Luminous Landscape Forum
Equipment & Techniques => Digital Cameras & Shooting Techniques => Topic started by: Abdulrahman Aljabri on July 01, 2015, 04:49:09 pm
-
How many bracketed pictures do you take to do HDR for complex interior spaces? Also, what increments do you use? Can you get away with shooting 7 pictures at 1.6 stop increments and get the same quality HDR file as you have shot 11 pictures at 1 stop increments?
Thanks
-
How many bracketed pictures do you take to do HDR for complex interior spaces? Also, what increments do you use? Can you get away with shooting 7 pictures at 1.6 stop increments and get the same quality HDR file as you have shot 11 pictures at 1 stop increments?
Thanks
it was said that Fuji was using (in cameras like S5 Pro with "SuperCCD") 2 set of sensels, absolutely identical otherwise, but separated by 4 stops by a simple mask on top of one set /so the sensor was otherwise, underneath, a simple Sony one/ to reduce the amount of light reaching them... so just 2 exposures separated by 4 stops shall do... and 3 will get you to absolute HDR nirvana of postprocessing... however which camera does allow 4 stops of separation in some automated mode ? so you can get by with 3 stops x 3 shots.
-
How many bracketed pictures do you take to do HDR for complex interior spaces? Also, what increments do you use? Can you get away with shooting 7 pictures at 1.6 stop increments and get the same quality HDR file as you have shot 11 pictures at 1 stop increments?
Thanks
Assuming that you want to use HDR processing simply to compensate for deficiencies in the dynamic range of your digital camera sensor (rather than to obtain the grotesque "HDR" effects that were a fad a few years ago), then the answer depends very much upon your own sensor. It is a matter in which considerable improvements have been made in recent years.
With my Nikon D810 I find that 2EV increments are adequate and I have yet to encounter a scene for which more than 5 exposures at 2EV increments was not sufficient. (Processing in Nik HDREfexPro from within Lightroom)
-
Shooting 11 shots seems overkill, aside from eating up processing time.
I've been toying with the idea of using an Expodisc 2 and using it to determine my over and under shots for three exposures. Might pay to use it and get the proper exposure spike in the middle of the camera's histogram, and then spin the shutter speed off until the spike is maybe 20% inwards on the left and right edge (i.e. Over & Under shots.) of the histogram? That, theoretically, would allow for whatever oddities your sensor and your camera settings like Active-D lighting, etc. produce too.
If you were really good, you could also tune the three spikes with the WB and Tint Control as well to get the spike perfect without any RGB colors on the sides of the perfect spike indicating a WB error.
I might try this today on mine and see how this would work in practice for HDR speed settings on the D800E and older D7000.
Add:
I tried the above with the Expodisc 2. Once I got the proper 4 seconds (ISO 250 at f/14 in a dark garage.) exposure, the D800E Over was 3 stops, and the Under was 4 stops and both spikes were still about 20% inside the histogram ends. I would have thought 3 Over & 3 Under would place them in the same area near the ends, but +3 and -4 were the ones that worked and about 20% in at each end. Not as linear as I would have thought, and these were long exposures: 30 seconds (+3 Over), 4 seconds (Normal), and 1/4 sec. (-4 Under). Odd. ???
SG
-
How many bracketed pictures do you take to do HDR for complex interior spaces? Also, what increments do you use? Can you get away with shooting 7 pictures at 1.6 stop increments and get the same quality HDR file as you have shot 11 pictures at 1 stop increments?
Hi,
It depends on how high your quality goal is, and if the subject has smooth gradients.
When the brightness transition of a smooth gradient is covered by two different exposures, then the difference of the Signal to Noise ratio of the exposures must not be too large. Otherwise the transition will become visible as a sudden jump in noise pattern, and is relatively easy to spot due to the lack of detail.
