B&W digital workflow....what's yours?

Quote:

---

Originally Posted by nicnilov

Adam, what exactly are the results you're hoping for and what you're getting that doesn't match them?

When digitizing B&W negatives I consider getting the linear source image of utmost importance before any subsequent manipulation. It is relatively easy to get that using a scanner, but comparably difficult with DSLR. A DSLR produces a linear raw file but pretty much any software working with raws automatically applies a profile, and with it a tonal response curve. It makes the picture look pretty but messes up all the shadows and highlights tonality. Although it is possible to interpret a DSLR raw linearly, it is not as straightforward. In the end it is about gamma control which is not difficult using profiles in Photoshop.

---

I believe you can get the linear file output using the manufacturers conversion software. Ie, do not use ACR or similar when importing the raw file. Another program I have will read the raw file with no manipulation or non-linearizing the file and then you can save it as a tiff file in that form. But this program is for astronomy.

Quote:

---

Originally Posted by nicnilov

I've used NegativeLabPro quite a bit with B&W negatives (both scanned and DSLR) but my current take is that it doesn't really bring many benefits there.

Automatic processing of a VueScan raw dng in NLP sometimes does give pleasant results but not in a predictable way. That is, even with all settings off and Linear tone selected, the tonal distribution varies depending on the image content. Sometimes parts of the tonal range get so compressed as to become quite unusable. But this is not how NLP is supposed to be used with B&W.

The recommended way is to first convert the raw dng to a tiff (e.g. in VueScan) and run that tiff through the NLP's gamma tool before giving it to NLP itself. Then the results are nice and predictable but there are a couple of issues with this approach. First, the gamma tool saves the tiff as a new file with the histogram taking only about 40% of the usable range (for gamma 1.0, linear). For (the required) 16bit files this may not be horrible but this still is a destructive step. Second, after already doing most of the work manually, we end up in the NLP UI where all we can do is to use the same controls already available in Lightroom.

Maybe it does make sense and I'm missing something, but it seems that just opening the same original tiff in Photoshop and assigning a grey space with a chosen gamma is quicker, easier, less destructive, more flexible, and completely controllable.

This is not to negate the NLP benefits for color negatives, which are immense. For someone going after a particular look, fast, NLP with B&W may also be a good choice. My priority at this step is a linear (and soft - gamma corrected) source, matching the physical negative's natural tonal distribution as much as possible. Subsequent steps take care of the look.

---

I never scan to a dog file. I use vuescan , scan as 48bit color neg, save output as linear raw tiff. Then I got PS and work it there. I have a fairly straight forward workflow, but it has many steps. But I am quick nowadays having done it for a while.

Quote:

---

Originally Posted by Greg

Rough workflow, and I am sure that I am leaving out some steps...

First use a light box that has extremely even illumination
Camera a D850
Lens depends on the size/format of the negative
Mask out atop the light box any area that is not being photographed
Pretested to determine the best f/stop to use for each lens
Expose manually with one setting when stitching digital files together (11x14)
Exposure determined by histogram in camera, I never use auto exposure
Always shoot RAW at the highest res setting
Open in Camera Raw and sometimes tweak a little bit
Open in Photoshop and desaturate the RGB file to make B&W
Invert to work with a positive image
Almost always minor changes with Image>Adjustments>Curve
Any modifications made in additional layers, new layer for each different modification
Since my digital negatives are 1:1 with the film sizes, no one color channel is sharper than the others
Always save the digital file at max res in case in the future will be printing bigger

---

I wouldn't just desaturate and adjust the file in 16 bit. I would adjust it as the full 48bit file and at the end use a BW or channel mixer adjustment layer. better adjustability at the end.

