RAW scanning in Silverfast, with with Epson V850

[rpagliari]

I like using the 16->8 mode when scanning B&W negatives because I can use the negafix to adjust for the specific film I'm using.

I found it very hard to emulate the results from negafix using what Silverfast refers to as RAW format.

The advantage of RAW is more flexibility I suppose, but using curves and levels I always and up with awful results, so I've always used 16->8.

Does anyone has suggestions/advices?

[Sasquatchian]

There is no raw format when scanning film, at least in the terms you're probably referring to - like that from a digital camera. Your best bet is to scan flat without clipping either the highlights or shadows using 16 bits per channel then using Curves Adjustment Layers to tone your image to your satisfaction. If your results are "awful" using Curves and Levels, there are plenty of tutorials out there to show you how to use them effectively. Hint: Levels are just Curves with three adjustable points and two hidden fixed points and a lot less flexibility than Curves. Once you master Curves, you'll never look at Levels again. No need to. Also, maybe post a couple of examples of what you're talking about so we can see and give better advice.

[rpagliari]

There is no raw format when scanning film, at least in the terms you're probably referring to - like that from a digital camera. Your best bet is to scan flat without clipping either the highlights or shadows using 16 bits per channel then using Curves Adjustment Layers to tone your image to your satisfaction. If your results are "awful" using Curves and Levels, there are plenty of tutorials out there to show you how to use them effectively. Hint: Levels are just Curves with three adjustable points and two hidden fixed points and a lot less flexibility than Curves. Once you master Curves, you'll never look at Levels again. No need to. Also, maybe post a couple of examples of what you're talking about so we can see and give better advice.

Do you see any advantage in using RAW format (Silverfast version of RAW). I've been using the 16->8 mode for a while now, and it seems just fine in post-processing.

[Sasquatchian]

What is 16-8 mode?

[Sasquatchian]

OK, I see that is when SF uses 16 bits internally and then outputs 8 bits. No, you don't want to do that. You want to scan internally at 16 bits and keep it at 16 bits, so whatever setting it takes to let that happen, as I don't have a copy of SF to look at and it's been a long long time since I've set anyone up on that. You also might want to try scanning as 48 bit RGB even though it's black and white. Often times these lower priced scanners, when using a single channel, are not using the best one and it's better to get a neutral RGB scan and then convert to Grayscale or do what I do, just leave it as 16 bit RGB.

[Alan9940]

I've been using Silverfast scanning software for nearly 20 years and the best results I've ever obtained are when scanning a raw linear file, then using ColorPerfect to convert in PS.

[Pere Casals]

You have to save files in TIFF format to conserve 16 bits. TIFF, always TIFF !!

[Alan Klein]

There is no raw format when scanning film, at least in the terms you're probably referring to - like that from a digital camera. Your best bet is to scan flat without clipping either the highlights or shadows using 16 bits per channel then using Curves Adjustment Layers to tone your image to your satisfaction. If your results are "awful" using Curves and Levels, there are plenty of tutorials out there to show you how to use them effectively. Hint: Levels are just Curves with three adjustable points and two hidden fixed points and a lot less flexibility than Curves. Once you master Curves, you'll never look at Levels again. No need to. Also, maybe post a couple of examples of what you're talking about so we can see and give better advice.

I'm scanning with a Epson V600 and Epsonscan software. I adjust levels so there is no clipping during the scan. But all other adjustments I do in post processing. Can you explain how to use curves instead of levels during the scan.

[Sasquatchian]

Alan - I don't know what your level of comfort is with Photoshop. The best way tone your black and white scan is with a series of Adjustment Layers. Curves is almost always the best tool as it allows you to have more control over the tonality of the image than any other tool in Ps. You can orient the Curves dialog so that the shadows are at the lower left and the highlights are at the upper right. That should be the default but not always and old time pre-press guys like them to be the opposite.

