The Great Single or Multiple RAW Exposure Debate

Nikolai

David_S85 wrote:

CS2 has 32-bit HDR. Question: Has the HDR function improved in CS3?

They say it did, but I didn't play with it much myself. Thus far having a single RAW developed for different exposure settings and following blend in PS pretty much served up all my needs. For instance, you would hate to see the original of the following image:

Yet just two different exposure from ACR made a decent image out of it (I hope so at least).

Having said that - everything else they did say they have improved they really did, so I don't have any reason to doubt HDR was an exception...

Michiel de Brieder

Nikolai wrote:

For instance, you would hate to see the original of the following image:

OT:
Dear Nik,

That is a *georgeous* image!!!!!!!!! It is THAT good that it doens't even matter that the path is centered and walks away straight down the middel, it just fits!

Cheers mate hope everything is okay in Socal!

Nikolai

Michiel de Brieder wrote:

OT:
Dear Nik,

That is a *georgeous* image!!!!!!!!! It is THAT good that it doens't even matter that the path is centered and walks away straight down the middel, it just fits!

Cheers mate hope everything is okay in Socal!

WOW! Look's who decided to show up!
Long time no see mate!

Appreciate the sentiment:-)!

Let's go to Sid's place and chat :-)

rutt

Nik, I've been claiming that it's never necessary to do two different conversions from raw. So far, I haven't failed a challenge. If you give me access to your raw file I'll see if I can prove that again.

Andy

rutt wrote:

Nik, I've been claiming that it's never necessary to do two different conversions from raw. So far, I haven't failed a challenge. If you give me access to your raw file I'll see if I can prove that again.

This is a great challenge to throw down to our Artist-in-Residence Marc Muench

rutt

Andy wrote:

This is a great challenge to throw down to our Artist-in-Residence Marc Muench

Understand, it's a different matter to merge more than one shot to get HDR. I'm just saying that if you have a single RAW with all the info, you can convert just once, not lose any of that info, and use PS to recover highlight/shadows and merge, whatever. And (as least as of CS2) the tools in PS are much more mature and complete then the ACR tools.

Andy

rutt wrote:

Understand, it's a different matter to merge more than one shot to get HDR. I'm just saying that if you have a single RAW with all the info, you can convert just once, not lose any of that info, and use PS to recover highlight/shadows and merge, whatever. And (as least as of CS2) the tools in PS are much more mature and complete then the ACR tools.

buk buk buk

rutt

Andy wrote:

buk buk buk

No capiche "buk". But I'm guessing it can't be a compliment.

Anyway, this business about multiple ACR conversions vs PS methods is a hijack. Moderators, maybe you can put it in it's own thread, possibly starting with Nik post and great photo.

Look, I'm not claiming that I can do better than the someone using multiple conversions, just that I can do as well. I'm not claiming my method will be easier for everyone, just for me (and maybe for some others.) People like to work in different ways. I'm bad with brushes and like to think of ways to avoid them. I know some basic applied color theory and can figure out how to do a lot of things from first principles. On the other hand, I don't know the filter set at all, for example. I know that's a lot more natural for some people (Shay?)

So I'm not claiming I know the one true way. Just that I know a way.

Andy

rutt wrote:

No capiche "buk". But I'm guessing it can't be a compliment.

Anyway, this business about multiple ACR conversions vs PS methods is a hijack. Moderators, maybe you can put it in it's own thread, possibly starting with Nik post and great photo.

Excellent idea

Look, I'm not claiming that I can do better than the someone using multiple conversions, just that I can do as well. I'm not claiming my method will be easier for everyone, just for me (and maybe for some others.) People like to work in different ways. I'm bad with brushes and like to think of ways to avoid them. I know some basic applied color theory and can figure out how to do a lot of things from first principles. On the other hand, I don't know the filter set at all, for example. I know that's a lot more natural for some people (Shay?)

So I'm not claiming I know the one true way. Just that I know a way.

:rutt

rutt

Actually, there are two issues:

1. Multiple vs one exposure. Until we have true HDR cameras, I have to admit that there are times when only multiple exposure will do.
2. Multiple vs one raw conversion from a single shot. Some people are quite addicted with multiple conversions and some sort of blending from multiple images in PS. I know this method can work and work well. But I believe that I can always do just about as well with a single conversion, begin careful not to lose any data to clipping or the default S curves.

I shot a stained glass window last winter for the Xmas card of a local church (Tiffany and they are very proud of it.)



