The GIMP for processing B&W

J

jamiewakeham

Long time lurker
Local time
8:44 PM
Joined
Jan 6, 2005
Messages
362
Location
Oxford, GB
View My Gallery
Hi all.

Busy sorting out the selected 60-odd photos that I took of a friend's wedding recently. For the first time, I've scanned everything and am using the GIMP (don't have PS) to try and process my images. Yes, it stops me having to ban my girlfriend from the blacked-out bathroom, but I'm finding it a bit frustrating...

I've uploaded a typical image to my gallery. I've done my best with it - applied (IIRC) +10 brightness and +15 contrast, followed by the GIMP's recommended unsharp mark.

The image just isn't doing it for me yet. I've got some blacks in there, and just about some whites, but the whole murky grey of the midtones is annoying me greatly. I'd rather see more black blacks and more white whites than all the murk.
I know what I'd do if I was looking at a test strip from an enlarger, but...

Anyone got any suggestions for a complete photoshopping ignoramus as to how to get the results I'm after? I rather suspect that I need to work with some sort of histogram here, but I figure that you guys can tell me what to do a damn sight faster than I can work it out for myself!

Thanks ever so.
Jamie
 
I'm not familiar with your software, but if I were faced with this problem I'd use the brightness curve, which is available in Photoshop and in most other post-processing packages, such as Picture Window Pro, which is the one I use for this stage of the process.
If you can, I'd keep my image in 16bits (for B&W) or 48 bits (for RBG) until after you've done this stage, and also color balancing and saturation mods if applicable. Then you can switch to 8bits / 24 bits for sharpening (if needed) and printing.

On a brightness curve, the input image is considered to have a constant contrast gradient which is a straight line running from bottom left (black) to top right (white). The histogram, superimposed on this, may or may not run across the full tonal range. If it does not, then you are not going to get deep blacks or clean whites. Usually you should therefore pull the bottom left of the curve to the right until it just meets the left end of the histogram. Also pull the top end of the curve left until it just meets the right end of the histogram. At this point, your image may look OK, or you may feel contrast is too low (for example) in the mid tones. If this is the case, you must tweak the curve from a straight line into a sort of S-shape, so that it becomes steeper in the middle and flatter at top and bottom. Exactly how you do this varies with the software you're using. It takes a little practise to get the hang of what works.

Hope this helps. If you need anything more, please jsut ask. No guarantees, but I'll do my best!
 
In GIMP you want Layer -> Colors -> Levels

That will give you a histogram and level control.

See Gimp Guru for other tips for The GIMP
 
Last edited:
Thanks, both.

I've uploaded a new version, where I've simply used the 'pick black' and 'pick white' tools and then set the gamme to the central midtone; it still needs a little tweaking to my eye, but it's a damn sight better!

Cheers,
Jamie
 
kmack said:
In GIMP you want Layer -> Colors -> Levels

That will give you a histogram and level control.

See Gimp Guru for other tips for The GIMP
Click to expand...
I considered myself TERRIBLE in Photoshop and even WORSE on GIMP. But lately, I've been using Gimp and trying to break that learning curve. I can't afford Photoshop or a new computer.

LEVELS. I just learned levels this summer and all my shots I've scanned look soooooo much better. I start out by converting to monochrome, then using levels, and then, if I have to, use curves and then contrast. Then do that sharp, resize, unsharp thing you did. How I did without levels...I don't know, I don't even want to look at my old scans.
 
I always save my original scans.
When I make adjustments using GIMP I then save them to a new file. That way if I learn a new trick, or want to try something different, I always have the original to work on.

One other trick, scan your negitives twice, I save them as A & B. That way you have a dup that you can use for masking grain aliasing.

Also -- see this thread
Non Destructive Sharpening -- a great tip from Fedzilla-Bob
 
Last edited:
eric said:
I considered myself TERRIBLE in Photoshop and even WORSE on GIMP. But lately, I've been using Gimp and trying to break that learning curve. I can't afford Photoshop or a new computer.

