B&W, JPEG, laziness

[fivetonsflax]

Hi folks.

 

I use a Canon 20D and a Mac. I prefer to shoot JPEG, for convenience's sake. But lately

I've gotten into shooting black and white. Now, as far as I can tell, a B&W JPEG has only 8

bits of tonal range. That's weak! So should I shoot RAW (and archive in DNG or TIFF), or

shoot color JPEG and convert to TIFF before taking out the color?

 

Any of these options is more hassle than my current workflow (or lack thereof). If I could

fully automate the process, I could live with it. I don't mind using Photoshop or iView

Media Pro or whatever else fits the bill.

 

Are there are other lazy B&W digital photographers out there? How do you cope?

 

Thanks.

[pieris_berreitter]

A color JPEG also has 8 bits of tonal range. You would gain nothing (other than the ability to post-filter) by shooting color JPEG.

 

In B&W you have no chrominance variations to distract from luminance variations, so careful control of luminance range is even more important than with color. Shoot raw.

[steve_bingham]

Shoot color and convert to B&W later in PS. Now you have the best of all worlds. By shooting raw you will increase resolution, reduce chromatic aberrations, better balance color temperature, and reduce jpg artifacts, halos, etc. Then when you convert to B&W you can adjust color channels to emulate the use of filters and produce wonderful B&W prints full of exciting tonal values. Whoops. I forgot. If you are lazy this won't work. It will require some studying. Sorry. :^)

[fivetonsflax]

Thank you both for your answers.

 

Steve, I don't mind putting work into an image that I want to print -- I just don't want to

have to fiddle extensively with every single image before I can even look at it to evaluate

it. See what I'm saying?

 

Pieris -- are you sure? It seems to me that you could map 24-bit color into 24-bit

grayscale, but I could be wrong.

[ken_hughes4]

"Then when you convert to B&W you can adjust color channels to emulate the use of filters and produce wonderful B&W prints full of exciting tonal values"

 

I'll believe it when I see it.....

[pieris_berreitter]

Beautiful b&w prints are made every day starting from color digital files. But back to the original question: with b&w you are only concerned about luminosity. Assume a colorless scene shot in color. There are 8 bits per channel but each channel reads the same amount (giving a shade of gray). There is no way you can extract more than 255 distinct levels of brightness from this image. When we say an image is "24 bit" we mean "24 total bits, consisting of 8 bits each of R, G, B".

 

If in doubt, do a quick test; compare a desaturated color JPEG to a B&W JPEG. The camera may actually be applying a contrast curve that you find to be superior to your desaturation results.

[digitmstr]

I wouldn't be so quick to dismiss the 20D's ability to record images in B&W - having said that, if you shoot in color you'll always have BOTH available to you. While you can convert color to B&W the opposite is not possible. For that reason alone I'd shoot in color.

 

However, if for whatever reason you had to shoot in B&W to satisfy your "laziness" :) why worry? When you feel lazy shoot B&W JPG, when not shoot RAW color and save in TIFF.

[michael_spencer3]

Ben Rosengart wrote: "...If I could fully automate the process, I could live with it".

 

Hm, I think that the "fully automated" process is called JPEG! RAW is not for the lazy or for folks who want stuff automated. It is for the exact opposite -- folks who really and dearly love post-processing and who are willing to fiddle and tweak until the cows come home.

 

Steve Bingham write "... By shooting raw you will increase resolution, reduce chromatic aberrations, better balance color temperature, and reduce jpg artifacts, halos, etc".

 

Steve forgot to mention that you will also grow back your hair, lose those extra pounds without excercising, AND eliminate those irritating sales calls during your dinner hour. Did I mention low cost drugs???

 

Be Calm, Be Happy,

 

Mike Spencer

[fivetonsflax]

Pieris -- most scenes aren't colorless. You are right that a colorless scene recorded as a

JPEG will have no more than 256 levels of brightness. So will a colorful scene that's been

converted to a B&W JPEG. But you could convert a color JPEG to a B&W TIFF or something

else, and have more than 256 levels of brightness. (Not 255 -- you forgot zero.)