I've been able to push the exposure differential to two stops maximum, without running into trouble, but I routinely shoot at 1.33 stop intervals to be on the safe side when extra processing will amplify the differences. Usually 7 bracketed exposures is enough for very contrasty natural light situations, on a few occasions 1 or 2 extra exposures are welcome. 7 stops at a 1.33 EV interval, adds/replaces 9.3 stops of better quality dynamic range to the 11 - 13 that the camera already has and reduces the shadow noise to 1/3rd of the single exposure while adding some better (color and structure) detail. That's already a lot of improvement.
Cheers,
Bart
-
How many bracketed pictures do you take to do HDR for complex interior spaces? Also, what increments do you use? Can you get away with shooting 7 pictures at 1.6 stop increments and get the same quality HDR file as you have shot 11 pictures at 1 stop increments?
There is no fixed rule since it heavily depends on the blending software/method you use to mix your shots.
With optimal HDR software, more than 3 or 4 shots for ANY situation is just unnecessary, and 7 or 11 shots is just plain crazy. You'll not only get the same quality with 7 pictures at 1,6 intervals than with 11 at 1 stop intervals, but the same with 4 pictures at 3 stops intervals.
Unless your camera is very old, optimum software will be able to produce the same quality of results allowing 2-3 stops between shots. If your software is not optimum (e.g. Photomatix), ok you probably need more crazy shooting.
Regards
-
Hi,
I think so. My cameras used to have ample DR, so I have seldom resorted to HDR. I would say that two exposures can be quite enough, one for the shadows and one for the highlights, perhaps a third one with "proper" exposure?
An earlier posting of mine may be worth checking: http://www.getdpi.com/forum/medium-format-systems-and-digital-backs/53815-testing-real-world-dynamic-range.html#post630479
Best regards
Erik
Can you get away with shooting 7 pictures at 1.6 stop increments and get the same quality HDR file as you have shot 11 pictures at 1 stop increments?
Thanks
-
it was said that Fuji was using (in cameras like S5 Pro with "SuperCCD") 2 set of sensels, absolutely identical otherwise, but separated by 4 stops by a simple mask
3,6 stops actually:
(http://www.guillermoluijk.com/article/superccd/histosr.jpg)
(http://www.guillermoluijk.com/article/superccd/comp.jpg)
But those sensels weren't identical however. The highlight sensels (R) were much smaller:
(http://www.guillermoluijk.com/article/superccd/superccd.jpg)
The original article: FUJI SUPER CCD. IN CAMERA HDR (http://www.guillermoluijk.com/article/superccd/index.htm)
Regards
-
But those sensels weren't identical however. The highlight sensels (R) were much smaller:
well, for that I 'd assume you need to have an actual photo of the sensor layout, not a marketing illustration... simple logic says that it is much easier to deposit something that let "3.6EV" less light on surface of the regular Sony Semi manufactured CCD than bother to make a non standard (CCD on top of that) chip in low quantities (unless you are Sigma of course, and Fuji is not - but is well known by its games with toppings, like X-Trans).
-
There is no fixed rule since it heavily depends on the blending software/method you use to mix your shots.
With optimal HDR software, more than 3 or 4 shots for ANY situation is just unnecessary, and 7 or 11 shots is just plain crazy. You'll not only get the same quality with 7 pictures at 1,6 intervals than with 11 at 1 stop intervals, but the same with 4 pictures at 3 stops intervals.
Unless your camera is very old, optimum software will be able to produce the same quality of results allowing 2-3 stops between shots. If your software is not optimum (e.g. Photomatix), ok you probably need more crazy shooting.
Regards
Guillermo, What software do you recommend? Jim
-
Hi,
The HDR in LR6 is quite OK, I would say. LumaRiver HDR is another favourite of my.
Best regards
Erik
Guillermo, What software do you recommend? Jim
-
Guillermo, What software do you recommend? Jim
I couldn't say since I use my own software for optimum blending (Zero Noise). SNS-HDR seems to work very well.
Regards
-
How many bracketed pictures do you take to do HDR for complex interior spaces? Also, what increments do you use? Can you get away with shooting 7 pictures at 1.6 stop increments and get the same quality HDR file as you have shot 11 pictures at 1 stop increments?
