HDR: bracketing in camera vs. bracketing in post

SloYerRoll

I've taken some interest in HDR lately and have a simple question that will probably branch into many more.

With all the exposure latitude that RAW files have. Is there that much of a difference between bracketing in camera versus adjusting exposure in post with a RAW file?

I wouldn't mind taking three+ shots to get a good HDR image. but if I can take one well exposed shots and stretch it in post to make three+ shots covering the entire dynamic range of the image...

Thoughts?

BTW:
I shoot with a D200 and D300 if it makes any difference.

BPerron

I think you are talking about just changing the exp. in PP in something like PS and getting the different exposure that way...if that is the case then...

In order for HDR to work you need to take at least 3 different bracketed shots with the camera...Simply by adjusting the exp. in PS will not yeild much...cause you have already captured the only detail, highlights and shadows that is possible at the exp...it would be like saying welll niose does not matter cause I can just fix it in PP...

I hope this helps...let me know if you have any other questions.

dgrinjackson

I'm certainly not an expert, but I think the Noise example doesn't really correlate.

With noise, you actually have lost all the data in that specific pixel(s).

Assuming you're talking RAW PP....If you take a properly exposed photograph and you see that you have not blown out the pixels or have underexposed the original shot so that the pixel(s) have gone completely black then you have not lost any of the information on any of the pixels.

When you bump up or down the original photo by one stop and you see that, again, no pixels have lost information by being blown out or by having gone completely black then I don't see difference between bracketing the exposure in the camera or bracketing the exposure in PP.

Regardless you still get three images at +1, 0, -1 exposure and you have not lost any of the data/information in any of the pixels.

At least that is how I see it, but I might be missing something.

BPerron

dgrinjackson wrote:

I'm certainly not an expert, but I think the Noise example doesn't really correlate.

With noise, you actually have lost all the data in that specific pixel(s).

Assuming you're talking RAW PP....If you take a properly exposed photograph and you see that you have not blown out the pixels or have underexposed the original shot so that the pixel(s) have gone completely black then you have not lost any of the information on any of the pixels.

When you bump up or down the original photo by one stop and you see that, again, no pixels have lost information by being blown out or by having gone completely black then I don't see difference between bracketing the exposure in the camera or bracketing the exposure in PP.

Regardless you still get three images at +1, 0, -1 exposure and you have not lost any of the data/information in any of the pixels.

At least that is how I see it, but I might be missing something.

My point with the niose example is that it is much better to get what you want from the camera not from post processing same goes for HDR...I will find the information on this, just might take a min.

ziggy53

Look at it this way, with a singular RAW file you will never have more HDR than any RAW file converter can yield. Many RAW converters "will" allow the compression of both highlight and shadow to yield something like a rudimentary HDR in 8 bit output.

True HDR goes beyond this to create not just compressed highlights and shadows, but highlights and shadows mapped into previously visually unavailable regions of the image.

The best HDR software does indeed import RAW files as part of the mapping process, so that the resulting images contain (potentially) much more tonal information in both crushed shadows and blown highlights than you could ever create from a single RAW alone.

That said, I do use singular RAW files to allow some recovery of both shadow and highlight when required, which is very much like an HDR treatment.

ziggy53

I almost forgot, RAW files may not contain very well balanced color tones in the deep shadows and bright highlights, especially in white balance outside of daylight and flash. The reason for that is because of how the Bayer filters and RAW conversions work and that they are both usually optimized for daylight capture (5500K or so).

HDR should create more accurate color tones in those extremes than compressed singular RAW can provide.

SloYerRoll

Thanks for all the comments. Your comments confirm my thoughts. But why not ask

ziggy53 wrote:

I almost forgot, RAW files may not contain very well balanced color tones in the deep shadows and bright highlights, especially in white balance outside of daylight and flash. The reason for that is because of how the Bayer filters and RAW conversions work and that they are both usually optimized for daylight capture (5500K or so).

HDR should create more accurate color tones in those extremes than compressed singular RAW can provide.