Quote:

---

Originally Posted by Peter De Smidt

Adam asked this same question in another thread. I asked for some examples, but he didn't produce any. In any case, nicilov is right, if you are camera scanning negatives, then you want to change the profile used in the raw processor to 'camera neutral' or similar in Adobe or 'no profile' in Capture 1. Set black and white points, making sure not to clip image information, and adjust a bit to taste, always erring on the side of leaving more information rather than less. Final tweaks are in Photoshop, as it has much better masking and manipulation functions than a raw processor. Once in Photoshop, spot the file. I then use channel mixer to pick the best channel, which is usually the green channel. I keep the file in sRGB, as that's what I'm going to use on the web. Use curve adjustment layers for dodging and burning using layer masks. I usually add a vignette. I tone the image. I like a slight magenta tone to everything but the highlights. I end by tweaking overall tonality with a curve. Save master file. Resize and sharpen for web, changing to 8-bit color at the end.

https://www.dropbox.com/s/1jvyhxifd4...1200.jpg?raw=1

---

Great image.

How do you spot a file in photoshop? Are you just using the eye dropper curser to see what values you have in the image?

I mainly use the healing brush on a blank layer to get rid of dust spots. I generally don't use an eyedropper tool with bw.

I do the same. I thought you were meaning you were measuring your image for brightness levels. My bad.

Ok. Here's an update. Thank you for jumping in on this one.

I think my problem was this....after I'd shoot to negative with the DSLR the imported file into Affinity Photo has a color cast on them. I wasn't expecting that, but it makes sense.

So what i have to do is still convert the "color" RAW file and dump all the colors to make it black and white. So once I did that I had a true b&w image. Then the rest of it worked out ok.

I was asked to post samples but I can't seem to upload them. They just don't take. I must be doing something else wrong!!

Now that all said, I'm not 100% happy with my images, but I'm in the right ballpark now.

I'm going to follow thru with a few of your pointers.

Before converting to gray scale use color balance in PS to remove cast. Works pretty good.

If you upload a file it has to be small. See the guidelines for max image size. Click the picture frame icon in tool bar to upload from computer.

You can also post on Flickr then select sharing>BBCode, then choose size copy then cut and paste directly into body of your post/reply.

@Steven,

Quote:

---

I believe you can get the linear file output using the manufacturers conversion software. Ie, do not use ACR or similar when importing the raw file. Another program I have will read the raw file with no manipulation or non-linearizing the file and then you can save it as a tiff file in that form. But this program is for astronomy.

---

That generally is not the case. Even very flat manufacturer profiles, e.g. Nikon's Camera Flat do have non-linear tonal response curve, so Nikon's Capture NX-D, for one, does not allow linear interpretation as far as I'm aware, neither ACR or Lightroom. Some third-party raw converters, e.g. RawTherapee do allow true linear interpretation. This makes sense as full linearization generally is far from necessary in typical processing workflows.

Quote:

---

I never scan to a dog file. I use vuescan , scan as 48bit color neg, save output as linear raw tiff.

---

The reason to scan to raw dng is that it can be loaded as a source to VueScan at a later time and resaved in a different format, e.g. raw tiff without having to scan again. Adds flexibility.

Quote:

---

I wouldn't just desaturate and adjust the file in 16 bit. I would adjust it as the full 48bit file and at the end use a BW or channel mixer adjustment layer. better adjustability at the end.

---

This is a good idea when the source image is a color one. If it is a scanned B&W negative, any color that is there is not coming from the image itself. No reason to mix it into the adjustments.

Adam,

Quote:

---

Originally Posted by AdamD

Ok. Here's an update. Thank you for jumping in on this one.

I think my problem was this....after I'd shoot to negative with the DSLR the imported file into Affinity Photo has a color cast on them. I wasn't expecting that, but it makes sense.

---

Your approach is valid. The color cast is coming from the mismatch between the camera white balance selection and the temperature of the light lighting the negative. One way to reduce the color cast is to use custom white balance set against your light source. Also if your camera has WB character adjustment, make sure it is Neutral and not e.g. Warmer. But some color will be there anyway, so in the end it is not a big deal.

As @Steven said, set custom white balance on a transparent film spot before converting to B&W.