Thanks
It all depends on the DR of the scene and the DR of the sensor. Most Nikon sensors have a DR of around 13 EV, of which 7 have very low noise. So in effect you can easily have steps of 2 EV. Now let us come to the DR of the scene. If you have a bright out side visible through the window and you want that and relatively dark interior, you can calculate the range, by using your meter in spot mode. Once point to the bright light outside and once to the dim light inside. What I do
. Assume the Light difference is 18 EV
. Assume that 2 EV steps are sufficient
. Assume that you want to use only 6 EV of your DR,
then
. The DR difference is 18-6 = 12
. Steps = 2 EV
. So shots required = 12/2 = 6
That way you will get absolutely clean shadows. In case you feel that the useful DR of sensor is 10EV instead of 6EV, you adjust the number of shots accordingly.
-
I have been shooting Bracketed images and using them for Tone Mapping since 2008 so I may have information about the technique that you may find useful.
In my humble opinion, unless you are a mathematician or a physicist I would not concern yourself with how Dynamic Range is calculated, I would stick to the photography. Assuming that you currently shoot with a DSLR. this is how I work and a methodology that produces consistent IQ:
1. If at all possible, always shoot from a rigid tripod using the Mirror Lockup option and an electronic or Cable Release to shoot your frames
2. Select the correct Aperture for the lens you are using and use as low an ISO as possible
3. In poor light (in any light really), set the Focus manually using Live View plus Zoom or Focus Peaking if it is available i.e. switch off the Autofocus on the lens. In poor light many Camera Bodies in the low to mid range of a manufacturers range are compromised in their ability to "lock on" to subjects when shooting in a Burst Mode. In a burst of 8 to 10 frames a proportion of the frames may be "soft". You do not want ANY "soft" frames
4. Use the appropriate Metering option on the camera to correctly expose for the brightest source of light in the scene you are capturing. This setting should be equivalent to your 0EV position
5. Produce a range of RAW images around this 0EV position. I would suggest that you start by using -2EV, -1EV, +1EV and +2EV. I have seldom needed a greater range
6. Take the five RAW frames into your Post Processing software and very carefully, fully process the 0EV frame such that the White Balance is correct and any blown highlights are treated. It is important that you treat any Digital Noise at this stage as well. Make sure that the processing Histogram stretches from the extreme left hand edge of the Y axis to the extreme right hand side of the Y axis (maximise the DR)
7. Copy these settings and then apply them to all five frames
8. Export the five images as 16-Bit TIFF files to your hard drive
9. Begin by importing all five 16-Bit TIFF files into Photomatix Pro software and then use the Tone Mapping or Fusion option to produce an image
10. An explanation of how to use Photomatix Pro software is too lengthy to include here so I will assume that you know what to do. As a guide line, if you are after a final end result more akin to the output available as Photo Merge to HDR in LR6, then stick with the options termed Balanced, Photographic and Neutral, which are displayed on the right hand side of the image displayed
11. When you have adjusted the image as seen in the Photomatix Pro software to your satisfaction, click on Apply and use Save As to move the 16-Bit File to your hard drive
12. Open the Tone Mapped or Fused 16-Bit File in Adobe Photoshop and then fine tune the end result as you would any other image
Note that I often select the minimum number of frames for processing in Photomatix Pro that result in the outcome that I want to achieve. More is not necessarily always best and usually the more frames that are combined, the more accentuated any Digital Noise becomes.
Do not be put off by the negative comments made about Photomatix Pro in this Post. I have found it to be excellent and certainly far more flexible in terms of what can be achieved VS Photo Merge to HDR in LR6 which is probably still more than enough for most people.
If you want any more information then send me an e-mail address so that any points can be addressed away from this Website. Inevitably there will be a proportion of people who see it differently and I don't have the time and the stomach to get embroiled in heated arguments. Hope this helps you. What I have described produces images of consistent quality time after time in my hands.
-
It occurred to me that I have not posted an image to show you the Technique in action. Here is an image shot on Abermar Beach, Pembrokeshire, West Wales, Uk, in May this year. It uses three frames which have been processed with Photomatix Pro Vs. 5.0.5a. Here I have used the Canon EOS 5D Mark I with the Distagon T* 2.8/21 ZE at f/5.6 and ISO 200. It may look better viewed Fullscreen. I used Nik's Color Efex Pro 4 to finish off the image. It has made a fine print.
