Please teach me BW conversion

[sagar]

Well I have seen some fantastic BW conversions here. But whatever I do, I just can't get that BW look and tone and my images look just pale, flat or just BW missing those tone. :banghead:

Here are some of my BW samples. I personally think these images could look much better if converted properly

http://www.flickr.com/photos/sagarjoshi/sets/72157624910544621/

Please help me learn BW conversion and some tips.

 

[tokengirl]

First of all, what software are you using to do the B&W conversion?
IMO, Nik Silver Efex Pro is the best software out there for B&W conversion. It's "Control Point" feature is truly amazing for dodging & burning and selectively controlling contrast. The "vignette" tool is also unique in that it lets you place the center of the vignette anywhere you want it.

I think that 2 of the 4 photos you linked to are wonderful photos and the B&W conversions that you did are not bad as a starting point. I am referring to "Path..." and "Rainy Evening in Paris". It would be helpful if we could see the original photos before cropping or conversion to judge how a B&W conversion might be handled for maximum effect.

 

[mediumcool]

Some aesthetic comments:

*signs of winter* is well-seen but too contrasty for my taste; the only reasonable detail is in the core background—the leaves burn out and the corners (burned?) are distractingly dark so the only texture in them pops out white.

*Falling leaf...* has the sort of tonality I really like—that rich grey (that’s how we spell it in Australia) B/G with plenty of texture to keep the eye busy. I'd like to see a bit more richness and detail in the leaf though ...

*path* is a little too grey and therefore a bit dull for me; I also find it a touch over-sharpened.

*Paris* is much too soot-and-whitewash. The burned-put details on the buildings and the solid shadows of the people and their brollies do not provide sufficient detail to keep my eye on them. I find it pleasing only in the wet pavement.

BTW, I made my first BW print in 1969, so I do have some experience, and I have taught the Zone System.

The old BW adage of “expose for the shadows and develop for the highlights” is inverted with digital shooting.

Digital is more like colour transparency film, where exposure is based on keeping the highlights from breaking and letting the shadows fall where they may (possibly using reflectors and lights to help—look at any pro movie shoot and there are lights/reflectors everywhere to make the shot look *natural*).

A question; are you using RAW or JPEG? If you use RAW you have much more control over tonal range when converting.

Interesting link about pulling detail from extreme under-exposure.

 

[mediumcool]

Also, most people do not calibrate their monitors (on this forum I suspect many more do so than elsewhere), so images can look very different on different systems.

I use Macintosh computers which have reasonable calibration control built-in; Windows has utilities to do a similar job I believe.

But hardware calibration is by far the best—I calibrate to a gamma of 2, a compromise between the old Mac standard gamma of 1.8 originally tailored for the print industry, and the video standard of 2.2 long used on the PC.

 

[Bob]

My basic process for B&W conversions.
disclaimer, everybody has their own "secret recipe".
First, do a raw conversion to a 16 bit tiff. make sure that there are no blown highlights or shadows other than what is intended. The image might look a little flat at this point. That is ok, in fact maybe preferable.
Second,
Open in Photoshop and add a B&W conversion adjustment layer. Click auto and look at the relative tones of the different previously colored areas.
Adjust these to taste until there is good tonal seperation in those areas of the image you consider significant.
Third,
Add a curves adjustment layer. Most classic film "looks" are due to a Hurter and Driffeld curve (density vs exposure) with a rather long toe and shoulder. Most digital images are more linear over-all than that and film curves are actually simulated. So now you get to emulate your favorite film curve. Forgive me for what some might consider toe and shoulder reversal, but it is to facilitate explanation.
For a "normal" look:
Start by shifting the black point so that you get some good deep blacks. Then add an adjustment point near the black point and drag it down and perhaps a bit to the right to lower the toe contrast and increase the mid-tone contrast.
Now adjust the white point just to the clipping point of the brightest part of the image. then add a shghlight adjustment point and drag it up and a bit to the left to emulate the shoulder of the curve. this further increases the mid-tone contrast.
Tweak these points to taste.
Since these are all layers you can fiddle with them to fine-tune the result.
Voila!
if you would like some grain, then stamp visible to create a layer with all of these adjustments applied on top of the curve layer. then filter/noise/add noise to taste. Dial down the opacity of this layer to where you want it.
That ought to get you in the ballpark.
-bob

 

[pfigen]

The vast majority of the black and white conversions I do are done with a combination of Channel Mixer, Curves Adjustment Layers and doses of hi-radius, low amount USM. I've done enough comparisons to the B&W Adjustment say that in most cases the Channel Mixer just looks better. In addition, the B&W Adjustment has a fatal flaw. All too often it causes posterizations and rough tonal transitions that are not there in the CM conversions. Oh, yeah, it looks good at first glance, but zoom in on a high res image and you'll easily see how crappy the gradations are. Good idea. Not quite there in implementation. In my opinion, another tool that was added without fully being tested by people that really understood what to look for.

For those who were raised on black and white film, and remember how filters afftected the tonality, the Channel Mixer will make the most sense.

 

[TRSmith]

All of these suggestion work and can even be combined. I don't think there's any one "right" way to do it. It does seem to help if you have a clear vision of what you're after though.

