement is something we like to keep under wraps!
Mine are blue and brown.



Cheers, Tony.

Didn't know you were an OES lover too. Wonderful image. I forgot the time when ours was that fuzzy and small, and the "mooneye" startled me at first.
--
http://www.flickriver.com/photos/newmikey/

Dr. Leonard said:

Why not just take one photo in RAW, make 2 copies and adjust the exposure in a RAW processor to +1 and -! and then combine the 3? One pic, no ghosting. I've used this on a couple of occasions but I suppose it can't be that simple. I'm guessing exposure correction in RAW is not quite the same as "when taken". I suppose a single shot can have 0 in the shade and 255 in the light and so no correction would be possible in PP. I think it would work however if the single shot retained information at the extremes.

Click to expand...

This is a well-known approach that works great if the DR of the scene is more than the JPG DR of your camera but within the DR of your sensor, and if you can expose the image so that supposedly blown highlights and shadows are recoverable by the raw converter. True HDR would still result in less noise and better colours in the lifted shadows.

There is another less-hassle approach, to use a raw converted capable of doing something like this in a single conversion. For a while, DxO OP has been doing this. Also the latest versions of LR/ACR are are substantially improved in this regard. Maybe also the 32bit-depth version of Raw Therapee?

brucet said:

First not all photos are the same so there is no one solution.

I use AfterShot Pro to convert NEFs to 16 bit tiffs. This gives the best, for me, files to work on. I then open those 16 bit tiffs in PSPx4. Then the following workflow.
Levels.
Fill Light/Clarity.
Curves.

Then, subject to each photo, I'll attend to any noise issues. Then if needed I'll go back a make selections on any area that needs specific attention.
Finally USM.

BUT. And it's a big BUT. Every photo needs a different approach. I'm not a big fan of following a set routine for every photo. Only occasionally will I batch a group of photos.

I'm not adverse to dropping a 16 bit tiff into Photomatix and seeing what the out come may be. Works sometimes and not other times. Only one way to find out. I have tried dropping an NEF straight into Photomatix but haven't been happy with the results.

HDR, Tonemapping, Pseudo, call it what you like. I've seen great paintings done with a 6 inch brush. Doesn't mean it can't be done. Only means that you like the end result regardless of how you got it.

If I had one bit of advice to anyone taking on HDR/Tonemapping in any form. That would be to learn to recognize what photos can or need to be HDR'ed or Tonemapped.

With my work I utilize both bracketed HDR shots and tonemapping extensively.

http://www.cre8ivephotography.com

good luck.

Click to expand...

Thanks. Again: How much time does one need to get the job done with your method? Typically, or on average. Roughly.
--

D5000 - Nikkor AF-S DX 18-105mm F/3.5-5.6G ED VR - Nikkor AF-S 50mm f/1.4 G - Nikkor AF-S 70-300mm VR - Canon PowerShot S95 - Canon PowerShot S3

My 'usual' work flow. These are all subject to no major changes being needed!!

Approximate times.

Load the NEF and hit my preset in ASP. 5 seconds.
About 4 seconds to save an NEF to 16 bit tiff in ASP. 4 seconds.

Subject to PSPx4 already being open.

Open the 16 bit tiff. 2 seconds.
Levels, auto. 4 seconds.
Fill Light/Clarity. This is the one setting I play with. 6 seconds.
Curves. I have setting I usually leave Curves on. 4 seconds.

Now if the photo needs some noise work that may take 10 - 20 seconds depending on how much work needs to be done. This can sometimes be 10 seconds. Some times a minute if I'm being fussy.
USM. I have a setting I leave it on. 4 seconds.

Save. 3 seconds.

So all up about a minute per photo.

Keep in mind that I select my photos in ASP. If they make it past that point I usually know what they will need in PSPx4.

The key in ASP is having a few Presets to use depending on what you have and what you are looking for.

BUT and there's always a BUT. But it depends. Depends on each photo. If you look at my web site ( http://cre8ivephotography.com ) you will see much of my work is similar. So most of the time I know my camera settings and work flow settings. They don't vary much. The thing is that much of my 'fine tunning' is done afterwards. Then comes the real editing for me.

Now on every day photos and snaps the above work flow covers 90% of what I'm looking for.

Remember though that this is just a guide and there will always be photos that need fiddling with.

Good luck.

I see you've received a lot of opinions on this but here's another...

The reason the multiple shot HDR works better is because each individual shot is exposed for a particular part of the scene. The more dynamic range in your scene... the more exposures you should be using. If there is quite a bit of DR, then 3 shots isn't enough... 7 or 9 is more commonly used.

That way the darker areas are exposed correctly and the brightest areas are too. Saving all the detail that is present. If you just take one shot and lighten and darken it to get your multiples... then you're losing details in those areas that your camera could have retained if exposed correctly, and usually adding noise to the dark areas as you pull them brighter.

"Ghosting" can be handled in Photoshop, after the fact.

The whole idea of HDR is to make up for what your camera sensor lacks. Total Dynamic Range. When you look at a scene with your eyes, your brains adjusts the light and dark and you don't really notice the drastic differences... but sensors aren't advanced enough to offer the dynamic range required to do the same thing. They can only properly expose for a certain amount. So blending (tonemapping) the best parts of each exposure you take ultimately creates an image closer to what you saw with your eyes.