-
Very nice image.
I would appreciate an explanation of how you approached the choice of 0 EV exposure in this image, referring to your previous post: "Use the appropriate Metering option on the camera to correctly expose for the brightest source of light in the scene you are capturing. This setting should be equivalent to your 0EV position."
What would be the brightest source of light in this image? I assume you do not mean the setting sun but some other feature? What metering option did you choose and why?
thanks for your post and explanation
-
Thanks for your enquiry. These RAW images are now archieved but I have got the external hard drive out of storage to remind me of how I produced this image.
I normally look at the BBC Weather Website to establish the exact time, by postcode, for either Sunset or Sunrise on the day I plan to shoot. I arrive at least 1 hour before the predicted time and set up the equipment for the shoot. The Camera/Lens is precisely leveled using a Manfrotto Hotshoe Spirit Level (the Canon EOS 5D Mark I does not have an inbuilt electronic levelling device), hence the position of the camera equipment is confirmed long before sunrise or sunset. From years of experience, I know that the Distagon T* 2.8/21 ZE produces its highest resolution corner to corner at an aperture of f/5.6 so this is set in Aperature Priority mode. Focus is then established manually and an appropriate ISO setting made, in this case I used ISO 200. The ISO reflects the balance between adequate Shutter Speed and the potential for Digital Noise. The lower the ISO used, normally the better. As the appointed time nears, I shoot a series of Frames at +1/3 increments up to +2EV using Centre Weighted Average Metering (from experience you will know that with the sun filling the majority of the frame, that the camera will tend to Underexpose the image. hence the + increments). I review each frame looking closely at the Highlight Alert and the shape of the Histogram. In the case of this sunset, +1 1/3 EV was the setting that was optimum in terms of offering the best Histogram and avoiding excessive highlight issues. At an appropriate time, when the disc of the sun was where I wanted it to be, I shot five Bracketed exposures at -2, -1 +1 and +2 EV either side of my theoretical optimum of +1 1/3 EV. From experience I know that I will get a decent final image from a combination of all or some of these Frames, it is trial and error. In the end I used only the - 2/3 EV + 1/3 EV and the the +1 1/3 EV Frames. The final image out of Photomatix Pro was fine tuned for Levels/Curves, Saturation etc. and Sharpened selectively for the foreground detail. I finished the processing in Nik's Color Efex Pro with the White Neutraliser as far as I can remember. Blue was also tweaked in the RGB area. I hope that you can make sense of this. Good luck with your photography. If I had been planning to use just a sinle RAW Frame for PP, then I would have worked only on the + 1 1/3 EV Frame. From experience I know that I will get a superior outcome using multiple frames into Photomatix Pro software. It is just my personal opinion of course and many, many others will cry foul for me using Photomatix Pro. Had I not told you then you would have been no wiser. By the way, for this image, the shutter speeds were good so the images were captured in bursts VS Mirror Lockup. I use Mirror Lockup mostly for architectural images, particularly indoors. If you still own a Nikon D700 or the Canon EOS 5D Mark I then don't get rid of it. Both cameras can still produce superb images if you maximise their strengths. Both cameras are vitually a steal second hand as most people belive that they will produce better images with higher Megapixel camera bodies. In reality this may not be the case as they are very much more technique dependent. Below is a Grab Shot of what the Histogram for the exposure selected to represent my baseline looks like. Bracketing -1EV and +1EV either side should be sufficient.
-
This is the view from the High Altar, through the Choir to the Nave inside the 11th century Winchester Cathedral, Hampshire, England. This uses 5 Frames. Distagon T* 2.8/21 ZE. Shot August 2013.
-
I have been shooting Bracketed images and using them for Tone Mapping since 2008 so I may have information about the technique that you may find useful.
