b&w conversion

[mrbubbles 0]

Ive been playing with a different way to get better bw conversions from my files. In raw i desaturate 100% then adjust the tonality with temperature and tint. Then i open in ps for final adjustments. Theoretically, i think i lose less information this way,I think the raw file is basically a greyscale image filtered for color. I seem to get a very pleasing and broader tonal range. Does this make sense, or is this in my little mind?

[acroporas 0]

You are thinking that since you dont care about the colour if you just treat each RAW pixel as 1 greyscale pixel instead of 1/3 of a colour pixel then you should see a significant gain in actual resolution. Right?

 

The would work great but because of the RGB filter covering the sensor, the image must be converted to colour even if you want a B&W image. There is no way around this. Even when your output is B&W you must apply the colour filter or you will get wierd patterns because of the colour filter.

 

I tryed to illustrate this. For example, you photograph a red wall. If you dont convert it to colour first you will get a mostly black image with grey spots.

 

I dont think there is any harm in converting to B&W in raw other than lack of editing options later, but I dont think you gain much over converting it to a colour 16bit tiff and then converting to B&W later.

[mrbubbles 0]

thanks, it seems like i get a better tonal range, then say using channel mixer, or converting mode to greyscale. Maybe its just subjective, or maybe its just less pixel degrading steps to do in ps. Try it and see what you think

[acroporas 0]

Are you working in 16bit when you do the B&W conversion in photoshop?

[Roulston 0]

I have purchased the following B&W plugin for Photoshop which is great. While you can achive the same with channels, this plugin takes a lot of the guess work out. Allow sthe application of particular color filter, multigrade paper selection etc.

 

http://powerretouche.com/

[frogfish 5]

If you make your image b/w when you convert from RAW, you're using someone else's general formula for translating RGB values into shades of grey. Sometimes it will work well, on other images maybe not so hot. Same thing if you use a plugin (though that may give you a few more options), or convert to a color space and then change mode to B/W in Photoshop. I think the formula Photoshop uses is something like 6 parts green to 3 paarts green and 2 parts red, or whatever - it's not really important. Another approach is to convert your image in LAB color space and just take the "L" channel - it's usually will turn out to be a pretty decent b/w version of your image.

 

Simpler, and often better, just to take the GREEN RGB channel. How? Image > Apply Image, set the source as the green channel and the target to the RGB composite. Instant B&W that may be better than Photoshop's canned conversion.

 

The reason this often works better than Photoshop's mode change algorithm (or converting the image from RAW into B&W) is that the Bayer screen your digital camera sets up twice as many green pixels as red or blue ones, so there will usually be a lot less noise in the green channel. and green often works well as a good B&W analogue. So you can sometimes get a cleaner image than something that averages all three channels according to some formula. But ther times, just relying the green channel doesn't work. Photoshop's canned conversion algorhithm doesn''t work for every image either, and it is almost never the best b/w conversion possible.

 

Why? Convert your image to LAB space. Click on the L (luminosity) in the channel menu and you'll see a b/w averaging all channels. Click back on Lab again and look at the color version. Did the image go lighter (too light) or darker (too dark?) Where did you lose contrast that you need to make the image work in b&w?

 

Better yet, open the info menu (still in LAB color space). It will give you LAB values and CMYK values for any point in your image. Move the cursor around and watch the L value in LAB. You can probably easily find two areas that are adjacent and contrasting colors ( blue and red, orange and green, yellow and blue, whatever) that have hugely different CMYK values (of course) but happen to have very similar "L" values. When you convert your image to b/w using any canned formula, those two contrasting colors will likely end up the same, or a very similar shade of grey.

 

The ultimate answer is that every image needs its own unique b/w conversion if you want optimal results with a full range of tonal values and contrast in all the places the image needs. Think about what someone like Ansel Adams, or Edward Weston, would be looking to achieve if he were handling your image. Try taking a photo of a green bell pepper with your digital camera in RGB, use Photoshop's canned algorithm to convert it to B&W. Does it look anything like Edward Weston? I didn't think so.

 

I'm sure a Plug-In would do better, but you don't really need it. But the way to get that optimal conversion can involve making what may seem strange moves (channel blending, apply image, applying steep curves to different plates in different color spaces and then blending these results back into other channels. But there is a logic to it. The important thing torealize is that there is no one single formula or easy answer that applies to all images. If that's what you want, just stick with Photoshops mode conversion, or just use the LAB L channel or the RGB green channel. It won't look BAD, but....

 

B/W (and duotint) conversons are covered in several different books on using Photoshop, but the best discussion of the strategies and techniques for b&w conversion I've read is in Dan Margulies' Professional Photoshop, which I consider a must-have book for all serious Photoshop users.

 

Frogfish (Robert Delfs)

[mrbubbles 0]

thanks frogfish, ill try those techniques. i have been experimenting with different conversion methods. i like chanels, but i would like to get better tonal range. for prints, a program called imageprint by colobyte software is incrdible but expensive and a pain to use. besides experimenting is part of the fun

[frogfish 5]

Have fun. Here's a couple of things that may help with getting the tonal range you're looking for.

 

1) Focus on the unwanted color. If you photograph a red wall, the red channel in RGB will be probably be almost maxed out - most of the detail, texture, and the gradations toward neutral that add up to 3-d modelling features in color will be in the other two RGB channels.

 

If the main subject happens to be a green bell pepper, then the green channel will be maxed out and the detail and shaping tones will be in the red and blue channels. Or you might want to convert to CMYK and see if most of the detail might be waiting for you in the magenta plate.

 

You want to eliminate or at least weaken channels with the most noise, and reduce the weighting of any channel that is maxed out and doesn't contribute useful data. Sometimes that can be accomplished by just trashing the channel that corresponds to the dominant color of the main subject.

 

While you're working, converting back to LAB once in a while to look at the L (luminosity) channel will give you an idea of where you are. Remember, while you're working with the different color channels, it doesn't matter what color things become - it's ok if the green bell pepper turns purple or orange or whatever, if that's what generates the tonal range in black and white that you're looking for. But that can be hard to see, which is why checking the L channel in LAB or looking at L values in the Info palette can be helpful.

 

LAB is Photoshop's "home" color space - when you change mode from RGB to CMYK or B&W, what actually happens is that the image is converted to LAB and then to the target color space, so switching to LAB doesn't cost much in terms of image degration. (Backing up after checking the L channel doesn't cost anything at all, and it's also faster.)

 

Robert