This was merged from 5 separate shots. No reshoot possible. But now I wish I'd used some sort of bright backlight so that I never had to use the longest exposures which show too much glass surface.

Gives you a whole new respect for those books you can buy at European cathedrals with the great pictures of those windows.

LiquidAir

rutt wrote:

Understand, it's a different matter to merge more than one shot to get HDR. I'm just saying that if you have a single RAW with all the info, you can convert just once, not lose any of that info, and use PS to recover highlight/shadows and merge, whatever. And (as least as of CS2) the tools in PS are much more mature and complete then the ACR tools.

I'm with Rutt on this one. Do one conversion in ACR to 16 bits making sure you don't lose any dynamic range and you can get the same effect in Photoshop as you would get with multiple conversions. When you convert a HDR file to 16 bits in Photoshop you have the option for some local contrast enhancement (sometimes called HIROLAM sharpening) as part of the conversion; I think that is primarly what gives HDR its distinct look. However, you can get that same effect with USM if you know what you are doing.

To me, an image only really counts as HDR if you capture more dynamic range than the sensor on your camera. Currently digital SLRs typically capture about 8 stops of dynamic range. Combining three bracketed exposures can expand that range to 10 stops; 6 exposurs will give you 13 stops. Multiple conversions from ACR will never give you more than the 8 stops your sensor captured.

Remember that no how many stops you capture, a print will never have more than about 6 stops (if that). A simple S curve is the easiest way to compress the 8 stops (or more if you use HDR) from your camera to the printed 6 stops. Over-comressing the global contrast and then enhancing local contrast with USM or masking is a different approach which works well with some kinds of images and typically gives the "look" normally associated HDR.

I can try to put together a tutorial on local contrast enhancement if anyone is interested. It'll take me a week or two, I have family visiting at the moment so I don't have much spare time.

jfriend

We need images and technique descriptions to try

rutt wrote:

Look, I'm not claiming that I can do better than the someone using multiple conversions, just that I can do as well. I'm not claiming my method will be easier for everyone, just for me (and maybe for some others.) People like to work in different ways. I'm bad with brushes and like to think of ways to avoid them. I know some basic applied color theory and can figure out how to do a lot of things from first principles. On the other hand, I don't know the filter set at all, for example. I know that's a lot more natural for some people (Shay?)

So I'm not claiming I know the one true way. Just that I know a way.

Technically, this seems to come down to two things:

1) Does a RAW image actually contain more data than can be expressed in a 16-bit TIFF and is that extra data valuable in PP.
2) If both ways can work equally well, which way is easier and faster for any given person?

I like to stay away from religious arguments of RAW vs. non-RAW and just focus on the quality of the result and the speed of getting the result.

Further, we could probably argue both of these questions in the "theoretical" for years. Things will only really be settled for me when I've tried a bunch of images both ways and compared the results both in terms of quality and speed.

With my current skill set, I know that certain types of images are easier and faster for me to get quality results by combining two RAW developments from the same image. I don't know if that is because the RAW image has superior data to work with or I just haven't learned the proper techniques for doing it in CS2 from a single 16-bit image.

So, let's bring on some example images and bring on some descriptions of good technique for both ways so we can all play and form our own opinions/arguments.

Let the games begin.

Duffy Pratt

I'm with Rutt on this one. At least with the Raw developers I have, an old version of ACR and DPP, I can do much more with PS techniques than with two different raw conversions. I've seen people get very good results with multiple conversions, but I don't think I've ever seen anything done with multiple conversions that I thought could not have been done with some basic PS techniques (most notable using false profiles).

Duffy

rdlugosz

My general opinion is that all techniques have their place and the more you know then the more options you'll have when trying to find the "right tool for the job".

Note that with the Fill Light and Recovery sliders in the new Lightroom/ACR4 I've found it far less often that I need to do a blend of two RAW exposures. I believe the algorithms are doing something similar behind the scenes on your behalf.

Nikolai

rutt wrote:

Nik, I've been claiming that it's never necessary to do two different conversions from raw. So far, I haven't failed a challenge. If you give me access to your raw file I'll see if I can prove that again.

Oops, somehow I lost subscription to this thread and haven't seen its development. Sorry...

John, I'll send you a link to the RAW file in question later... If I don't please remind me...

Apart from that... It is fairly possible one can do without multiple exposures. However, just like in case with i2e, if it saves me 95% of the time while only losing 5% of the ultimate quality, I think I'm game...