In my humble opinion, unless you are a mathematician or a physicist I would not concern yourself with how Dynamic Range is calculated, I would stick to the photography. Assuming that you currently shoot with a DSLR. this is how I work and a methodology that produces consistent IQ:
4. Use the appropriate Metering option on the camera to correctly expose for the brightest source of light in the scene you are capturing. This setting should be equivalent to your 0EV position
5. Produce a range of RAW images around this 0EV position. I would suggest that you start by using -2EV, -1EV, +1EV and +2EV. I have seldom needed a greater range
6. Take the five RAW frames into your Post Processing software and very carefully, fully process the 0EV frame such that the White Balance is correct and any blown highlights are treated. It is important that you treat any Digital Noise at this stage as well. Make sure that the processing Histogram stretches from the extreme left hand edge of the Y axis to the extreme right hand side of the Y axis (maximise the DR)
If you want any more information then send me an e-mail address so that any points can be addressed away from this Website. Inevitably there will be a proportion of people who see it differently and I don't have the time and the stomach to get embroiled in heated arguments. Hope this helps you. What I have described produces images of consistent quality time after time in my hands.
I am curious why, if you are metering for the brightest source of light in the scene as your "base" exposure, you are shooting frames at -1 and -2EV. Also, if the base exposure is metered for the brightest source of light in the scene, why would the highlights be "blown"?
-
If you are shooting an architectural interior, say for example inside a cathedral looking up the Nave to the Altar, you will be presented very often with multiple light sources, a large area of glass behind the altar and very often oblique lighting from high up in the vaulted roof and in addition artifical lighting, sometimes pointing directly at the camera i.e. these are difficult images to shoot. You may be using Centre Weighted Average Metering and that will give you a starting point but you will be wise to take a series of frames at + and/or - 1/3EV steps and review the resulting images on the back of the camera. You will be looking at the shape of the Histogram and for the subtle changes that these incremental adjustments have made.
The second point is that if you are shooting in a Bracketed Mode, you will have to accept the fact that the camera will shoot either side of your optimum EV i.e. it will automatically shoot at +EV Fame and a -EV Frame.
Thirdly, as mentioned in one of my previous posts, it is very useful to have these frames as you may get a superior result by using only a few of the exposures in Photomatix Pro i.e. a Mix and Match approach
Don't get too bogged down in these technicalities as you can still repair problem areas of the Tone Mapped image in Photoshop at the end if you have to, but it helps not to have a totally compromised starting image coming out of the Tone Mapping process. I hope that you can understand what I am trying to explain. It just works for me 99% of the time. Just run with the basic ideas and see what you get.
-
How many bracketed pictures do you take to do HDR for complex interior spaces? Also, what increments do you use? Can you get away with shooting 7 pictures at 1.6 stop increments and get the same quality HDR file as you have shot 11 pictures at 1 stop increments?
Thanks
I t wil depend on the luminance range of the scene and the dynamic range of the sensor. I prefer to do my HDR shooting in one stop increments from -4.0 to +4.0. While this vastly exceeds the normal luminance range of almost all subjects it also allows me to decide which subset of those images will make the best composite.
My favorite Lightroom too for extended dynamic range work, especially for architecture and real estate work where pleasing and realistic color is preferred over the acidic candy colored distorted look is Timothy Armes' Enfuse plug-in from http://photographers-toolbox.com/products/lrenfuse.php
Armes also recommends this e-book http://photographyforrealestate.net/enfuse/ by Simon Maxwell. After reading for the past few days I concur with his assessment.
-
Interesting and worth investigating, especially as I have LR6 anyway. The Photo Merge to HDR offering in LR6 is "ok" but it does not really offer the degree of fine tuning that I would like. With Photomatix one does have to reduce the suggested defaults by a considerable degree and my suspicion is that many just don't bother. As presented, they can produce some OTT results. I will have a closer look at your suggested Plug In. When shooting interiors where you are presented with a host of natural and artifical lighting, adjusting the Colour Temperature of your Frames can be very important. I suspect that many individuals shoot on Auto White Balance and do not alter the sometimes variable and unrealistic Colour Temperatures that the camera selects. I try to remember the type of stone used and work with that as a basis for any alterations.
