It looks fake because it is. We already compress the natural brightness of light, HDR compresses it further. Full sun on a white wall is incredibly bright, brighter than any projection system can manage.

The lie is further compounded by the field of view. To realistic recreate a city view you'd need a very large screen or at least a system that wrapped the image around the viewer.

We can look into deep shadows and see inside them. We generally have to shift our gaze and wait for the pupils in our eyes to adjust. Shift our gaze back to something in direct sunlight and we can be temporarily blinded. Without huge advances in projection systems and very tolerant audiences this is limited to an interesting creative effect and one to be used with caution.

Bob.

Its not the High Dynamic Range you dislike, its the Tone Mapping to display the high range on low range media. All tone mappings associated with HDR that I have seen kill the local contrast.

What should happen, is highlight and lowlight/shadow regions should be identified and the brightness of each region (not pixels) should be adjusted while preserving the contrast within each region. The trick is to do this without introducing halos.

The Shadow/Highlight effect in Adobe Premiere etc. is the best I have seen for video editors, but unfortunately is applied to low dynamic range material that has been exposed for the highlights, and so the shadows show excessive grain/noise.

Furthermore, we don't usually want to be able to see everything in every shadow. We should only reveal what is in the shadows when it is important. Shadows also invoke an air of mystery. Rembrandt did this well.

Several years ago, I posted my technique for achieving HDR using Vegas:

OT: Polavision

My two captures were done with a Workprinter, but they could just as easily come from two cameras sharing a common lens, split via a prism, or some other way to get two images from the same plane, exposed differently. You then combine the two in Vegas as I describe in the post. It worked incredibly well, in my project, to take an absurdly high-contrast scene (taken with Polavision instant movie film) and turn it into something that looked "normal."

For those who don't want to click on the link, or who get lost, here is the "recipe" from the middle of the post linked to above:

1. Put the bright, over-exposed capture on the top track. Set that track to parent.

2. Put a Mask fX on this top track (or put it on the event: it doesn't matter).

3. Set the compositing mode of this first track to Multiply (Mask).

4. Create a second track and make it a child to the first track. Put the darker capture on this track, and line it up so the frames match with the first capture.

5. Put the brighter capture once again on the third track. This track should NOT be a child, but should be a "regular" track.

The only thing left is to play around with the mask controls on the fX on the first track. You set the mask to Luminance. Leave the bottom two sliders alone. Then, move the first two sliders around until you get just the highlights from the darker capture, but all the midtones and shadows from the brighter capture.

Voilà!