LaserSoft Imaging

Any tips on creating smoother gradients when scanning B&W negatives with areas in images such as open skies shot with red filter on lens.
I keep getting speckly gradients when the gradient range is great, but it doesn't look anything like the film grain. The film was 35mm Ilford Delta 400, and I tried both 16bit Grey and 48bit colour scans at 2800dpi using an Epson f3200, and Silverfast AI. I don't use any sharpening during the scan, but found that using GAIN did some smoothing, but I didn't like the effect, and using Neat Image or Remove Noise in Photoshop only resulted in ugly banding. I have seen some posts relating to scanning as HDR, will this help?

One of the issues may be the resolving ability of the Epson scanner. If i remember correctly, most of the scanners in that era could resolve somewhere around 1800 - 2200 ppi . Also, the Epson scanner driver may be including some sharpening on its own. i don't know how to turn that off.

Try scanning at 1600 ppi, color, 16 bit per channel and see if that helps any. Look at the individual histograms and see which is contributing the noise in the image and apply noise reduction to it only.

HDR is giving you the "raw" bits from the scanner, after the driver has done its basic work. You will need a copy of HDR or DC to work with the file and use Silverfast tools.

If you are seeing color speckles, it may also be scanner sensor/electronics noise. If that scanner will do multisampling or can do Silverfast's implementation of the same, you may be able to cancel out a good amount of the random noise. 4x sampling seems to be a reasonable balance of time and effectiveness on my CS5000.

Thanks for your reply. Having had another look at the scanned negatives, and checking the the histograms, and then trying every possible combination, I'm convinced now it is the Ilford Delta 400 film grain that is being resolved as patchy when using the Ilford Delta 400 choice in negafix. Although the film is Delta 400, I got way smoother results when I tried scanning as other film choices, in fact the difference was quite remarkable when I scanned the negative as either Neopan or PanF.

I tried reducing the resolution as you suggested, but lessening the resolution did not adjust the patchy effect for the better, and the f3200 scans optically to 3200 anyway. I've attached a 100pxX100px section of one scan with its patchy resolution of the film grain as an example.

This is chromagenic B&W ? It may just be the dye clouds. Do you get the same response with other films (color neg films) ?

The trouble with scanning negs is that the dark areas, which have most highlight detail, are hardest for the scanner to detect above the noise limitation of the sensor and electronics. Thus highlights tend to have most electrical noise (speckles) from the scanner, which may be what you're seeing. Also, the flat bed scanners tend to have higher Dmins than a film scanner compounding the issue.
Since the problem is in the dark areas of the negs, best bet is to use multi-sampling which will minimise the effect of random noise in the dark areas by canceling it out and thus maximise the detail in your highlights.

i have read about some scanner noise issues being somewhat helped by cleaning up the power to the scanner (or taking it off of a "UPS" which generally adds high frequency noise on the power lines and ground lines). Using a surge arrestor with high frequency filtering may help. Consumer UPSs usually synthesize the sine wave with a chopper circuit that , if not heavily filtered injects high frequency noise.

Try blowing out the highlights more than the autoexposure does and see if that helps some.

Another approach is to scan it as a positive and invert applying your own curves. As you saw from using different film profiles you will get very different results and you can edit those profiles to get closer to a desirable image if you use negafix.

Or if you are my age, just taking off the glasses helps tremendously ;-)

Thankyou very much for your reply, the info is very helpful. I have only noticed this effect with this batch of 35mm Ilford Delta 400, as I'm more used to scanning 645 and 6x6 format films, and usually it is Neopan, FP4, or TMax. With those emulsions, the results are silky smooth, even when choosing an 8 bit greyscale output. The scanner I use is an Epson film scanner (f3200), it is not a flatbed. I tried scanning the same negatives with the twain driven Epson Scan software in Photoshop as an experiment, that Epson software is just plain awful to use, but its results were without the blotchiness that I got from Silverfast AI and the Delta 400 negafix profile, so there must be some interesting stuff going on with that profile. I will follow your advice, and overide the auto exposure settings and see what happens, and work towards a useful curve for this film. However, I only had a couple of films of Delta 400, as i just got a batch of FP4 for my next set of work. Thanks a lot.