I usually start out with a fairly simple Curve that just sets highlight and shadow points (using the Color Sampler Probes in the Info Palette) and maybe hits a bit of overall tonality. Then I move to more targeted Adjustment Layers, making the adjustment, then immediately filling the mask of the Adjustment Layer with Black, then slowly painting that adjustment back in using an appropriate brush. By painting on the mask, you can not make a mistake, painting in the adjustment with white and painting it back out with black.

The advantage of Curves over Levels is that you can target any part of the tonal range and use up to 16 adjustment point on the Curve to make fairly complex and effective corrections. You can lighten an area while simultaneously adding contrast exactly how YOU want to and not be severely limited by the three control points of Levels.

There are many youtube tutes out there about how to use Adjustment Layers, but the basic idea is that with an Adjustment Layer is infinitely undoable and does not permanently apply its adjustment until you finally flatten all the layers of the image. Which, of course, I don't recommend that you do. Generally keep a layered Tiff or .psb file if your files are larger than 4 gb and save off a copy for printing or online distribution. There is very little reason to ever use the native .psd Photoshop format anymore as tiff does everything that .psd will do except make duotone, tritones and quadtones for offset press and tiff doubles the file size limit of .psd from 2 gb to 4 - and lets you save with zip compression on both background and layers if you want which results in longer save times but a smaller file than .psd.

[Jeff T]

I've used Silverfast raw 16bit (4passes) to get the file then use Silverfast HDR Studio to process the file to 16bit TIF or 8bit TIF files with dust removal feature and all the color adjustments on the fly. Processing the raw files is much faster than scanning for raw files. Raw files are huge and slow to scan but will maximize your effort if you don't want to rescan the film in the future. Silverfast Negafix is a big disappointment, and will not accurately emulate the film signature colors and the choices available are limiting. I always have to tweak the colors in Silverfast HDR Studio to get get what I need and it's much more efficient than messing with Photoshop.

[Sasquatchian]

The advantage of Photoshop over any sort of scanning program for post production, of course, are the adjustment layers and advanced masking tools. Layer Masks alone are worth the price of admission and once you learn the tools, it ain't "messing" around, it's getting stuff done quickly and right.

[Alan9940]

@Sasquatchian, I've used PS since v5.5 so I feel pretty comfortable with the software and what it can do. I was responding the the OP's request for help when scanning B&W negatives. Like Jeff T, I wrestled for years with Negafix and never found any setting that suited my taste; I wound up simply turning it off! Scanning to linear raw files in Silverfast can be confusing, but it's certainly more straightforward, if you stay in the Silverfast family via HDR Studio. However, after much testing and trial-n-error I stumbled across ColorPerfect and have never used anything else for years.

[rpagliari]

The problem with scanning at 48 bit is that the file becomes huge at 2400dpi. Should I reduce to 1600?

[rpagliari]

Alan - I don't know what your level of comfort is with Photoshop. The best way tone your black and white scan is with a series of Adjustment Layers. Curves is almost always the best tool as it allows you to have more control over the tonality of the image than any other tool in Ps. You can orient the Curves dialog so that the shadows are at the lower left and the highlights are at the upper right. That should be the default but not always and old time pre-press guys like them to be the opposite.

I usually start out with a fairly simple Curve that just sets highlight and shadow points (using the Color Sampler Probes in the Info Palette) and maybe hits a bit of overall tonality. Then I move to more targeted Adjustment Layers, making the adjustment, then immediately filling the mask of the Adjustment Layer with Black, then slowly painting that adjustment back in using an appropriate brush. By painting on the mask, you can not make a mistake, painting in the adjustment with white and painting it back out with black.

The advantage of Curves over Levels is that you can target any part of the tonal range and use up to 16 adjustment point on the Curve to make fairly complex and effective corrections. You can lighten an area while simultaneously adding contrast exactly how YOU want to and not be severely limited by the three control points of Levels.

