ok, i'll post a HDR Merge tutorial in Adobe Photoshop CS3. First of all, I wan't to say that's i'm sorry for those who have slower internet connections. I'll post the pictures at a pretty big size so you can see the menus and everything.
first of all, we have 3 exposures of the same scene (3 raws):
so, we have the 3 exposures. launch PS CS3 and then launch the hdr merge automated process in ps here:
from the file browser, select the 3 raw files. if the shots where made handheld, check the "Automatically allign pictures", otherwise you can leave it unchecked (it's faster and only when using a tripod, like me). click ok and then wait a few seconds/minutes (depending on the computing power). you'll get something like this:
click ok, and you'll have this for now:
now, the image is represented in 32bit floating point, that's way beyond what a jpeg can save and the monitor can display. there's aprox the equivalent dynamic range to a human eye. we'll have to make the tone mapping and conversion for out hdr.
to do that, click here:
from 32 bits we want to convert it to 16 bit (tiff format). later we'll do the 8 bit conversion, but that's pretty much nothing compared to 32 -> 16 bits.
so you'll get something like this:
as you can see, the picture is pretty much flat in terms of lighting because there the same amount of detail everywhere (shadows, mids, highlights). so here we make the tone mapping. select the "local adaptation" menu' and look at the tonning curve. when it's straight, the picture looks "flat". to "fix" that, we'll "throw" away some detail in shadows and highlights. so, we'll make the curve looking like this:
it's still looking a bit "flat" but we'll fix that in post-processing. click ok and then wait for photoshop to make the hdr conversion. what you have now it's a 16 bit image that we'll post-process.
first, duplicate the base layer and on the duplicate apply a "adjustment layer, channel mixer". check monochrome, and input the following values for RGB: 54, 54, 32. select ok and set the blending option of the layer to "multiply". then press "ctrl + shift + alt + E" on the top layer to flatten what we've got so far.
then, image -> adjustments -> shadows/highlights tool, with a value of 50% for shadows. click ok.
it's still looking a bit dull but we're fixing that right now, before we're done:
duplicate the top layer, apply a high pass filter (filter -> other -> high pass) with a radius of 80px (depending on the image size) and set the blending option of the last layer to "soft light". this will give the image contrast and texture.
flatten the image, change the mode from 16 to 8 bits and it's ready to be saved as a jpeg.
the final result, following the steps i just wrote above can be seen here:
hope this will come handy to many of you. i like PS better than photomatix because it's much easier to keep the noise in control and get much cleaner images.
sorry again for the high res pictures, hope it'll be ok.