 

Here's a thought experiment. You have a color JPEG -- 8 bits per channel. Let's say it's

only 1 pixel in size. Take the least significant bit from each channel, average the values,

and record the result as a three-bit value. Then repeat for the second-least-significant

bit, and so on for all the bits.

 

You now have a 24-bit value. It only contains 16 bits of information -- put another way, it

has 2^16 possible values, not 2^24. But that's a significant improvement over 2^8.

 

(Why 2^16? Because those three-bit values will always be 0/3, 1/3, 2/3 or 3/3. You're

recording the number of bits that were turned on in that position, but not which channels

they came from. It's not that this operation doesn't throw away information, it's that it

throws away 8bpp not 16bpp.)

 

Ack, my head hurts.

[jeff medkeff anchorage, a]

<p><i>There is no way you can extract more than 255 distinct levels of brightness from this image.</i>

 

<p>Of course raw (on a 20D) will get you 4095 levels per channel. That's got to provide a lot more elbow room in processing than 255 levels. (Of course the 12 bits are mapped [after conversion] into a 16 bit format.)

 

<p>Ben, I also have an interest in automating this. I've thought about writing some photoshop action that takes a TIFF, and presents me with five or six variations of the same image based on different desaturated channel mixer settings. Has anyone here fooled around with such an idea?

[fivetonsflax]

Well here's a related question.

 

Has anyone got a technique in Photoshop that can convert a color JPEG into a B&W image

with more than 256 levels?

 

I tried something extremely naive: open the JPEG, change to 16-bit mode, and either

desaturate or change to grayscale. Judging from what I could see, this did not work. (I

eyeballed it side-by-side with the same JPEG, also desaturated/grayscaled, but still in 8-

bit mode.) If my previous ramblings were correct, though, there oughtta be *some* way to

do it. (Leaving aside the ridiculous notion that something might be theoretically possible

but not possible in Photoshop. ;-)

[ken_hughes4]

"Beautiful b&w prints are made every day starting from color digital files"

 

average maybe, have yet to see beautiful.

 

problem is that you are shooting bw with the restrictions of color capture, digital or not. digital/color film has less of a tonal range for capture than bw film. I can pull 11 stops with normal dev.--up to 15 stops with dilute development, not to mention how filters affect the image (at the moment of exposure) as compared to it being an afterthought. going into the computer is not going to gain information that was not already there.

 

This is why after having used photoshop almost as long as it has been around and having printed all my images digitally-- both bw and color-- for several years, I still shoot bw film when I want a bw image.

 

face it, the demand for high end bw from a digital camera is just not there. bw for digital cameras will remain a stepchild to color for a long time if not forever. chances are not good that anyone will ever make a digital camera dedicated to fine bw quality.

[fivetonsflax]

Ken -- a filter doesn't add information, it removes it. So it doesn't matter, from an image

quality perspective, whether you put the filter on your lens or do it with Photoshop.

 

As for digital black-and-white, you're probably right in general, but I still want to try it.

 

I can't help wondering how nice it would be to try a digital camera with no Bayer filter.

Shouldn't that buy you something like a stop and a half of sensitivity with zero extra

noise?

[Spearhead]

Ken's comments show an assumption of a single methodology of printing black and white, and it's important to point out that many prints do not require every tone in the universe to look good. In fact, many great digital black and white prints have been made that are high key, low key, high contrast, rather than Ansel Adams long tone prints. Daido Moriyama's prints, just to give one reference, have very different requirements than "15 stops."

 

There are things that can be done with inkjet prints that are extremely difficult with traditional printing, for example, printing on watercolor paper. One can get new variants on "alternative process" prints, most of which also don't meet Ken's requirements but can have a beauty of their own.

[ken_hughes4]

sorry Ben, nitpicking about the physics of filters does not change my opinion. working on an image before or during it's being taken is very different from working on an image after it is taken.

