Frustrated about HDR and dynamic range?

[Hans Kruse]

Hans, first - thank you for starting this thread...

as i understand :
you used 2 photographes in your sunset example;  say 2 stops different in exposure. A= Light and B= dark
The lightest one A you have developed in two different ways - only adjusting the  exposure in LR. So you have now two versions of Photo A
As a result you end up with 3 images and you mix them- as three layers-  together in photoshop.

( correct me if i am wrong)
if i am correct; the conclusion is that LR or ACR at the moment is unable to get one perfect version of photo A.
It is something i also conclude from my own experience with ACR.
it looks like that Abobe did not see coming the enormous dynamic range of some sensors- maybe in a next version it will do better.

Your method  is about how i deal with enormous dynamic range problems.. I am good with photoshop and find HDR programs usually not doing so well in getting a natural looking photo.

btw your sunset mix is very well done.

I would say that LR is an excellent tool to handle high dynamic situations and the limitation is in the sensors in our cameras. Some sensors are way better than others, but the sensor still has a limit and this is the situation I'm discussion here.

So in post #7 in this thread I show 4 pictures:

#1: The unedited version of the -1EV picture from the bracket sequence. The original bracket sequence was -3EV, -2EV, -1EV, 0EV, +1EV. The purpose of showing this version was simply to show how dark this is in the shadows.

#2: The edited version of the -1EV done in Lightroom using highlights and shadows sliders with exposure and some grad filters.

#3: Is the +1EV picture from the bracket sequence and edited with the same settings as the -1EV picture and adjusted for the difference in exposure. Lightroom can automatically by the short I mentioned (shift-alt-cmd-M) adjust the expost in all other shots selected relative to the most selected. The exposure in Lightroom of the -1EV was +0.9 and the +1EV one was adjusted to -1.18 and this is a difference of 2.08 where the exposure difference in the camera from bracketing should be 2EV. The exposure in the camera of the -1EV was 1/13s, f/16 and ISO 100 and the +1EV was 0.3s, f/16 and ISO 100. All done on a tripod.

#4: The manually merged result done in Photoshop. And as mentioned since the pictures look identical except for the blown highlights and noise in shadows teh merge is simply painting in the missing highlights around the sun which is done in 5 seconds.


Attached to this post is the 1:1 view of a part of the picture. #1: -1EV, #2, +1EV and #3: merged result.

[Hans Kruse]

Hans, by adjusting the exposure you are defeating the purpose of shooting the brackets in the first place. Also, I use LR/Enfuse to do the blending (merging) right in LightRoom.   - Jim

The adjustment of exposure was in post processing from different exposures from the bracket sequence. You could from the same exposures blend automativally using LR/Enfuse or the 32 bit Photomatix plugin I mentioned. They both work by exporting a TIFF file which is blended by an external program and then the result is brought back into Lightroom. It is not done IN Lightroom. Maybe one day Lightroom will feature blending directly in Lightroom of different exposures.

So what I describe here is exactly the same is these other blending methods in principle except that the blending is done manually by me in Photoshop from two fully edited files.

I have used LR/Enfuse from time to time but mostly before I discovered the method described in this thread. Sometimes I have been happy with the results from LR/Enfuse sometimes not and the same with the Photomatix plugin. I found them both to be superior to the 32 bit HDR Pro in Photoshop.

[Peter_DL]

I'm interested in experience in other blending methods in Photoshop.

Similar approach here
(but just with 1 Raw file). Anyway, the blending procedure is based on a Luminosity mask
with some interesting refinements in steps 10 to 12.

Peter

--

[Hans Kruse]

Similar approach here
(but just with 1 Raw file). Anyway, the blending procedure is based on a Luminosity mask
with some interesting refinements in steps 10 to 12.

Peter

--

Yes, there were many of these approaches around before Lightroom and ACR got the local edits and tone mapping controls. Also many used to make HDR tone mapping from single RAW files, but all that is history now in my opinion.

[Bart_van_der_Wolf]

Similar approach here
(but just with 1 Raw file).

Hi Peter,

When the procedure is based on a single Raw, then it's something quite different, it becomes a way of tonemapping with different gamma adjustments, but not with less noise.

Shooting actual exposure brackets means that more photons get recorded in the shadow exposure. More photons give a higher S/N ratio, less noise. Blending that will always be superior. The only thing that differs from the point of tonemapping is the difference in gamma, but that can be adjusted in various ways.

Cheers,
Bart

[Tim Lookingbill]

Thanks for the clarification with the 100% view crops, Hans. That helped quite a bit in my getting an idea on the benefits of layer blending bracketed exposures.