And I think some healthy experimentation is a worthwhile investment of time. Crank those sliders all over the place to exaggerate effects and see what they do. For instance, if you're using the Channel Mixer method, you can manipulate the color to subdue or enhance portions of the image before implementing the Channel Mixer. Lots of ways to release your inner B&W!

 

[Jack]

The most recent B&W adjustment layer in CS is essentially a channel mixer anyway, allowing you to adjust tonality in your B&W conversions based on colors. I do what Bob does. Here, I would reiterate that if there's a "secret sauce" it is the final output curve with a toe and heel that matches your favorite B&W film or your artistic vision -- and usually it's a much higher midtone contrast setting with flatter toe and shoulder than a typical digital file has. The only other secret suace tip is to boost the saturation of your color image by about 8 or 10 points before the B&W conversion, as this allows a more definitive separation of tones while tweaking the channel slider sliders in the CS B&W adjustment layer tool.

Cheers,

 

[pfigen]

Jack - While I agree that the B&W is a sort of Channel Mixer, that fact that it adds more color options to the mix is also its weakness. The smoothness of transition between those added color tweaking options, and how they actually blend on a high res image, are the ultimate downfall. Add to that, that most users make there B&W adjustments looking at an image that might be displayed at 16.7% or less on screen, makes the artifacts impossible to see. Sure, everyone can and should use whatever method they want, but they should also be aware of the limitations of the method they choose as well. My first reaction to B&W was, WOW, this is freaking cool. That initial enthusiasm was soon tempered by reality and I move almost exclusively back to CM. It just looks better.

 

[Jack]

Like anything, one needs to learn to use the tools to maximum benefit while remaining non-destructive. Yes it is "easy" to impart artifacts in any B&W conversion that uses selective color channels, especially if you boost the Blue channel. Again, careful application will keep those to an acceptable level in the image.

 

[meilicke]

This is a great conversation, thanks all.

 

[Jack]

and then there is the gold standard:

:ROTFL:

(PS: Image is actually an in-cam B&W jpeg from my Panny G2 with 45 macro lens shot wide open. CS work includes curve per discussion above, slight edge burn, a little D&B here and there, and a slight perspective tweak... )

 

[mediumcool]

Another technique I have adapted from my film days is dodging: because the PS dodging tool can be set to affect the highlights, mid-tones and shadows respectively, used subtly it can improve or simulate texture in shadows (I mostly use it set to highlight, and never to shadow).

Worth trying if a large and distracting shadow area seems completely filled in.

 

[sagar]

Thank you folks, this is indeed very helpful discussion. Here is the original color image of one of the photos. If you don't mind, would you please play with it and help me show how it would look in BW if you did the conversion with your methods.

 

[mediumcool]

Just whipped this up (bit of layering/masking going on)

 

[mediumcool]

How I did it.

Opened image in PS2.

Quickly inspected each colour channel (Command-1, Command-2, and Command-3). Often Green looks great, but here RED and GREEN were about equal (BLUE tends to look dark and also noisy)

Opened Channel Mixer dialogue, set to mono output with 50% each of RED and GREEN, based on preliminary inspection above — changed to 30% R and 70% G to make up 100%.

Wanted more detail in highlight up the top so selected all (Command-A) > copy to new layer (Command-G) and set its transmission to multiply, but gave it away when I couldn’t quickly mask out the effect on the darker areas (using the image itself as a layer mask). Wasted those seconds.

Next, duped layer again, and used dodger to brighten blurred brollies and tried to separate the legs and feet from the B/G (fun aspect of this shot). Used this new layer so as not to destroy original.

Then duped the dodged layer again, applied 3 pixel Gaussian blur and set its transmission mode to lighten.

Then put a Levels adjustment layer on top and pulled the middle slider down to 0.90.

This is on a Mac, so substitute the Control key on the PC.

 

[mediumcool]

A variant with a gradient layer mask applied to the blurred layer so the foreground glint (?) is not haloed.

Subtle but there.

 

[Bob]

View attachment 37688
My version.
B/W adjustment layer, auto
Curves adjustment layer - brighten mid tones, masked out some background areas
Curves adjustment layer, increase midtone contrast, long toe, long shoulder
Stamp visible
selective sharpening to lady near car to help visual point of interest and some foreground reflections
-bob

 

[mediumcool]

Bloody interesting (Australian adjective) how you can use so many different techniques to solve problems/create your own vision. Bob, I like what you’ve done, but find it a wee bit contrasty for my taste (at least on my monitor—I do love zones 4 to 7!).

Like what you did to the woman in centre BG with brolly; I just dodged to brighten it a bit but the sharpening is another way of creating attention. When you think about it, it’s quite a busy image.

Just googled B+W adjustment layers—CS3, the version after mine, was the first to get it.

What time is it over there? 11.18pm on a humid late spring night here in South Australia.

 

[mediumcool]

Bob

I’ll have to modify my contrast statement after another look at both our attempts; the romantic ghosting I have engineered pushes grey out from the highlights so much that I think if it wasn’t for that, our contrast ratios would be much closer.

I was trying for a fin de siècle look (hey, it is Paris!).

Too late in the evening; too much inexpensive domestic chardonnay! Hic!