LEVELS. I just learned levels this summer and all my shots I've scanned look soooooo much better. I start out by converting to monochrome, then using levels, and then, if I have to, use curves and then contrast. Then do that sharp, resize, unsharp thing you did. How I did without levels...I don't know, I don't even want to look at my old scans.
Click to expand...

1) I scan my B&W negs as generic color using Vuescan, saving as TIF.
2) Open file in The Gimp.
3) Crop if needed.
4) Adjust levels.
5) Desaturate.
6) Convert to Grayscale.
7) Check levels again.
8) Save as XCF (Gimp Native Format) file.
9) Clean using clone tool as required.
10) Save again.
11) Scale Image for whatever use (web, printing, etc).
12) Apply USM.
13) Save as JPG.

I agree, levels are the key. Sometimes I simply adjust the black and white points myself, and occasionally move the center (midtones) slider on way or the other. Sometimes I try the 'auto adjust' and be ready to hit 'reset' if it goes wonky. I also sometimes try doing the eyedropper to set the white and black point on the image itself, but that seldom works well for me - I often get whites that are blown out and detail lost. But I play around until I get something I like.

I understand that 'curves' are more powerful than 'levels' in The Gimp, as they are apparently also in PS. But I have not yet taught myself to use much more than the routines I've mentioned above, so I leave that be for now unless I'm just playing around.

I try to always apply USM after resizing an image, and I play with the amounts until I get something that is a subtle increase in sharpness, but does not appear to have been sharpened. Could be I'm too subtle with it, but I like the effect overall.

That's about all I do with The Gimp and film. I also use it with my DSLR and RAW files, but that's another story.

I realize that The Gimp is not as powerful as PS, but it and Linux are free - more money for film, GAS, and computer hardware. Not to mention beer.

Best Regards,

Bill Mattocks
 
bmattock said:
1) as TIF.
2) Open file in The Gimp.
3) Crop if needed.
4) Adjust levels.
5) Desaturate.
6) Convert to Grayscale.
Click to expand...

I learned something already!!
I didn't knwo I can adjust the leves before I convert to Grayscale. Is that better? Or is there an advantage to doing that than convert to grayscale and then levels?
 
eric said:
I learned something already!!
I didn't knwo I can adjust the leves before I convert to Grayscale. Is that better? Or is there an advantage to doing that than convert to grayscale and then levels?
Click to expand...

I used to desaturate and then adjust levels and then convert to grayscale. Somewhere, I read that I should do the levels first, so I started doing that.

I'm not positive it makes a difference, but it also doesn't seem to hurt, so I keep doing it.

When converting a negative from color to B&W, I have also (at times) experimented with Layers, choosing the RED layer and deleting the other two. Then saving, closing and reopening, then adjusting levels and converting to grayscale. Just kinda sorta trying to simulate as if I had a red filter on the lens and was shooting B&W to begin with. I'm probably being an idiot, but it is kind of fun.

LOL!

Best Regards,

Bill Mattocks
 
GIMP is extremeley powerful for what we do, and its free like beer!

There are several ways to do what you describe and the procedure bill mentions is pretty simple and very very effective.

Curves are a tad more powerful tahn levels in the sense that you can affect the overall look, orf the whole picture or you can do localizad edits.
 
bmattock said:
I used to desaturate and then adjust levels and then convert to grayscale. Somewhere, I read that I should do the levels first, so I started doing that.

I'm not positive it makes a difference, but it also doesn't seem to hurt, so I keep doing it.

When converting a negative from color to B&W, I have also (at times) experimented with Layers, choosing the RED layer and deleting the other two. Then saving, closing and reopening, then adjusting levels and converting to grayscale. Just kinda sorta trying to simulate as if I had a red filter on the lens and was shooting B&W to begin with. I'm probably being an idiot, but it is kind of fun.
Click to expand...