Thank you very much for this post Ziggy. Rendering colors acurately is very important to me, so this little nugget prooves especially useful

SloYerRoll

I know the answer to the following question can be preferential. But I'm giving photomatix a shot for HDR software. Any other contenders out there worth checking out? I like what I've seen so far as far as output. But allot of my shots are still in my head so I don't know how photomatix will handle them

LiquidAir

Typical RAW files from modern DSLRs capture a bit over 8 stops of dynamic range. In my experience the RAW files contains between half and a full stop more dynamic range than JPEG generated by a conventional RAW conversion (i.e. no hight recovery or fill light in ACR) and occasionally I'll find that multiple RAW conversions from the same file is the easiest way to recover that extra range. However, if the scene you are trying to capture has more than 8 stops from the highlights to the shadows you must use multiple exposures to fully capture shadow detail without blowing out highlights and, even if you don't need more than 8 stops, you can often get lower shadow noise and better color rendition in both the shadows and highlights if you start with multiple exposures.

ziggy53

SloYerRoll wrote:

I know the answer to the following question can be preferential. But I'm giving photomatix a shot for HDR software. Any other contenders out there worth checking out? I like what I've seen so far as far as output. But allot of my shots are still in my head so I don't know how photomatix will handle them

I tried Photomatix but wound up purchasing "Dynamic-Photo HDR" because I liked the interface better, it can do "pseudo" HDR and most importantly because I like the HDR effect better from it. The process is a bit unique, but once you understand what it wants and how it works, it is really easy to operate.

BPerron

LiquidAir wrote:

Typical RAW files from modern DSLRs capture a bit over 8 stops of dynamic range. In my experience the RAW files contains between half and a full stop more dynamic range than JPEG generated by a conventional RAW conversion (i.e. no hight recovery or fill light in ACR) and occasionally I'll find that multiple RAW conversions from the same file is the easiest way to recover that extra range. However, if the scene you are trying to capture has more than 8 stops from the highlights to the shadows you must use multiple exposures to fully capture shadow detail without blowing out highlights and, even if you don't need more than 8 stops, you can often get lower shadow noise and better color rendition in both the shadows and highlights if you start with multiple exposures.

Beat me to it...Thanks for saying what I was trying to say...it has been a very long day....

Slinky0390

I've always been under the impression and I've always read that in-camera bracketing is better than changing the exposure of a single image. Like someone said, with a single image, you've already captured as much detail as you could, so changing the exposure doesn't do much. When you change the exposure in the camera, you expose different things. When you have lets say -2, you'll pick up more of the darker details, but when you're +2 you'll get the brighter ones. Essentially, you're covering a greater dynamic range, which is what HDR is. Someone correct me if I'm wrong.

BPerron

Slinky0390 wrote:

I've always been under the impression and I've always read that in-camera bracketing is better than changing the exposure of a single image. Like someone said, with a single image, you've already captured as much detail as you could, so changing the exposure doesn't do much. When you change the exposure in the camera, you expose different things. When you have lets say -2, you'll pick up more of the darker details, but when you're +2 you'll get the brighter ones. Essentially, you're covering a greater dynamic range, which is what HDR is. Someone correct me if I'm wrong.

Nope you are exactly correct. I believe most single exposure captures between 6-8 EV's while an HDR merged will cover about 14-16 EV's. Also you are editing in 32 bit floating point which gives you a better platform to PP with.

ziggy53

Slinky0390 wrote:

I've always been under the impression and I've always read that in-camera bracketing is better than changing the exposure of a single image. Like someone said, with a single image, you've already captured as much detail as you could, so changing the exposure doesn't do much. When you change the exposure in the camera, you expose different things. When you have lets say -2, you'll pick up more of the darker details, but when you're +2 you'll get the brighter ones. Essentially, you're covering a greater dynamic range, which is what HDR is. Someone correct me if I'm wrong.

I know what you are thinking and I agree wih the comment, "Essentially, you're covering a greater dynamic range, which is what HDR is."

I suggest that if you try to shoot with an FEC of -2 you will retain more detail and tonality in the "highlights", and FEC +2 will retain more detail in the "shadows" (in the final composite or HDR, whichever technique works to your effect).