 

Jeff, you make huge assumptions as to what is "acceptable" to me, I was making the point that a color to bw conversion does not equal "B&W prints full of exciting tonal values" as someone else claimed, it's just not true. I never said that every print had to have 15 stops, just making a point about the advantages of bw film over digital and how it pertained to what was said. to say that I require every tone in the universe for a print to look good is just plain silly.

[ken_hughes4]

Ben, I guess a good way to put it would be to ask you if you really think that an image on your screen in photoshop holds as much information for conversion to bw as the real world holds for me with a camera loaded with bw film?

[Spearhead]

All I was doing, Ken, was following your logic in your 2:38 post.

[ken_hughes4]

Jeff, my 2:38 post does not stand alone as my only comment but is part of a larger "conversation" so my response would be that you took me out of context.

[fivetonsflax]

Ken,

 

I said that doing the filtering in post-processing didn't matter _from an image quality

perspective_. You seem to be saying that it's different from an artistic perspective. I don't

disagree in the slightest.

[jeff medkeff anchorage, a]

<p><i>[do] you really think that an image on your screen in photoshop holds as much information for conversion to bw as the real world holds for me with a camera loaded with bw film?</i>

 

<p>I've rarely seen such a misleading statement. There is a difference of many orders of magnitude between the information content of a real-world scene and the amount that is recorded by b&w film. There is even a difference of many orders of magnitude between the information that could potentially be captured by b&w film and that is actually captured by b&w film.

 

<p>The amount of information captured by B&W film must be far closer to the amount of information captured by a 20D than to the amount of information available in the real world scene (but this should be no surprise, considering how poor b&w film is at recording information). I personally suspect that in most scenes, a 20D image captures less information than a black & white film image, which is to say that in terms of significant bits recorded, film sucks less. Which moves us on to the next question: Who cares? If the doctrines of the Holy Church of Black & White don't recognize that filters subtract information no matter where in a process they are used, or that potential information capture in B&W is quantitatively orders of magnitude different from information availability in the real scene, then I think we can afford to be less than impressed with the factual basis of those doctrines and dismiss them as fairy tales.

 

<p>More importantly, I see no reason why Ben's wanting to fool around with B&W and his 20D should cause anyone else's britches to bunch up in a manner either pro or con. If we could get away from making such thunderously obvious, uninteresting, and intrusive statements as "b&w film is different," then we might have a shot at discussing workflow here in the <strike>B&W Film Forum</strike> <b>digital darkroom</b> forum.

 

<p>Can anyone suggest an improvement upon this: shoot raw, convert to 16 bit TIFF, use the channel mixer with desaturation on to find the contrast you want. For Ben's benefit, there are ways to automate the raw conversion and, inspired by his question, I'm working on a channel mixer action to take some of the tedium out of sliding the sliders on that tool which I'll give away if it bears fruit.

 

<p><i>I can't help wondering how nice it would be to try a digital camera with no Bayer filter. Shouldn't that buy you something like a stop and a half of sensitivity with zero extra noise?</i>

 

<p>I dimly thought that Kodak actually made such a camera. In any case, all my science grade cameras lack both Bayer filters and IR blocking filters. I can't vouch for the amount of sensitivity gained but it sounds like you are in the right ballpark, and you are correct that you get zero extra noise. Lack of color sensors also has some implications for antialiasing and you get a filter less prejudicial to high frequency detail, as well. On the other hand, you get a one-trick pony of a sort that is disadvantaged even more than a film camera loaded with B&W film.

[Spearhead]

Jeff's comments seem to say it all.

 

Regarding my earlier comments (note this separator to make it clear that people know what I am talking about), I'm no mind-reader. There are standards for writing, including devices that make it clear when there is a topic change. However, you still managed to ignore the rest of my comments.

 

But I'll let the other Jeff's comments say it for me.

[pete_su]

Kodak did make a B&W only version of one of their F5/EOS-1n based bodies,the 620 or

something. I think you can still buy them, but they are expensive.

 

They also weigh about 10 pounds and are the size of a small laptop.

[ken_hughes4]

"You seem to be saying that it's different from an artistic perspective"

 

no, I am saying that information lost with color/digital from the moment of exposure is information that is still available to someone who shoots bw film from the start.