-
Here's one that may surprise you. Six RAW frames shot in 2013 from a Panasonic LUMIX DMC LX3 camera inverted on a tripod. The 0EV was shot at 5.1mm, ISO 125, f/2 and 1/30sec. Many people would write off this little camera as too inexpensive to be any good. I would love the to try the LX100 with the larger Sensor. The RAW conversions and Geometric corrections were made in LR6 and I then used the Photo Merge to HDR utility. This is the beautifuiiy vaulted ceiling of the 12th century abbey at Malmesbury in Wiltshire, England. The 16-Bit Master File holds up quite well to a close examination @ 100% and probably would print well.
-
Hi,
My personal experience is that two or three exposures, three to four EV apart, can be quite enough. I am pretty sure that has also been indicated by Jeff Schewe and Eric Chan.
I don't do HDR very often, but I found that the in Lightroom 6 does a very good job. LumaRiver HDR is also a program that I found much to my liking.
Best regards
Erik
-
Thanks Erik, I will have a look at your recommended HDR software.
-
Thanks to everyone for their feedback.
I did some extensive testing while shooting my projects. I mainly used the Canon 5D mark 3 files and on some occasions the Sony a7r. All the files were processed using Adobe Photoshop CS6 Merge to HDR tool in 32bit mode and saved as tiff.
I think as most have pointed out there is no set number of pictures that will achieve the best results. I found that exterior architecture pictures in sunset do not require more than 7 stops range. On the other hand some interior spaces required 11 stops range to capture the difference between shadows and inside and highlights out side in the view.
Most often I required more negative exposures than bright ones. For example with the exterior picture I did not need more than +2 for the shadows but needed -4 to tame down the bright light sources.
One interesting observation was that I found adding too many bright exposures caused banding in white walls. Sometimes adding extra bright exposures is needed because the shadows are too dark and the extra information from those files will improve them. The solution is to create two HDR files and blend different parts with masks in photoshop.
All in all I find my self shooting from 7 to 11 exposures to capture the entire range and in 1.3-1 stops increments to insure the smoothest results.
Increments wise, there isn't really much difference between 1.3stops and 1stop. You can shoot 7 1.3stops exposures and obtain the same results from 9 1stop exposures. I just find shooting single exposures easier to calculate. However, 1.6 and 2stops provides less clean results. For example, with equal exposure range you will be able to brighten shadows to the same extent but the shadows will be a little less smooth.
-
All in all I find my self shooting from 7 to 11 exposures to capture the entire range and in 1.3-1 stops increments to insure the smoothest results.
Increments wise, there isn't really much difference between 1.3stops and 1stop. You can shoot 7 1.3stops exposures and obtain the same results from 9 1stop exposures. I just find shooting single exposures easier to calculate. However, 1.6 and 2stops provides less clean results. For example, with equal exposure range you will be able to brighten shadows to the same extent but the shadows will be a little less smooth.
Can you show an example that demonstrates that the shadows are less smooth when shooting at 2 stops invervals than when doing it at 1 stop intervals? or it is just a perception you have?.
On the other hand, more shots also mean a higher risk of losing sharpness. Unless your images are perfectly aligned, something very difficult to achieve even if you align them in software (real world is not subject to perfect 1-pixel shifts), the more shots you do the more areas in the final composite will be provided by a blending of 2 or more shots, and these progressive areas will loose sharpness.
This is a sample of a HDR machine-gun-shooter guy trying to convince others that the more shots the better. He provided an image where using 7 shots was less sharp than using just 3 shots:
(http://www.guillermoluijk.com/misc/nitidez.gif)
Regards!
-
...You'll not only get the same quality with 7 pictures at 1,6 intervals than with 11 at 1 stop intervals, but the same with 4 pictures at 3 stops intervals.
http://www.sns-hdr.com/forum/viewtopic.php?f=7&t=4138
The gradients are less precise when the step is large. I do not recommend step greater than 2ev.
