Photoshop CS2 HDR and Beyond

My brother, an excellent photographer who has won a few ribbons on dpchallenge.com (philup), called me the other day asking me about HDR in CS2.  I teach a web graphics class and use Photoshop CS all the time, but I had no clue what he was talking about.  I began to investigate and soon realized that I may want to experiment as well.I made the plunge and upgraded my CS version to CS2 (I usually skip a release before I upgrade, but I was intrigued enough to jump the gun a bit this time).  I'll starty by giving a brief summary on HDR images.  While I don't fully understand the 32 bit architecture, it doesn't take a genius to figure out that these files hold more information that their 8 and 16 bit counter-parts.  I don't want this to be a lesson in the 32 bit format, but instead a lesson on how to help with mini photos using the HDR method.We typically work within our Photoshop canvas directly from files created from images snapped from our digital cameras.  That's why we never really think about the bit depth of our images, because we use what the camera gives us upon opening our image - 8 bits.  HDR combines multiple 8 bit images to create a single brilliant 32 bit image.The main thing we painters gain is true to color images.  We also get less shadow and more detail, but we lose a bit of actual shadow depth.  This process also takes away the guessing game for our camera EV setting or exposure and we could possibly eliminate some post rendering steps that sometimes strip our minis of color and depth.  This has to do with the ability of the HDR process to resample the high intensities versus the low intesnities and find a balance for a true life photo of a model not normally viewed under our photo lights.You may go through this exercise and decide that HDR is not for you for the benefit or lack thereof, and admitingly I won't do it for every model, but it does have a place.  Namely when you just want to show clean lines and natural looking painted fades and highlights while not relying on the model 3D and dark areas to do it for you.  This is truly for the hard-core painter who spends a lot of time on such details, unlike someone like me who paints to play (most of the time).  To start the process of HDR in Photoshop CS2, I will begin by taking 5 pictures (HDR recommends at least 3 or more).  I took these pictures at -2, -1, 0, +1, and +2 EV, which usually equates to an f stop of 2 in either direction to achieve.  You must have a tripod, use a timer and no flash on.  There are other steps with bracketing too, but you can read net articles about this based on your own camera.For my example, I'm using a square softbox made of 8x11 picture frames glued together where the glass has been replaced with trace paper.  I use 4 60 watt GE daylight bulbs.  One on top, two on each side and one in front beside my camera lens.  I shot this with a Canon Rebel (6mb) and a 50mm Macro lens on auto white balance (risky, but it worked)  Just be cautious of your white balance changing on you mid-stream, it could have weird results.  Parameter 1 default Canon setting.Let's start by looking at my initial set of 5 images to be used in creating the master HDR image.Below is the result with no manipulation on my part.  You can also see that Photoshop detected my EV from the image.Now Photoshop does not give us much color correction or any other adjustments I usually go through with minis, so we'll have to dump the 32 bit depth now (that's OK, we used it for what we needed it for) and go to an LDR method (8 bit).  In Phososhop, choose Image, Mode, 8 bit.  I accepted the default for the Method, Exposure and Gamma, by just increasing the Exposure a bit since I lacked that +3 EV image, but if you're good at Curves, you can use your HDR image one last time to get a nice tonal value for your image before you start your typical method of image fixing.  Since I do not have a good white balance set on my firmware, I always fix that first in Photoshop.  First I have to find my whitest white and blackest black areas of my image.  I use Photoshop's Threshold option for this.  I goto Image, Adjustments, then Threshold.  If I slide my threshold to the left, I can tell that my blackest point is just below his left rib cage.  (I make a mental note).Now to get the whitest point I slide the glider to the right and make a mental note that his whitest point is between his legs.Now that I have my mental notes, I need to fix the Curve of my image (blasted cameras sometimes).  I goto Image, Adjustments, then Curves.  The example below shows the buttons I clicked on to set both the black and white areas of my image.  This will fix any kind of white balance issues you may have as well.  The black circle is to set the black point and the white circle is to set the white point.Notice how that really cleaned up that yellow cast on my model.  Now I want to sharpen it up a bit using CS2's new Smart Sharpen.  I goto Filters, Sharpen, Smart Sharpen.  I set my Amount to 100 and kept the Radius 1.  Yours will vary depending on your taste, camera lens quality and camera parameter setting.One thinig I've noticed on Photoshop's HDR capability is the lack of contrast afterwards.  I have played with it a bit and found that adding a bit of contrast (Image, Adjustments, Brightness & Contrast) helps a lot with giving it more of a 3D look rather than looking so flat.  Lastly I'll do the typical background manipulation using the magic wand and gradient tool found in the forums here for reference, however I prefer to do my Contracting and Feathuring BEFORE I inverse my selection because I don't want the selection cutting into my mini and I use fairly light backgrounds so there isn't a halo effect.  I also want to work the same exact steps I did above, but this time with only one my starting images (the +1 EV image) so you can see the difference between the two.  Left Image = HDR'd imageRight Image = Starting single image at +1 EVAs you can see, there are a few differences of the two images.  Most people honestly prefer the image on the right.  I know that my model does not look more like the model on the right though, so I prefer as a true reflection of my model, the image on the left.  The left image does look a bit flat too because of the lack of depth in shadows, but the details are so much more vivid and not drowned out by shadows and lack of values.  Notice the areas in the mud near the left foot and between the legs.  Notice also the subtle fade in the skulls on the left compared to the more harsh transition of the fade on the right image.  Just to point out a couple things.The longest part of this method is the preparation up-front.  Once you have your images, the Photoshop part is not bad at all and you may find that you've been able to eliminate steps during your post-processing phase now since your color and value should be close to exact.XC