There are many youtube tutes out there about how to use Adjustment Layers, but the basic idea is that with an Adjustment Layer is infinitely undoable and does not permanently apply its adjustment until you finally flatten all the layers of the image. Which, of course, I don't recommend that you do. Generally keep a layered Tiff or .psb file if your files are larger than 4 gb and save off a copy for printing or online distribution. There is very little reason to ever use the native .psd Photoshop format anymore as tiff does everything that .psd will do except make duotone, tritones and quadtones for offset press and tiff doubles the file size limit of .psd from 2 gb to 4 - and lets you save with zip compression on both background and layers if you want which results in longer save times but a smaller file than .psd.

The nice thing about negafix is that it tries to emulate the specific film in use. I'm not sure I could do that with curve adjustments. Did you find curve adjustments for different types of film online?

[Sasquatchian]

There's really no such thing as emulating a black and white film. Remember when you printed in the darkroom, you chose the developer, the length of time to develop, the paper type and grade, or used multiple grades in the same print sometimes, or created soft or hard edged masks using Pan Masking Film (yeah!) all to make your print look like YOU wanted it. Pretty much the same way now in a hybrid workflow. YOU decide how YOU want it to look and use the tools Adobe gave you to create that. Remember their old ads from the mid 1990's - "If you can dream it, you can do it" Well, it's pretty much that but we've come a long way since then. The same goes for color neg. There's no right or wrong and no preset for a specific film is going to get you what you want. You need to decide and use the tools to craft that image to your liking. Black and white is simple. Color negs are more challenging because no two are alike, and simple changes in exposure, development, emulsion batches, lighting conditions, etc., mean that you have to use manual intervention to massage the file in the right direction. That orange mask has to be made neutral and that is not the same on every neg, even on different exposures from the same roll. YOU have to decide where to set your highlight and shadow values and whether midtones need to be neutral or not.

There's no need to try and find curves online. You just make your own to suit your own images. It's really not hard at all. Just make sure to use Adjustment Layers and not apply the curve directly to your background image. Between Curves and Hue/Saturation you can do whatever you need on color images and Curves is all you need for black and white. I've used about half a dozen different scanning packages to scan, or attempt to scan color negs, and, by luck I guess, by far the easiest and the best (and there are always problem child images) has been the old drum scanning app for the Howtek scanners, Trident. More recently I've played around, just for kicks, with copying color negs on a light box with a Canon 5DSR and a good macro lens. Using that, I've done the inversions and orange mask removal both using Photoshop and also entirely in Capture One. They both work pretty well, and frankly, I was surprised at just how well using the controls in C1 worked for this. Still, my preference is always to drum scan, but I also realize that very few of us have their own. There are also some clever ways to use layers and blending modes to subtract out the orange mask on a color neg in Ps. This may indeed be what some of these other programs are doing behind the scenes, but we'll likely never know that for sure. Anyhow, lots of options out there and it's easy and cheap to experiment.

[rpagliari]

OK, I see that is when SF uses 16 bits internally and then outputs 8 bits. No, you don't want to do that. You want to scan internally at 16 bits and keep it at 16 bits, so whatever setting it takes to let that happen, as I don't have a copy of SF to look at and it's been a long long time since I've set anyone up on that. You also might want to try scanning as 48 bit RGB even though it's black and white. Often times these lower priced scanners, when using a single channel, are not using the best one and it's better to get a neutral RGB scan and then convert to Grayscale or do what I do, just leave it as 16 bit RGB.

That way your image would triple in size right?

[Sasquatchian]

Your file size doubles when it goes from 8 bits to 16 bits, not triple. It's all the layers and adjustment layers that really ad up, but that's the price of admission these days and big hard drives are cheaper than ever.

[Sasquatchian]

And if you do save out 16 bit per channel files, you have the choice of using no compression, LZW compression or ZIP compression. Use ZIP, not LZW. If you use LZW and your files are not true 16 bit per channel - like so many scanners that are really 12 or 14 bit internally and then pad the file up to 16 bits, the file sizes actually get bigger instead of smaller, weirdly enough. Use ZIP. Actually good advice across the board.