I see your technique got rid of the vertical banding in the shadows which wasn't evident in your previous full frame versions.

This appears to work best on shots with flat areas of tone where finer detail such as the grass can suffer from over softening. I was under the impression newer sensors on higher end cameras would render cleaner fine detail in the shadows. It looks pretty much but a bit cleaner than what I get on my 6 year old 6MP Pentax DSLR.

[Peter_DL]

Hi,

- just avoid a misunderstanding, my response was referring the PS blending part, as an alternative to
>>Add an adjustment layer and brush the areas where the highlights are blown.<< (from the initial post),
and without intending to question that bracketed exposures do provide a better database than one single Raw file.

Peter

--

[Hans Kruse]

Thanks for the clarification with the 100% view crops, Hans. That helped quite a bit in my getting an idea on the benefits of layer blending bracketed exposures.

I see your technique got rid of the vertical banding in the shadows which wasn't evident in your previous full frame versions.

This appears to work best on shots with flat areas of tone where finer detail such as the grass can suffer from over softening. I was under the impression newer sensors on higher end cameras would render cleaner fine detail in the shadows. It looks pretty much but a bit cleaner than what I get on my 6 year old 6MP Pentax DSLR.

No matter what camera you will have scenes with high enough DR such that a single RAW file will expose noise and maybe too much noise if you print large and do not want to see the noise and would like not to have too much clipped highlights. In essence cases where you otherwise would use HDR techniques and manual blending. The reason you will use such techniques is to expand the dynamic range of the camera system to match the scene you photograph by using bracketed shots which is a quite normal technique to use.

Notice in my example anything else than the clipped highlight areas were all coming from the most exposed shot. I only had to brush a relatively small area. This means that the quality of the higher exposed shot comes through fully in almost the entire photo.

Notice all I'm presenting here is a way to blend manually in Photoshop without the traditional work and difficulty.

Manual blending is a way to get more control over how blending is done than automatic blending which I mentioned in the beginning of this post.

To make it really clear how much I needed to brush in the example see the rough outline of what I brushed away from the most exposed shot to reveal the lesser exposed shot.

[Hans Kruse]

- just avoid a misunderstanding, my response was referring the PS blending part, as an alternative to
>>Add an adjustment layer and brush the areas where the highlights are blown.<< (from the initial post),
and without intending to question that bracketed exposures do provide a better database than one single Raw file.

What I'm trying to get across is that with the presented approach I do not need to do really editing in Photoshop except a bit of 100% brushing with the top layer to reveal what is under that layer. This means it is quick to do and all the editing decisions are made within Lightroom only and if you have to redo an image then the final steps of copying the edit to another exposure and the blending in Photoshop are very mechanical and can easily and quickly be redone.

[Hans Kruse]

This appears to work best on shots with flat areas of tone where finer detail such as the grass can suffer from over softening. I was under the impression newer sensors on higher end cameras would render cleaner fine detail in the shadows. It looks pretty much but a bit cleaner than what I get on my 6 year old 6MP Pentax DSLR.

The approach means that it is not needed to do any noise reduction on the images as long as the most exposed shot is well exposed for the shadows. This also means there is no softening af any structures. The amount of DR when this approach is needed clearly depends on how large the DR of the sensor is, but as Bart points out there is an inherent noise in low exposed shadows regardless of the sensor. So given then camera you have (new or old, Canon or Nikon D810) it will vary, however there is always a threshold where it will be needed. Well, at least as long as we use a global shutter on our cameras 

[jarnoh]

Lightroom is actually already doing tone mapping with its Shadow/Highlights/Clarity etc sliders.  When Process Version 2012 was released (LR4), I actually compared it to results on few published papers on HDR tone mapping, and the results were superb.  For extreme situations, the range provided by these sliders is a bit limited, but you can sometimes work around it by adding local adjustments on full frame.

For combining the exposures, you might also want to try http://jcelaya.github.io/hdrmerge/ which combines several frames of RAW data as a float DNG.  You can then do all your processing in Lightroom with a single merged exposure.

[Hans Kruse]

Lightroom is actually already doing tone mapping with its Shadow/Highlights/Clarity etc sliders.  When Process Version 2012 was released (LR4), I actually compared it to results on few published papers on HDR tone mapping, and the results were superb.  For extreme situations, the range provided by these sliders is a bit limited, but you can sometimes work around it by adding local adjustments on full frame.

For combining the exposures, you might also want to try http://jcelaya.github.io/hdrmerge/ which combines several frames of RAW data as a float DNG.  You can then do all your processing in Lightroom with a single merged exposure.

