Attached Link: http://www.photoscientia.co.uk/Grain.htm

I was off looking for something else and found this. The link goes to an article about why some results with film scanners look really "grainy," much worse than what one would expect from the film. The discussion seems to make sense, (some of the math folks out there may know about Nyquist sampling, Fourier transforms, etc.). The best solution seems to be to manually de-focus very slightly while making the scan. Just enough to blur the grain, without losing too much image detail. Always scan at highest resolution available, and then reduce in Photoshop. Then, any image sharpness that may have been lost will be recovered with unsharp masking. [Trying to fix it after scanning, by blurring or despeckeling, just makes thinks crepier.]

This looks really interesting Misha...off to read it now....

What I found interesting was his statement about scans being the hardest when the film is dark and has low contrast.

Duh... An underexposed piece of film will not print well either. Even though there is more latitude in a traditional darkroom, the bottom line is that there is no substitute for a good exposure.

It would be nice if there was a publication date as some of his references made me think that the information was not real current.

I will say that I have tried the soft focus thing before as my scanner allows for manual focus adjustment, and that it can help. However, it is a trade off, and the amount of noise reduction gained does affect the overall image sharpness and tonal qualities.

It is still best to stick with ISO 160 or lower speed films if you intend to scan most of your work and expect high quality prints.

Grain and exposure are of course related. Can't turn a bad exposure into something good. I think the author is just making that observation here. It does not seem to be the central point. I also don't think the author is just saying slow speed film is good, high speed is bad. I think the point is that conventional prints or projected slides made from higher speed films (sometimes) look far better than digital scans made from these films, and he is trying to understand the mechanism in order to improve things. Several people have commented here that their scans, of nominally high quality materials, seem to look too grainy. [Then again, maybe the author is just a dufus... It is frustrating not having a curriculum vitae coming with random google hits.] Approximating a continuous image with equally spaced samples is the same sort of thing as sampling continuous music with periodic time samples; much of the same mathematics applies. If there are brightness variations in the image, (in this case on the film), that occur closer together than the sample interval, then these variations cannot be adequately approximated. These higher frequency components "beat" with the sample interval. At best, this adds uniform noise. At worst, the beat noise is not uniform, but has visible patterns. The analogous effect in music is a "harshness," or audible tone artifacts. In music, the way to reduce these artifacts is to low-pass filter the source, prior to sampling. Ideally, this removes all variations changing faster than 1/2 the sample rate, so nothing is left to beat with the sampler. In imaging, this low-pass filtering could be accomplished be a "blurring screen" sandwitched between the film and the sensor, or by simply throwing things slightly out of focus. Actually accomplishing this, without the adverse tradeoffs Alpha mentions, is problematic. I'd rather have it handled in a controlled manner by the scanner manufacturer. Perhaps the key point the author makes is that the problem has to be addressed at the scanner hardware, and cannot be fixed by later photoshop processing. One suspects that modern scanner manufacturers are well aware of this, and employ some sort of blurring in the hardware. The Applied Sciencefiction GEM techniques, used in several modern scanners, do not seem to be just image processing, but seem to also vary the focus. [BTW, this may also be why it is not important to have TOO good a lens on a digital camera... At this point, we expect Rork to jump in, "See, analog is better, and I like vinyl music too." Both good points. ;-) ]

Attached Link: Grain Surgery" target="_blank">Grain Surgery

Here is a couple of samples from a Demo version of a program called Grain Surgery that I just downloaded. (I have included a link for anyone interested)

This first image is a photo that Niko uploaded the other day called Kid. The grain (actually dye clouds because it is color) is very strong. I asked what film was used,but as Niko is never around on the weekends I have not gotten a response yet. The photo next to it has been run through Grain Surgery using the program's default settings and then sharpened in PS7 using an unsharp mask with a threshold of 3, a 0.7 radius and a 90% sharpen setting.

The photo was at 72 PPI when I got it, so the question was how well does it work at a 300 PPI setting...

This image was shot with T-Max ISO 1600 pushed to 6400 (That is 2 full f-stops). It was hand processed in T-Max developer by yours truly, and scanned on a Nikon LS-2000 film scanner at 300 PPI. I ran a cropped portion of the scan through Grain Surgery again at the default settings. Once the filter had run, I reduced the image to 72 PPI for uploading and applied an Unsharp Mask with a threshold of 3, a 0.7 radius and a 150% sharpen setting on both the original and the filtered version. I did have to convert the photo to RGB in order for the filter to work and the speed at which the filter ran diminished exponentially with the high resolution scan.

For a first try with the filter, and having read nothing in the instructions, I am fairly impressed. I think most people would find the results of acceptable quality. I will test it some more over the next week or so and form a final opinion by then. The only thing I really wish would be that they didn't have a stupid screen over the image and would just have the demo time out instead. It would be a lot easier to see the results without a visual distraction.

Wow! Seems to work quite well! Have to check out that link.

Not only does this program remove grain, but it also lets you selectively add grain with much more control then photoshop's built in alternitives. Why would one want to add grain to a photo you ask???

All B&W films have a unique grain characteristic. Anyone adept at working with and evaluating prints from B&W film can see and spot grain qualities immedietly. One of the major complaints that purists have had in digital work is the lack of grain detail in prints. Some companies have made filters to address this before, but in all honesty they were very expensive and the results were not that great. Again this program has impressed me. Just wish it wasn't $200.00.

BTW... This was shot with T-Max 100 and developed in D-76 by Dobby the House Elf aka Alpha... The lack of visible grain in the original is due to not over-agitating the film during processing.

Grain Surgery is a wonder (I hold it in as high esteem as I do the gift from the gods the MIGHTY "Healing Brush". :) One of the really great things about Grain Surgery is that it actually has 3 seperate uses. The one it's most often used for is the removal or minimization of grain. But like Alpha says above, it can also add grain (you can choose grain templates - ie. Tri-X 100, Tri-X 400, etc.). The third is the one that I just love the sound of (but haven't used yet), you can actually input an image into it with grain that you like and it will sample/analyze the grain structure and then attempt to replicate the look of that grain onto a 2nd image. The grain aliasing issue is still a current one. There is a recent (6 months or less) article on it that I read about 2 week ago. I'll see if I can find that link somewhere. -=>Donald

I googled around a bit... I was not aware that this is such an inflammatory issue. Links out there on the web get pretty violent!

...pretty stunning results. Could not even imagine there could be program doing results like ... I agree you Alpha, I like what I see. :o) As I'm getting digital soon, this would be really nice way to get more personal results for b&w photos. Prize is pretty salty though. I have played with Photoshops grain a bit, it's not as good as this seems to be, the grain there is, lifeless. .n

@Alpha, indeed....there's no substitute for good exposure. And film thickness seems to make some difference too.