-
I saw some paper where they estimated the DR of some lens on a Canon 20D to be limited to 10-20 stops due to flare, depending on the size of the bright patch (a single bright dot giving the largest DR).
When your lens is limiting DR, exposure bracketing wont fix it.
-h
-
http://www.sns-hdr.com/forum/viewtopic.php?f=7&t=4138
specifics of that particular application ? did you notice bad gradients with Fuji raws where masked sensels were separated by more than "step greater than 2ev" ?
-
Hi,
There was a thread discussing that issue. Iliah Borg said that captures would be limited to 11EV due to lens flare. I doubt that figure is absolute, but 11 EV essentially lies in line with what I can see in RawDigger in most of mine high luminance range of images.
The included single exposure image is quite OK, but noise in the darks is a bit high, still manageable with noise reduction. HDR version gives better shadow detail. Shadow detail is often not very visible so it can often take a lot of noise reduction.
Best regards
Erik
I saw some paper where they estimated the DR of some lens on a Canon 20D to be limited to 10-20 stops due to flare, depending on the size of the bright patch (a single bright dot giving the largest DR).
When your lens is limiting DR, exposure bracketing wont fix it.
-h
-
Hi,
My take is really that we need one image exposed for the highlights, with essentially no clipping. Raw processors don't give useful indications on clipping as they can apply both curves and tone mapping to the histogram. Using a tool like RawDigger gives accurate info. Personally I find that in camera histograms are accurate enough for me.
Than I need at least one exposure for good detail in the darks. I would say Guillermo has a good point on having fewer images may be better. I would also guess that it can depend on software. Also, having more exposures can help with ghost elimination but also create more ghosts to eliminate.
Software may also matter. I mostly used the HDR built in Lightroom, but I also have Lumariver HDR and of course the one in Photoshop.
Best regards
Erik
Thanks to everyone for their feedback.
I did some extensive testing while shooting my projects. I mainly used the Canon 5D mark 3 files and on some occasions the Sony a7r. All the files were processed using Adobe Photoshop CS6 Merge to HDR tool in 32bit mode and saved as tiff.
I think as most have pointed out there is no set number of pictures that will achieve the best results. I found that exterior architecture pictures in sunset do not require more than 7 stops range. On the other hand some interior spaces required 11 stops range to capture the difference between shadows and inside and highlights out side in the view.
Most often I required more negative exposures than bright ones. For example with the exterior picture I did not need more than +2 for the shadows but needed -4 to tame down the bright light sources.
One interesting observation was that I found adding too many bright exposures caused banding in white walls. Sometimes adding extra bright exposures is needed because the shadows are too dark and the extra information from those files will improve them. The solution is to create two HDR files and blend different parts with masks in photoshop.
All in all I find my self shooting from 7 to 11 exposures to capture the entire range and in 1.3-1 stops increments to insure the smoothest results.
Increments wise, there isn't really much difference between 1.3stops and 1stop. You can shoot 7 1.3stops exposures and obtain the same results from 9 1stop exposures. I just find shooting single exposures easier to calculate. However, 1.6 and 2stops provides less clean results. For example, with equal exposure range you will be able to brighten shadows to the same extent but the shadows will be a little less smooth.
-
link=topic=101682.msg843983#msg843983 date=1439322325]
http://www.sns-hdr.com/forum/viewtopic.php?f=7&t=4138
"The gradients are less precise when the step is large. I do not recommend step greater than 2ev."
Sebastian, SNS-HDR's author, is surely right just by adding "when using SNS-HDR" to his claim. This also happens with Photomatix and other blending algorithms like Enfuse's pyramids. But this is a fault of those pieces of software, which are sub-optimum at the blending stage doing quite extensive progressive blending, it is not a general truth about bracketing fusion.
Sub-optimum progressive blending algorithms have two consequences: they fail to get the best available information from the input RAW files so they need more shots for the same output blending quality, and they loose sharpness when a minimum missalignment between shots persists.
With modern sensors (and not so modern as my Canon 350D's), gradients are fine at 4 stops intervals, although I prefer not to go that far and bracket at 3 stops intervals.
Regards