Yes, Lightroom has the tone mapping built-in and that is what I mention in the first post in this thread. I also mentioned the different merge methods of bracketed shots and with final tone mapping in Lighroom. I also mentioned that this approach works sometimes and sometimes does not work so well. The approach I presented actually builds directly on this capability to tonemap a single RAW image (or any other format for that matter including 32 bit TIFF files). The method I presented is a way to improve IQ in cases where the tonemapping of a single image is suffering from quality loss in shadows (edither noise and sometimes color shifts).

Tonemapping in Lightroom as I do it (I just call it editing my photos consists for me in using the white balance, exposure, contrast, highlights, shadows, whites and blacks and tone curve globally and typically a number of grad filters. Often the radial filter and the brush for local edits. The assumption for the blending approach is that there is one photo in a bracket sequence that is edited like this to the final look. The blending itself is only to improve IQ based on another shot in the bracket sequence that it can be blended with. The edits from the edited picture is copied to the other picture in the bracket sequence and exposure is adjusted to match and then blending is easy and quick as I described.

Again I would like to emphasize (for those who jump into the thread without reading it all) that the method I describe here is not suggested to be a panacea for all cases. It does not necessarily replace the other blending methods I mentioned that creat 32 bit TIFF files or other blending methods that just create 16 bit TIFF files and with final tone mapping in Lightroom or manual blending in the traditional way in Photoshop.

I have described the method in detail and this is available in a PDF file which I'm happy to send to all who contact me via e-mail. The e-mail address can be found on my homepage in my signature.

[ErikKaffehr]

Hi,

My understanding is that the benefit of Hans way of doing this is that it mostly relies on the mapping methods in Lightroom. Compressing the tone scale can cause problems with excessive noise in dark areas.

What Hans does, according to my reading, is to mix two exposures so he can have clean darks and still avoid clipping in the sky. I used to do this by blending in darks from a longer exposure. Hans I think found a smarter way to do it. I still need to work trough his recipe but my impression is that he blends in the parts that would be overexposed. That approach seems smart to me!

Best regards
Erik

Yes, Lightroom has the tone mapping built-in and that is what I mention in the first post in this thread. I also mentioned the different merge methods of bracketed shots and with final tone mapping in Lighroom. I also mentioned that this approach works sometimes and sometimes does not work so well. The approach I presented actually builds directly on this capability to tonemap a single RAW image (or any other format for that matter including 32 bit TIFF files). The method I presented is a way to improve IQ in cases where the tonemapping of a single image is suffering from quality loss in shadows (edither noise and sometimes color shifts).

Tonemapping in Lightroom as I do it (I just call it editing my photos consists for me in using the white balance, exposure, contrast, highlights, shadows, whites and blacks and tone curve globally and typically a number of grad filters. Often the radial filter and the brush for local edits. The assumption for the blending approach is that there is one photo in a bracket sequence that is edited like this to the final look. The blending itself is only to improve IQ based on another shot in the bracket sequence that it can be blended with. The edits from the edited picture is copied to the other picture in the bracket sequence and exposure is adjusted to match and then blending is easy and quick as I described.

Again I would like to emphasize (for those who jump into the thread without reading it all) that the method I describe here is not suggested to be a panacea for all cases. It does not necessarily replace the other blending methods I mentioned that creat 32 bit TIFF files or other blending methods that just create 16 bit TIFF files and with final tone mapping in Lightroom or manual blending in the traditional way in Photoshop.

I have described the method in detail and this is available in a PDF file which I'm happy to send to all who contact me via e-mail. The e-mail address can be found on my homepage in my signature.

[Hans Kruse]

My understanding is that the benefit of Hans way of doing this is that it mostly relies on the mapping methods in Lightroom. Compressing the tone scale can cause problems with excessive noise in dark areas.

What Hans does, according to my reading, is to mix two exposures so he can have clean darks and still avoid clipping in the sky. I used to do this by blending in darks from a longer exposure. Hans I think found a smarter way to do it. I still need to work trough his recipe but my impression is that he blends in the parts that would be overexposed. That approach seems smart to me!

Your understanding is correct although I would remove the word mostly. The only thing that happens in Photoshop is brushing the highlight areas away to reveal the lesser (or not) clipped highlights in the lesser exposed picture. The editing is entirely in Lightroom on a single picture which then is the basis for the rest of the procedure.