Rather than using Desaturate try Filters -> Color -> Channel Mixer.
In the Mixer you can then click the Monochrome check box.

You can then adjust the filters to get the kind of effect you were trying to get with the layers dialog.
 
From someone who works with Photoshop, I can tell you that the three most useful things for my purposes are:

1. auto levels adjustment (or manual, per-channel adjustment if possible)
2. brightness/contrast curves ("curves" in photoshop)
3. channel mixer (not sure what the GIMP equivalent is)

I had problems with middle grays as well; these were often washed-out and muddy. Now, I like contrasty images, not so much a Rodinal effect all the time, but you get the idea.

...now, I should point out that I work mostly with scanned B&W negatives, both chromogenic and silver. I scan them as 24-bit RGB, which based on what I've read gives a potentially larger tonal range when I'm done (i.e., more initial information to work with). The "channel mixer" step above lets me enhance individual R,G,B channels BEFORE i "desaturate". I don't actually convert to grayscale until much later in the process; sharpening is the LAST thing I do, after any resizing, right before saving.

I don't know if this helps, as much as confuses; I just came off a night of scanning work, so it's all rather fresh in my head.


Cheers,
--joe.
 
A little footnote to the excellent advice here: unless you have a disk-space issue, save the final greyscaled images as RGB jpg not Greyscale jpg, for two reasons:

- You'll generally get better results from printing or sending out for prints if they're in RGB.

- You can use your editor to tone the prints -- sepia, bluetone, etc.

Gene
 
jamiewakeham said:
Hi all.

I've uploaded a typical image to my gallery. I've done my best with it - applied (IIRC) +10 brightness and +15 contrast, followed by the GIMP's recommended unsharp mark.

The image just isn't doing it for me yet. I've got some blacks in there, and just about some whites, but the whole murky grey of the midtones is annoying me greatly. I'd rather see more black blacks and more white whites than all the murk.
I know what I'd do if I was looking at a test strip from an enlarger, but...
Click to expand...

I use the yiny-yang contrast more than most people here will admit to - but even I use it mostly for red eye (or yellow eye if you're shooting your cat) reduction. But for what you're doing, dont use it - use Tools->Color Tools->Levels instead. Auto does a marvelous job most of the time, but you can always pick your white and black levels manually.
 
Man, I'm slow. I've used curves to bring stuff out of darkness but I've never gotten it to work like levels as advertised here.
 
OK, a *long* night in front of the computer has left me feeling a little more proficient here. Thank you all for your hints. I'm a lot happier with how my images look now (give me a while and I'll upload some from the Zorki...)

One last q - I've been leaving everything in its rgb format all the way through. Given that I started with B&W negs, is there any real need or benefit to my converting the image to greyscale? I can't see any difference when I do!

Cheers,
Jamie
 
GIMP currently has 8-bit depth per color channel (meaning there 256 hues available for each). The transition to 32-bit channels is expected only to the version 3.

Whith RGB you've got 3 channels, while grayscale mode leaves you with only one. 3 channels provide a bit more room for levels/curves, hence the common suggestion.

GIMP indeed has channel mixer, and most of the common tools/plugins available in Photoshop. Oh and it is free as in freedom, not just as in beer* 🙂

Shameless plug: I've also writen a small GIMP script providing "classic BW" and "Infrared" filters for desaturation of color images: http://registry.gimp.org/plugin?id=4504



[*] In free software world, meaning that source code is free and available as well, not just the binaries.
 
Chuckle, CJP!

I *can* rig up my own darkroom, but it involves borrowing an enlarger and then barring my girlfriend from the bathroom for long periods of time. She expressed a strong preference, last night, for my scanning, GIMP-ing and paying for a snappy snaps to print the files over the 'trad' way...

I am apparently still allowed to do my own dev, though 🙄

Jamie
 
You must log in or register to reply here.
Back
Top Bottom