Color to black and white conversion in Photoshop - best way?

[bob carnie]

Stephan

As well , I should add that this method is not one Margulis teaches at least in my first two sessions and I have not seen it in his books , It is something that I have tried as our lab is in the business of converting to black and white as we are one of six labs worldwide with this harmon wet fibre paper that works from digital files.
The man from mars technique is in his books and is a wicked method for colour work , and I have applied it to BW.
As I am in Chicago with him next week on another three day marathon, I will show him this method and get his feedback.

Bob

[bob carnie]

Working with the numbers is the cats ass, once you get it everything falls into place.
In fact LAB density runs from 0-100 which taken in steps of 10 works perfectly for those use to the Zone System.
Zero A Zero B cannot be any easier.
Nuetralizing colours is IMHO the first and most important step in the PS workflow and this is where Dan is insistant to look at the numbers and I believe he is right.
Today we use Lightroom 2 for a white balance before we open in PS , but we still check key points to see if everything is balanced.
After Lightroom we are very close and getting dead nuetral is easy if that is required, than move on to all the other steps.

Back in the old pre-press days before color profiles were fully worked out and integrated into the workflow Margulis's techniques made a lot of sense, i.e. color correcting by numbers, not by eye. In fact I still do that. I haven't read anything from him lately but I respect someone who does rather than theorize....

[Stephen Best]

It should be noted that I have done over a thousand conversions using this method and I am not relying on someones word but rather than practical usage.
Have you tried this method? If not I would consider giving it a go and see for yourself, I would be interested in your observations.

Personally, I no longer go anywhere near Photoshop's Lab mode. It has holes you can drive a truck through, but if it's working for you well and good.

Dan is one of the sharpest guys around when it comes to colour (as opposed the most of the dullards that write books on Photoshop) but the problem is that he's in pre-retirement mode and will do whatever it takes to cover up Lab's holes so he can go on teaching his classes without dissent. I made the mistake of thinking that his colortheory group was for the discussion of advanced techniques when in reality it exists for the promotion of the Margulis brand. That said, his fundamental thinking is sound, it's just that his techniques have shortcomings.

We're getting pretty off-topic here, but Lab's main problem (other than large-space-itis) is that the results are totally dependent on the final RGB space and inevitably you'll be doing a colorimetric conversion back to this. If this is different to your monitor space, basically you're working blind. Increasing colour variation is a great idea but, as I noted above, in reality you're getting increased variation in some parts of the space, but decreased variation in others ... not the desired effect.

If you're interested though in handling tonality and colour separately, check out Lobster (http://www.freegamma.com). An updated CS4 compatible version is available on request. This is what I use. To increase colour variation with this, apply an S-curve to the Chromaticity component.

[bob carnie]

Working with Dans methods can be the most difficult and hard to understand, but I am constantly checking my numbers when in RGB and when in Lab, I do not find variations as you may be suggesting.
Dan will say himself that working with LAB is like using a sledge hammer to put in a finishing nail and at other times the only mode one should be working in.
Personally I am using LAB curves to separate the colours and not to correct colour.. for that I convert back to RGB for any final colour tweaks or adjustments.
But for separating complimentary colours, sharpening, contrast control I think there is no better place to be in my opinion , but backed up by many , many clients whose work has been improved by very simple LAB moves.
Back to the OP request, I strongly believe a move to LAB before any conversion will give pleasant suprises and possibly some unpleasant but that is the fun.
My only hope is that when I reach Dans age , I am as sharp and competent as he is. I have never gone on those discussion groups that you refer to as each session is enough for me to walk away and practice , practice , practice.

Personally, I no longer go anywhere near Photoshop's Lab mode. It has holes you can drive a truck through, but if it's working for you well and good.

Dan is one of the sharpest guys around when it comes to colour (as opposed the most of the dullards that write books on Photoshop) but the problem is that he's in pre-retirement mode and will do whatever it takes to cover up Lab's holes so he can go on teaching his classes without dissent. I made the mistake of thinking that his colortheory group was for the discussion of advanced techniques when in reality it exists for the promotion of the Margulis brand. That said, his fundamental thinking is sound, it's just that his techniques have shortcomings.