[Hans Kruse]

As a side comment on this approach the method relies on Lightroom can render exposures done at different exposures in the camera, like e.g. 2 stops apart very precisely if the exposure is adjust accordingly in Lightroom. Especially if you use the function Match Total Exposures in the fucntion in the Settings menu in the Develop module. For shots that are 2 stops apart from the camera this function would normally adjust the exposure difference in the exposure slider in Lightroom to be close to 2 stops, but not exact since the camera does not provide infinitely precise 2 stops apart bracketing. So rather than "manually" adjust for the 2 stops in Lightroom, it is more precise to use the Match Total Exposures function from what I have seen.

So what I often do is to adjust what I think is the optimal exposure from a bracket sequence and then later realize that is may not be the optimal judged from the look after editing rather than the look before editing by judging clipped areas. This happens typically in shots with the sun in it or with very strong light from the sun on e.g. clouds. Then what I do is to select all shots in the bracket sequence and keep the edited one the most selected and then copy edits to all by shift-cmd-S and then adjust exposure to all by shift-alt-cmd-M. Then I look at each of the exposures and choose the one that I find most compelling. I may even continue editing on that one and if the IQ is satisfactory I'm done. If there are noise in shadows and I have high exposed shots in the bracket sequence then I will copy the edits to the other shots in the bracket sequence and normalize exposure again (using the short cuts mentioned) and then choose the exposure optimal for the blending approach and with acceptable noise and then do the blend. Then I'm done.

In order to have shots available for this approach it is necessary to bracket the shots in such situations and I prefer 1 stop apart in the brackets and at least 5 in a bracket sequence in high DR scenes. Often I will end up with an edit of a single RAW file, but I never know this when I'm shooting as the judging of what is best only can be judged in Lightroom after the editing is done. It is not possible (in my opinion) to judge this on the back of the camera when shooting. It is also my shooting philosophy not to bother with judging histograms etc. when I'm shooting and concentrate on getting the right compositions while I'm there and the light may be changing by the minute.

[Iluvmycam]

OP, nice if you know what your doing with PS... I don't. I'm an old film photog and know very little about computers. As such I only use LR.

Here is an image that had lots of sky burnout. I just dodge and burn on LR to bring it back. It is a single image HDR. I have a better version, but no link to it. This is about 90% done.


nsfw

https://upload.wikimedia.org/wikipedia/commons/thumb/7/76/Queen_of_the_Bikers%27_Mardi_Gras_Copyright_2012_Daniel_D._Teoli_Jr..jpg/709px-Queen_of_the_Bikers%27_Mardi_Gras_Copyright_2012_Daniel_D._Teoli_Jr..jpg

I do lots of single image HDR. Don't have the luxury of using a tripod doing doc work.

[ErikKaffehr]

Hi,

I guess that the OP, you and for that part me are shooting different pictures. For landscape Hans and I need credible skies. Lightroom does a great job of recovering blown out highlights.

Here is an example what I am trying to do:


This is shadow detail:


And this is same shadow detail with my technique (that differs from Hans's method a bit)


There may be need for less workarounds if we used a camera with very little readout noise.

Best regards
Erik

OP, nice if you know what your doing with PS... I don't. I'm an old film photog and know very little about computers. As such I only use LR.

Here is an image that had lots of sky burnout. I just dodge and burn on LR to bring it back. It is a single image HDR. I have a better version, but no link to it. This is about 90% done.


nsfw

https://upload.wikimedia.org/wikipedia/commons/thumb/7/76/Queen_of_the_Bikers%27_Mardi_Gras_Copyright_2012_Daniel_D._Teoli_Jr..jpg/709px-Queen_of_the_Bikers%27_Mardi_Gras_Copyright_2012_Daniel_D._Teoli_Jr..jpg

I do lots of single image HDR. Don't have the luxury of using a tripod doing doc work.

[EricV]

Isn't this just Guillermo Luijk's zero noise technique?
http://www.luminous-landscape.com/forum/index.php?topic=17775.0

[Hans Kruse]

Isn't this just Guillermo Luijk's zero noise technique?
http://www.luminous-landscape.com/forum/index.php?topic=17775.0

No. I had not even seen his technique and have not (yet) tried it. If you read the description again you will realize that my suggested techinque is quite different. It is also different from the traditional blending techniques I have seen. If you are interested in a detailed description send me an e-mail and I will send you the PDF.

[EricV]

If you read the description again you will realize that my suggested techinque is quite different....

I guess I should have been more precise.  The implementation techniques are of course different, in that one uses built-in LightRoom controls to equalize exposures, while the other uses custom software, and one uses Photoshop layers to combine frames, while the other uses custom software.  However, the basic idea and the end results are the same -- create an image where the darker (noisier) pixels come from an over-exposed (less noisy) frame, mathematically scaled back down to the proper brightness.