SloYerRoll

BPerron wrote:

Nope you are exactly correct. I believe most single exposure captures between 6-8 EV's while an HDR merged will cover about 14-16 EV's. Also you are editing in 32 bit floating point which gives you a better platform to PP with.

So your saying both jpeg + RAW images provide the same amount of stop range for a single exposure?

Read back to my original post. I think you guys confused to what I asked.

jfriend

SloYerRoll wrote:

So your saying both jpeg + RAW images provide the same amount of stop range for a single exposure?

Read back to my original post. I think you guys confused to what I asked.

Your answer is right in LiquidAir's excellent post #9. A single RAW file contains more dynamic range than a single JPEG file. Multiple RAW files with bracketing contain more dynamic range than a single RAW file. So, it all depends upon how much dynamic range your scene has and you need to capture. If you need to capture 10 stops of range, then you need multiple and bracketed RAW files. If you only need a little more than a JPEG has, then sometimes a single RAW file will do.

Also, the quality of the information in the darker tones can be noticably higher (more detail and less noise) when using multiple bracketed shots rather than a single RAW file. That's because the bracketed multlple shots can have one shot specifically exposed for maximum detail and lowest noise in the shadows whereas the single shot must "push" the shadows in order to try to recover detail there. This "pushing" results in lower quality shadows than a bracketed exposure just for shadows.

So, the real answer is it depends upon how much dynamic range you need to record and how important low noise and detail is in the shadow areas. I regularly blend multiple developments from a single RAW file to enhance my post processing, but when I'm thinking about it in the field and have time to compose on a tripod, I bracket because it can yield higher quality and more dynamic range. One is a shortcut that can still yield improved results, the other is the way to get the best possible result.

SloYerRoll

jfriend wrote:

Your answer is right in LiquidAir's excellent post #9. A single RAW file contains more dynamic range than a single JPEG file. Multiple RAW files with bracketing contain more dynamic range than a single RAW file. So, it all depends upon how much dynamic range your scene has and you need to capture. If you need to capture 10 stops of range, then you need multiple and bracketed RAW files. If you only need a little more than a JPEG has, then sometimes a single RAW file will do.

Also, the quality of the information in the darker tones can be noticably higher (more detail and less noise) when using multiple bracketed shots rather than a single RAW file. That's because the bracketed multlple shots can have one shot specifically exposed for maximum detail and lowest noise in the shadows whereas the single shot must "push" the shadows in order to try to recover detail there. This "pushing" results in lower quality shadows than a bracketed exposure just for shadows.

So, the real answer is it depends upon how much dynamic range you need to record and how important low noise and detail is in the shadow areas. I regularly blend multiple developments from a single RAW file to enhance my post processing, but when I'm thinking about it in the field and have time to compose on a tripod, I bracket because it can yield higher quality and more dynamic range. One is a shortcut that can still yield improved results, the other is the way to get the best possible result.

Thanks John. I acknowledged LA's (as usual) great answer. I think the thread lost integrity after the answer was already given

Thanks for that info. More good stuff to digest. I'm "evolving" into a tripod shooter more and more. So the bottom line if it is that I should get the exposures in the field and give my HDR software even more to work with. Even if I didn't use a tripod. the align layers command in PS is a great help in situations like this.

Thanks again.

BPerron

SloYerRoll wrote:

Thanks John. I acknowledged LA's (as usual) great answer. I think the thread lost integrity after the answer was already given

Thanks for that info. More good stuff to digest. I'm "evolving" into a tripod shooter more and more. So the bottom line if it is that I should get the exposures in the field and give my HDR software even more to work with. Even if I didn't use a tripod. the align layers command in PS is a great help in situations like this.

Thanks again.

I am not sure where the post lost integrity...More useful info. was given, like being able to tone map and work in 32 bit, as opposed to 8-12, or even 16 bit...which is a great thing as well

Another thing that is great about HDR is that because you are combining bracketed shots you will get some noise reduction

Another thing about bracketing for HDR is that even if you do not convert to HDR with the all the shots, you will still have a single good exposure...

