HDR and large format?

[Wallace_Billingham]

I am not a fan of negative film because it much grainier then my trannies.

If you are worried about grain in 8x10 negative film you will never be happy with the results of HDR at least using current software and hardware. The problem is that no matter what you do the process will introduce grain, noise, and artifacts in several steps. First off you will have the grain from multiple pieces of film that will show up in the scans. So if you have 3 sheets of film you will get 3X the grain. Also since some of the scans will be over exposed and some under that will introduce more grain than a properly exposed trannie.

Then when you go to make the scans you will get 3X the digital noise from the scans, and the noise will be much higher in the over/under exposed trannies.

Then you will get sharpness and artifact issues with the exposures themselves. HDR works best when shot with a digital camera because you can fire off many frames in a second with little to no camera shake electronically adjusting the shutter each time often automatically. With a LF camera you have to take the shot, put in the darkslide, remove the holder, adjust the shutter, put in a new holder, remove the darkslide, take the shot, put in the darkslide, remove the holder, adjust the shutter, put in the new holder, remove the darkslide and take the 3rd shot. All the while you better make sure the camera does not move at all, the wind does not blow against your bellows, the subject does not move, the lighting does not change etc. All of these things would result in sharpness and artifact issues when combined in HDR. And in the swamps of Florida that would be pretty hard.

The other issue is that the better HDR programs work best by reading the exif data embedded in the digital files so they can know the variables in exposure by reading the f/stop and shutter speeds. Your scans will not have that data.

I think you would be much better off just shooting 400 speed negative film and get crazy wide latitude with the trade off of grain

[Kirk Gittings]

Then when you go to make the scans you will get 3X the digital noise from the scans, and the noise will be much higher in the over/under exposed trannies.

Interesting, as running HDR on digital files results in virtually noiseless results. I am not saying that I am thrilled with HDR results in general but noiseless images are an obvious benefit. For example noise gets enhanced when you try and pull detail out of deep shadows, but in HDR you are using a exposure for the shadows that has very open shadows, resulting in virtually no noise in the shadows. The difference in digital imaging is dramatic in that regard.

[Brian Ellis]

I agree with Wallace. I have Photomatix and I don't even like it much with my digital camera - too much nose, artifacts, whatever for my tastes. And for large format there are so many additional problems, outlined by Wallace and others, that I don't think it would be very practical. Rather than using Photomatix or HDR with large format, I think you'd be better off just making two exposures - one for the shadows and one for the highlights - and merging them in Photoshop. Or even making just one exposure, then making two scans and merging them.

[David Luttmann]

Interesting, as running HDR on digital files results in virtually noiseless results. I am not saying that I am thrilled with HDR results in general but noiseless images are an obvious benefit. For example noise gets enhanced when you try and pull detail out of deep shadows, but in HDR you are using a exposure for the shadows that has very open shadows, resulting in virtually no noise in the shadows. The difference in digital imaging is dramatic in that regard.

I agree Kirk. Adding 3 negs does not produce 3x the grain. In fact, it has quite the opposite effect....grain is reduced. I used to use this method in the darkroom, stacking 2 shots of astronomical photographs shot on hypersensitized Kodak PPF400. Combining images smooths out grain.

[Marko]

<OT>

You don't need HDR to accomplish grain/noise reduction, you can do that by layer averaging. You essentially shoot several identical exposures and stack them up in Photoshop then adjust their individual opacity according to a simple formula. The alternative (for digital capture) is dark frame averaging, where you shoot one "dark" frame (with the lens cap on) using the same exposure as with the other frame(s).

</OT>

[Wallace_Billingham]

Interesting, as running HDR on digital files results in virtually noiseless results. I am not saying that I am thrilled with HDR results in general but noiseless images are an obvious benefit.

the reason for this is that digital files have exif data that tells the program what your ISO was set for and then the program uses that data to reduce the noise. Your film scans will not have such exif data and the NR software will have a much harder time dealing with both digtal noise and film grain

[Kirk Gittings]

I don't think that is true. Both the straight file and the file used for the HDR shadows have the same ISO exif data. The only difference being the exposure. In terms of shadow noise it is simply a matter of exposure with the better exposed lower noise shadows being used for the HDR shadows. This is clearly visible by just looking at the deep shadows in the original files.

[Emmanuel BIGLER]

So if you have 3 sheets of film..

For those who are ready to use 3 sheets of film, the tri-color separation process as advertised by several groups (including a very active French group) uses the very long characteritic curves of modern B&W films like Tri-X or HP5 to record colour images with a high dynamic range.
Actually a single HP5 film can record a very high dynamic range per se, but you need a damn' good scanner to extract the highest densities on film.

My understanding is that the limits of the actual dynamic range that can be recorded on an HP5 film are due to the internal flare of the lens+camera+film system.. and the performance of the scanner if you have to process the images digitally.

The tri-colour group on flickr : http://www.flickr.com/groups/92087504@N00/pool/

You can see some examples on Henri Gaud's web site
Systematic tests with various ISO sensitivities and N+/N- processing of HP5+ in HC-110.
http://trichromie.free.fr/trichromie...sement-par-iso
http://trichromie.free.fr/trichromie...x12-les-dtails

tri-color HP5 pushed to 6400 ISO just for fun : http://trichromie.free.fr/trichromie...ousser-un-film
http://trichromie.free.fr/trichromie...-film-suite-II

A recent example of the tri-color process, a portrait by Laurent Askienazy : probably not eligible as "HDR", but yes, you can do portraits with this strange process...

http://trichromie.free.fr/trichromie...rent-Askienazy

[Ken Lee]

If we wanted to get even more dynamic range, we could take multiple exposures with each color.

For example, a 9-color process: Red Green and Blue get 3 exposures each: one for the shadows, another for the mid-tones, and a third for the high values.

High-res scans of nine 8x10 negatives could possible strain one's computing resources - so it might be best to merge each color first, and then merge the final 3

[Wallace_Billingham]

.....Both the straight file and the file used for the HDR shadows have the same ISO exif data. The only difference being the exposure.....

Correct and the higher the ISO the more NR that gets applied behind the scenes. So if you have a 1600 ISO file it will get more than a 100 ISO file. This NR is set up to deal with digital noise which looks very different than film grain. One of the reasons that the Photoshop Actions/Plugins that add noise to an image to simutale grain look so bad.

Digital Noise combined with grain is very hard for software to remove and results in a loss of sharpness in the final image. Will you get a lot of grain and noise in the final image? Not really, but if the OP is bothered by film grain using 8x10 negative film than any added noise of grain would be too much.