Wordy and boring... but decided to add that for anyone that was reading along and new to HDR.

Now, if your goal is not to make a scene more realistic but more colorful, cartoonish, etc... then I agree that it doesn't matter how you get there as long as you are happy with the results.

All that said, I want to add: Where I said "can't", "don't", "shouldn't".... thats just my opinion. Nothing is "absolute" about any of this... except... that if you are happy with the results... it doesn't matter how you got there.

Some guy (joeyfoto) posted these in a thread years ago and I have it bookmarked... it is not mine ... but bookmarked because I've been to the actual place many times, and love the end result... I wanted to post it here. It's extreme HDR. And coincidentally all 3 are from 1 RAW, he claims.

Dr. Leonard said:

The usual process is to take at least 3 photos at different exposures and then combine in e.g. Photoshop or Photomatix. Unless the subject is absolutely static, I always get ghosting, even when using Canon's auto 3 shot with one click process.

Why not just take one photo in RAW, make 2 copies and adjust the exposure in a RAW processor to +1 and -! and then combine the 3? One pic, no ghosting. I've used this on a couple of occasions but I suppose it can't be that simple. I'm guessing exposure correction in RAW is not quite the same as "when taken". I suppose a single shot can have 0 in the shade and 255 in the light and so no correction would be possible in PP. I think it would work however if the single shot retained information at the extremes.
--
Descartes to his wife - "You annoy me, therefore I exist."

Click to expand...

--
So **** Clark dies in the same year we aren't suppose to have a new years....

Well played Mayans... well played...

brucet said:

My 'usual' work flow. These are all subject to no major changes being needed!!

Approximate times.

Load the NEF and hit my preset in ASP. 5 seconds.
About 4 seconds to save an NEF to 16 bit tiff in ASP. 4 seconds.

Subject to PSPx4 already being open.

Open the 16 bit tiff. 2 seconds.
Levels, auto. 4 seconds.
Fill Light/Clarity. This is the one setting I play with. 6 seconds.
Curves. I have setting I usually leave Curves on. 4 seconds.

Now if the photo needs some noise work that may take 10 - 20 seconds depending on how much work needs to be done. This can sometimes be 10 seconds. Some times a minute if I'm being fussy.
USM. I have a setting I leave it on. 4 seconds.

Save. 3 seconds.

So all up about a minute per photo.

Keep in mind that I select my photos in ASP. If they make it past that point I usually know what they will need in PSPx4.

The key in ASP is having a few Presets to use depending on what you have and what you are looking for.

BUT and there's always a BUT. But it depends. Depends on each photo. If you look at my web site ( http://cre8ivephotography.com ) you will see much of my work is similar. So most of the time I know my camera settings and work flow settings. They don't vary much. The thing is that much of my 'fine tunning' is done afterwards. Then comes the real editing for me.

Now on every day photos and snaps the above work flow covers 90% of what I'm looking for.

Remember though that this is just a guide and there will always be photos that need fiddling with.

Good luck.

Click to expand...

Thanks. Sorry but I'm not quite sure in which step you recover the hightlights and lift the shadows for which others use one shot pseudo HDR.

--

D5000 - Nikkor AF-S DX 18-105mm F/3.5-5.6G ED VR - Nikkor AF-S 50mm f/1.4 G - Nikkor AF-S 70-300mm VR - Canon PowerShot S95 - Canon PowerShot S3

Dr. Leonard said:

The usual process is to take at least 3 photos at different exposures and then combine in e.g. Photoshop or Photomatix. Unless the subject is absolutely static, I always get ghosting, even when using Canon's auto 3 shot with one click process.

Why not just take one photo in RAW, make 2 copies and adjust the exposure in a RAW processor to +1 and -! and then combine the 3? One pic, no ghosting. I've used this on a couple of occasions but I suppose it can't be that simple. I'm guessing exposure correction in RAW is not quite the same as "when taken". I suppose a single shot can have 0 in the shade and 255 in the light and so no correction would be possible in PP. I think it would work however if the single shot retained information at the extremes.
--
Descartes to his wife - "You annoy me, therefore I exist."

Click to expand...

It has been a few months since I have experimented with using one Canon RAW (CR2) image and process multiple times using DPP (the Canon Raw Processing Software) then use the processed 16 bit TIF files as input to the HDR software.

Today, I processed 3 images: one as normal exposure , one using +1.0, and one using -1.5.

The reason I used the +1.0 (instead of +1.5) was because the original exposure was metered to favor the Highlights.

I have limited experience with HDR processing so I chose to use the "defaults" without making any type selections or changes for this test.

The final HDR processed image actually is very nice and has accomplished somewhat more details in the shadow areas. Highlights (such as Clouds) plus color of the sky were also improved.

My opinion is: There definitely is a benefit for using one original RAW exposure in such a manner. I am NOT experienced with HDR processing. However,I am aware of the basis for originally making multiple exposures using the + and - settings -- and using a Tripod for better alignment, etc.... This would obviously be needed for a wide variance in Dynamic Range Exposures.
--
Vernon...

In PSPx4, after levels, go to Adjustments/Fill Light/Clarify.
You have two sliders. Use the Fill Light to play with your shadows.
Clarify can give you a HDR look if that's what you are after.

Keep in mind that PSPx4 has a few ways to open up the shadows.