EasyEd
9th of February 2009 (Mon), 00:26
Hey All,
OK so I been studyin on these HDRs. What I would like you all to comment on is do I understand this correctly? If not please correct me. I've also embedded a few questions.
Bur first - with respect to this HDR stuff and photography in general - I realize that ultimately it comes down to what you like to photograph and what you like the end product to - generally speaking - look like. In my case what I like to photograph is machinery and landscapes - not so much people or cityscapes or seascapes although like most people who carry a camera lots of different subject matter(s) get photographed. That said I strongly like realistic with additional detail to slightly overdone HDRs. Simply put I seek Ansel Adams style photography in colour - not so much in terms of subject matter - but in terms of composition, detail, texture and light - without being distracted by colour or having colour "flatten" the whole image. Is it so much to ask? You might consider my questions in light of this. PS - For now I use a G9.
And second - I understand that a digital image is nothing but a rectaangular series of small squares called pixels that have 3 numbers associated with each pixel - a number representing red an number representing green and a number representing blue - basically RGB. It is the variation in these numbers combined with the juxtapositioning of pixels next to each other that gives us pictures. I know this from processing NASA satellite (TM) data in real life.
So with respect to any potential HDR I think you have three major considerations - and a fourth issue which I've also touched on.
First - Exposure. It seems to me that the more contrast there is in a scene the greater the range of exposures you should take in order to capture the detail (meaning fewer 0,0,0 - black and/or 256,256,256 - white pixel values or do I have these backwards?) in the scene - and the greater the range of exposures the greater the number you should take. True? I suspect that for relatively low contrast scenes 7 or 9 or more exposures probably buys you very very little if anything. High contrast scenes more exposures is likely better - in order to capture detail in the scene. So a question I have is how do I know if a scene I'm shooting is high contrast enough to justify a wide range of EVs? The histogram? If so what is my clue? Single peaks? Multiple peaks? How tall the peaks are relative to each other? What?
Second - combining multiple EV exposures. Since all images are just pixels with unique RGB values the combining of different EVs is nothing but an exercise in math. I suspect there are a wide variety of math formulas for combining different EVs - probably ranging from simple averaging to various other mathmatical methods. My question is simple - what software has the widest set of options or perhaps more realistically what 2 programs give the widest set of options for combining different EVs? Does the number of EVs used matter with respect to what software you use?
Third - tone mapping. I don't know how to describe this however I've no doubt that it is nothing but more mathmatical manipulation of the pixel RGB values in an image. Clearly there are likely multiple tone mapping strategies. I suspect that some tone mapping strategies lead to more "realistic" images while some lead to more "overdone" images. Is there a software package or pair of software packages that give you the best of both worlds? Does any particular software package tend to do "realistic" better or "overdone" better? Is it better to use a program that allows you to use curves like FDRTools versus one that apparently does not like Photomatix? Does it matter?
Fourth - PseudoHDRs. I understand that a pseudoHDR derived from a single image is not a true HDR because you cannot add any information (meaning more detail in the very light or dark areas) from the original scene by re-exposing the RAW. What you got is all you got - because you cannot change an RGB value of 0,0,0 (black) or 256,256,256 (white) for any particular pixel in a picture to an accurate representation of what information that pixel might have contained if you had either over or under exposed it when you took an original set of pictures at different EVs. That said pseudoHDRs still can represent human eye reality very well and/or be very artistic. My question is - is there a tone mapping package that works best for creation of single image based pseudoHDRs? Are there any tone mapping only software packages?
TIA for any and all replies! They will certainly help in my understanding of HDR digital photography. It might help if you quote the part your responding to.
-Ed-
OK so I been studyin on these HDRs. What I would like you all to comment on is do I understand this correctly? If not please correct me. I've also embedded a few questions.
Bur first - with respect to this HDR stuff and photography in general - I realize that ultimately it comes down to what you like to photograph and what you like the end product to - generally speaking - look like. In my case what I like to photograph is machinery and landscapes - not so much people or cityscapes or seascapes although like most people who carry a camera lots of different subject matter(s) get photographed. That said I strongly like realistic with additional detail to slightly overdone HDRs. Simply put I seek Ansel Adams style photography in colour - not so much in terms of subject matter - but in terms of composition, detail, texture and light - without being distracted by colour or having colour "flatten" the whole image. Is it so much to ask? You might consider my questions in light of this. PS - For now I use a G9.
And second - I understand that a digital image is nothing but a rectaangular series of small squares called pixels that have 3 numbers associated with each pixel - a number representing red an number representing green and a number representing blue - basically RGB. It is the variation in these numbers combined with the juxtapositioning of pixels next to each other that gives us pictures. I know this from processing NASA satellite (TM) data in real life.
So with respect to any potential HDR I think you have three major considerations - and a fourth issue which I've also touched on.
First - Exposure. It seems to me that the more contrast there is in a scene the greater the range of exposures you should take in order to capture the detail (meaning fewer 0,0,0 - black and/or 256,256,256 - white pixel values or do I have these backwards?) in the scene - and the greater the range of exposures the greater the number you should take. True? I suspect that for relatively low contrast scenes 7 or 9 or more exposures probably buys you very very little if anything. High contrast scenes more exposures is likely better - in order to capture detail in the scene. So a question I have is how do I know if a scene I'm shooting is high contrast enough to justify a wide range of EVs? The histogram? If so what is my clue? Single peaks? Multiple peaks? How tall the peaks are relative to each other? What?
Second - combining multiple EV exposures. Since all images are just pixels with unique RGB values the combining of different EVs is nothing but an exercise in math. I suspect there are a wide variety of math formulas for combining different EVs - probably ranging from simple averaging to various other mathmatical methods. My question is simple - what software has the widest set of options or perhaps more realistically what 2 programs give the widest set of options for combining different EVs? Does the number of EVs used matter with respect to what software you use?
Third - tone mapping. I don't know how to describe this however I've no doubt that it is nothing but more mathmatical manipulation of the pixel RGB values in an image. Clearly there are likely multiple tone mapping strategies. I suspect that some tone mapping strategies lead to more "realistic" images while some lead to more "overdone" images. Is there a software package or pair of software packages that give you the best of both worlds? Does any particular software package tend to do "realistic" better or "overdone" better? Is it better to use a program that allows you to use curves like FDRTools versus one that apparently does not like Photomatix? Does it matter?
Fourth - PseudoHDRs. I understand that a pseudoHDR derived from a single image is not a true HDR because you cannot add any information (meaning more detail in the very light or dark areas) from the original scene by re-exposing the RAW. What you got is all you got - because you cannot change an RGB value of 0,0,0 (black) or 256,256,256 (white) for any particular pixel in a picture to an accurate representation of what information that pixel might have contained if you had either over or under exposed it when you took an original set of pictures at different EVs. That said pseudoHDRs still can represent human eye reality very well and/or be very artistic. My question is - is there a tone mapping package that works best for creation of single image based pseudoHDRs? Are there any tone mapping only software packages?
TIA for any and all replies! They will certainly help in my understanding of HDR digital photography. It might help if you quote the part your responding to.
-Ed-