We're getting pretty off-topic here, but Lab's main problem (other than large-space-itis) is that the results are totally dependent on the final RGB space and inevitably you'll be doing a colorimetric conversion back to this. If this is different to your monitor space, basically you're working blind. Increasing colour variation is a great idea but, as I noted above, in reality you're getting increased variation in some parts of the space, but decreased variation in others ... not the desired effect.

If you're interested though in handling tonality and colour separately, check out Lobster (http://www.freegamma.com). An updated CS4 compatible version is available on request. This is what I use. To increase colour variation with this, apply an S-curve to the Chromaticity component.

[Stephen Best]

But for separating complimentary colours, sharpening, contrast control I think there is no better place to be in my opinion , but backed up by many , many clients whose work has been improved by very simple LAB moves.

If you're comfortable with Lab, you'll find Lobster very similar. But without the side effects. You can also retain your layer stack from start to finish.

[bob carnie]

Thanks Stephen
I will look into this, retaining layer stack? I take it then that Lobster is a all RGB workflow space then? Can I go into a Luminosity channel to do specific tasks?
Does sound interesting
Bob

If you're comfortable with Lab, you'll find Lobster very similar. But without the side effects. You can also retain your layer stack from start to finish.

[Gordon Moat]

. . . . . Lab's main problem (other than large-space-itis) is that the results are totally dependent on the final RGB space and inevitably you'll be doing a colorimetric conversion back to this. . . . . . . . .

Dan Margulis use to write about this in a now defunct publication called Electronic Publishing. One other interesting method involving LaB was to create imaginary colours, which would then be converted to a CMYK space for final printing.

Dan was a big proponent of CMYK and advancing press and pre-press techniques. That was when PhotoShop was largely a tool for graphic designs (not web guys). Unfortunately as graphic design became more a temporary employment realm, pay levels went down, and every weekender got a copy of PhotoShop, the market for teaching CMYK based techniques faltered, and seminars dwindled.

RGB largely was pushed by a new group, though partially the background criticism was that photographers were not sharp enough to understand CMYK properly (you should hear how often I heard that comment (and still do) at printing industry trade shows). So teaching RGB methods is now quite big in seminars and the photo-educator market. Dan can differentiate himself from the crowd by offering LaB techniques. These are interesting directions, but as anyone who has worked with PhotoShop on a daily basis for more than ten years knows: there is no one method in PhotoShop that is always the best choice.

Seriously, look at the Channels, and understand what you are viewing. Then combine whatever comfortable and repeatable method you can find to best exploit your starting file state (scan or capture). The final goal is the printed results, so your choices need to consider how an image is printed. Ideally, I would advocate that photographers learn more about CMYK, but that often seems to fall on deaf ears.

Ciao!

Gordon Moat Photography

P.S. - Parts of this post are quite obviously my opinion. I have a great deal of respect for Dan Margulis' knowledge and teaching; he is one of the few truly knowledgeable printing gurus left.

[Stephen Best]

Thanks Stephen
I will look into this, retaining layer stack? I take it then that Lobster is a all RGB workflow space then? Can I go into a Luminosity channel to do specific tasks?
Does sound interesting
Bob

Yes, all RGB. As such it avoids the vagaries you get with the choice of RGB space when converting out of Lab. This is an area Dan totally ignores in his Canyon Conundrum book. Lobster's Luminosity component can be used for the same things you use Lab's Lightness channel for (tonality adjustments, sharpening, dodging/burning etc). Many (most?) Lobster users are Lab refugees ...

Getting back to the subject at hand, you could place a B&W adjustment on top of the Lobster layer stack. This way, you've got the Chromaticity component for increasing colour variation, Luminosity for final tonality and the B&W layer for choice of filtration. You've also got the individual colour channels available for blurring noise (if required). It's hard to envisage any permutation of the B&W conversion process that the above doesn't encompass.

[Kirk Gittings]

At Stephens suggestion a few weeks back I downloaded Lobster. I really have not had time to use it much, but it looks very promising.

[bob carnie]

I get the chance to be with Dan twice this coming month, I will bring up Lobster and get his viewpoints, I hope he dosen't smash my monitor on my head.
It could be much like asking him what he thinks about histograms, don't even go there.
I too am going to give Lobster a go and as Gordon says work with whatever suits the image. I have been learning cmyk workflow, as I plan to separate colour images with bwcontone film on my lambda for alternative negs and am getting very excited about all the possibilities.