Unfortuntately, I don't know anything about Darktable. A quick google suggests that (unless I misunderstand) Darktable bases its conclusions about over/under exposure on the preview image embedded in the RAW - not on the RAW data itself. I think you would need to check with a tool such as RawDigger to be sure.

The histograms (luminance and each colour channel) for the "Image A" JPEG show a very different story - none of the highlights are clipped. I know that ACR models values for clipped colour channels from the unclipped data for the other channels. PCDU must do that too or we'd be able to see the clipping in the JPEG.

Without being able to examine the RAW, I cannot draw any firm conclusions but - if reprocessing from a 0.5Mpx JPEG gets anywhere near to the result from generating an HDR from 14.6 Mpx RAWs, I really don't think that the case for using HDR is made, IMHO.

Andrew

Interesting.

This is the histogram for Nigel's Image A.




Checking the individual colour channels shows that none of them have blown highlights. How did you verify that there were blown highlights in the RAW?

Andrew

If you re-read what I wrote both in that comment and others, it should become clear that I was referring to my own images - not the one you have used as an example for HDR.


It is not obvious to me what the subject of the photo is or what you were trying to communicate when you took it. Pointing your camera elsewhere might have been more productive. As you used HDR, you must have set the camera up on a tripod and taken your time about composing and shooting the images. This suggests that the technical quality of the end product was central to your purpose, rather than content.


Whilst the difference between your HDR processed images and that derived from your on-screen Image A is quite small, what you say here must be true. No one would deny that processing a full resolution RAW is likely to produce a better result than re-processing an 800 pixel wide JPEG.

If the image has a dynamic range substantially greater than the camera can record, several potential solutions suggest themselves. Reframing the subject to avoid the problem is one. Using flash, ND grads, ETTR (but avoiding highlight clipping) or HDR could be others depending on the subject. Although, from time to time, I have experimented with HDR, it has not become a technique that I use regularly. In general, exposing so as to JUST avoid clipping highlights and then adjusting tonality in Lightroom seems to do the trick. It also means that, most of the time, I can avoid lugging my (far too heavy) tripod around everywhere I go.

Andrew

I'm not wholly sure that is true.

Earlier in the thread Nigel said "I don't deny that focussing on the artistic will ultimately lead to better results than less lower quality artistic results 'boosted' by processing - I don't believe that anymore than anyone else.".

The thread was diverted into a discussion about HDR.

Andrew

On the contrary, I offered a reason why I should NOT use HDR. My proportion of keepers is relatively low. I don't see any evidence in this thread that regular use of HDR would improve it.


Andrew

Nigel, I admire your perseverence.

Ultimately, though, I think my time is better spent trying to take better photos than trying to polish turds.

Andrew

Spot on!! I agree completely.

Andrew

I remember talking to a friend about a shot he was tryng to take of something in a cave (with a not_Pentax 5D MkII) - but where part of the cave entrance was also visible. I suspect that the clipping for that would have been so severe that HDR was the only option.

Andrew

Bear in mind that the image I produced used an 800 pixel wide onscreen JPEG as source, not a full resolution RAW - so a look at the "actual quality of the detail and textures at both extremes" just isn't possible.

Andrew

If you look at the luminance histogram for Image A as posted by Nigel (and the processed version of Image A that I posted), it doesn't look clipped at either end. This is also true for the red and green channels - but the blue channel is clipped in the shadows.

Andrew