This is a non-hdr image taken at the same time as the example images below. Note the slight overexposure of the sky, which was very dark and cloudy at this time.
I have been getting a lot of questions regarding the HDR images that I have taken and placed on my website. Before I get into the subject of this discussion, I want to explain what an HDR is. HDR is an acronym which stands for High Dynamic Range. When speaking about HDR photography, the images produced using this technique employ a very large range of lights and darks that cover the entire spectrum of exposure. This technique should be used when parts of a subject are dark while other parts are light and both parts have details which can add to the photo. For instance, a great example of when an HDR can and should be employed would be taking a photo of a sunset where you also have a subject besides the sky that you would like to properly expose in the photo. By utilizing an HDR, the colors of the sky will show up along with the subject in perfect exposure, where as, in a non-hdr photo, you have to chose whether to properly exposure the sky and underexpose the subject, or over expose the sky and properly expose the subject. HDR allows both subjects to be exposed correctly at the same time adding a lot more to the photo than the normal non-hdr.
With that initial background out of the way, I will focus the rest of this on what techniques to use when making an HDR. Plain and simple, the best HDR's are those that are taken over the entire exposure values (-5 to +5). Most entry level DSLRs have built in bracketing functions that allow the camera to be set up to take a certain number of photos across the different exposure values (EV) automatically by simply continuing to hold down the shutter button. Depending on how spaced out the EV values are set to, you can take a lot of varying exposed images. Personally, I try to shoot 9 frames with a 1 EV space between every image. This gives me photos with EV values of -4, -3, -2, -1, 0, +1, +2, +3, +4. Now, this leaves out the +/- 5 EV, but in most cases, there isn't much additional detail to be gained from these extreme values. The goal in this is to cover the entire range, which can be accomplished in as little as three images, or as many you can take with the smallest EV changes between photos.
This is a 9 image HDR composition with my "Storm" filter applied to it with no other post production.
Now, the more photos you take, the more chance there is for differences between images, which is essentially the problem that comes with trying to take several identical images of a subject with the only change being the EV value. When shooting a moving object, it is impossible to get a multi-exposure range by taking multiple exposures. Here is where taking the image in RAW format will be an advantage. A RAW image is the equivalent to essentially an electronic negative of the photo with absolutely zero post processing done to it in camera. Off the camera, the image can be artificially manipulated into multiple images with different EV values, thereby creating the entire exposure range. But, this is more of a pseudo-HDR than a true HDR because depending on the original settings of the image, a lot of the detail can be lost to over or under exposure.
This is the psuedo HDR RAW conversion of the same image with the same "Storm" filter applied. Note the lack of detail in the amount in the clouds in comparison to the 9 image true HDR shown above.
So, the decision on whether to take a multi-image HDR or a single exposure RAW HDR really all comes down to what kind of subject you are taking. If there isn't access to a tripod, a RAW image may be the best bet. If the subject is moving (could be as little as the wind causing leaves to move) a RAW image would be the best bet. But, if all things are constant and unmoving, then a multi-image HDR will give you the best results.
In conclusion, multi-image HDRs produce the best dynamic ranges over pseudo RAW converted HDRs that lose details in the most under or overexposured portions of them. But, sometimes a trade off must be made in order to produce an image that would otherwise be impossible to produce using the multi-image technique. Knowing which one is every situation comes from just evaluating what the subject is and whether it will stay still long enough to take a few pictures of it.