In any event, I can not wait to see pics of your HDR images.

BigAl

I often use DavidTo's tut on blending two images from a single RAW. It's actually easier to do this in Paintshop Pro than David's tut for PS.

To do exposure bracketing in the camera you need a *very* steady tripod, and that is not always available.

In this pic, the grass was *totally* blown out, but rescued by using the tut.

BPerron

BigAl wrote:

To do exposure bracketing in the camera you need a *very* steady tripod, and that is not always available.

Not really true...it helps alot but is not a must if you camera will do bracketing on its own...

Handheld HDR http://www.flickr.com/photos/smgallery/2280652846/
done by smgallery on Flickr...

Just do a search on flickr for handheld HDR and it wil bring up pages of handheld HDR.

BigAl

BPerron wrote:

Not really true...it helps alot but is not a must if you camera will do bracketing on its own...

Handheld HDR http://www.flickr.com/photos/smgallery/2280652846/
done by smgallery on Flickr...

Just do a search on flickr for handheld HDR and it wil bring up pages of handheld HDR.

There is a time lag between shots and if you don't have a steady hand, you won't have the pics lining up when you do the HDR, and you will end up with a rather soggy mess (been there, got the t-shirt).

BPerron

BigAl wrote:

There is a time lag between shots and if you don't have a steady hand, you won't have the pics lining up when you do the HDR, and you will end up with a rather soggy mess (been there, got the t-shirt).

I understand that there is lag...but like I said...do a search for handheld HDR and you will see that is very possible to create one with no problem...You just need to be able to hold still. The align images in photomatix and PS are very good at taking care of the differences between the photos...so it can be done and there are hundreds if not thousands of photos around the net that prove this.

SloYerRoll

BPerron wrote:

I understand that there is lag...but like I said...do a search for handheld HDR and you will see that is very possible to create one with no problem...You just need to be able to hold still. The align images in photomatix and PS are very good at taking care of the differences between the photos...so it can be done and there are hundreds if not thousands of photos around the net that prove this.

You can hand hold. But for those tack sharp shots that are properly exposed. I'm discovering you need a solid tripod.

BPerron

SloYerRoll wrote:

You can hand hold. But for those tack sharp shots that are properly exposed. I'm discovering you need a solid tripod.

For sure...that goes for most photography though...tripods are usually better not as convient, but oh well

Osprey

One benifit
There is one benifit to Pseudo HDR and that is the images that are to be combined are identical to each other spatially There is no movement which causes ghosting sometimes.

SloYerRoll

ANother thing that can help this is make sure you only adjust your shutter speed so your DOF doesn't change.

Slinky0390

but i found it better to put the camera into aperture priority (AV for Canon users). In regards to tripod vs no tripod, it all depends on how much light there is, if its a darker situation, handheld will be near impossible. Nothing to do with the thread, but this smilie is so damn funny :deadhorse

SloYerRoll

I don't know.. If this simple thread goes on much longer..

BigAl

SloYerRoll wrote:

I don't know.. If this simple thread goes on much longer..

Well, I'll give it a shot

Consider the 40d. In AEB mode, you should ideally also have the camera in burst mode - let's say the fast burst mode with 6.5 frames per sec. That means a delay of 0.154s per shot. In keeping with SloYerRoll, we want to keep the DoF constant, so we will shoot in Av mode, let's say F16 for a nice deep depth of field. For sunny F16 that means 1/125s for the base shot. For ease of calculation, let's use 1 stop AEB which means 1/250s for the underexposed shot and 1/60s for the overexposed shot. So, let's take our picture and according to the manual, the order is standard, delay, under, delay, then over:
1/125 + 1/6.5 + 1/250 + 1/6.5 + 1/60
=0.008+0.154+0.004+0.154+0.017
=0.337s

which means you need to keep that camera steady for more 1/3s to give the software less to do

Am I earning my keep at keeping the thread going ?

BTW, I've just started a new gallery where all the shots will be manipulated using DavidTo's tut mentioned in my first post.