 

 

 

"There is a difference of many orders of magnitude between the information content of a real-world scene and the amount that is recorded by b&w film. "

 

of course there is, I never said that bw film was capable of recording all of the information, only that more information is capable of being recorded on bw film as opposed to the amount of information that is available for a bw conversion from a color/digital file.

 

 

 

 

"If the doctrines of the Holy Church of Black & White don't recognize that filters subtract information no matter where in a process they are used"

 

of course filters subtract in both processes, only there is a significant difference in how much change takes place and the level of quality that comes of it, how about skin tones as an example?...maybe the doctrines of the holy church of digital could learn a thing or two.....

 

 

 

 

"I see no reason why Ben's wanting to fool around with B&W and his 20D should cause anyone else's britches to bunch up in a manner either pro or con"

 

how do you manage to comment when you haven't read the conversation... I never said anything pro or con about Ben doing color digital to bw conversions...more power to him. I took exception with the statement, "B&W prints full of exciting tonal values", sorry, I don't think that is very accurate. I also said more than once for those who don't read the entire conversation that I don't think that every image on the planet has to have a huge tonal range to qualify as a good print.

 

 

 

 

"B&W Film Forum digital darkroom "

 

since the conversation is still here I would suppose it is relevant, unless you are simply afraid of being challenged on your biases.

 

 

 

 

"you get a one-trick pony of a sort that is disadvantaged even more than a film camera loaded with B&W film."

 

further evidence of your bias, which makes your opinions irrelevant.

 

 

 

 

 

"However, you still managed to ignore the rest of my comments"

 

sorry, I don't think that I ignored any of your comments, please elaborate.

 

 

 

everybody gets in a huff, some even try to insult me and I am the one getting my britches in a bunch?? funny

[jeff medkeff anchorage, a]

<p><i>of course filters subtract in both processes, only there is a significant difference in how much change takes place</i>

 

<p>That is simply not true. A filter has a certain function, which it always carries out regardless of where it happens to be located.

 

<p>Does this mean that putting an optical filter in front of a B&W film camera lens is the same as applying an equivalent algorithm to a 20D's digital file? No, of course not - the results will be different, <i>even if</i> both the algorithmic filter and the physical filter have <i>perfectly identical</i> effects. Why? Because different media with different recording characteristics have been used.

 

<p>Although you don't get there through a correct understanding of information theory, you do have the right answer in the end. But your answer reduces, on analysis, to the claim that "B&W is different," and that is pretty obvious.

 

<p>What really bugs me is that you leave us guessing about the implication of this difference. If you tell someone, "there is a runaway bus headed straight for you," the obvious implication is they should get out of the way, and fast! Telling someone that they dropped a one hundred dollar bill has the obvious implication that they ought to pick it up unless they want to become a hundred dollars poorer. Telling someone "B&W film is different from digital" may be just as true as either of these mundane statements. But what is the call to action here, now, in this thread?

 

<p><i>and the level of quality that comes of it, how about skin tones as an example?</i>

 

<p>I believe that black and white film and its resulting images are both qualitatively and quantitatively different from digital and its images. But where are you going with this? The photographic equivalent of the bus I am to jump to avoid or the dollar bill I am to pick up is what, exactly?

 

<p>And indeed, <i>what about</i> skin tones? Can you explain what we're supposed to understand by your bringing the subject up? We all know perfectly well that B&W and color digital capture skin tones very differently.

 

<p><i>...maybe the doctrines of the holy church of digital could learn a thing or two.....</i>

 

<p>I personally find that church's doctrines to be more infuriating than the film church's doctrines. But my personal feelings aside, I know that neither church's members understand their medium, nor do they respect it. Both groups are posessed of large quantities of urban legend and myth which passes as fact about their mediums. Many are also intolerant, antisocial, and prone to taking the slightest statement as an indication that their favored medium is insufficiently valued. Such people might get snippy if I bring out my film cameras, my digital cameras, my B&W film, my color print film, my slide film.... What an insufferable bunch of meddlers. I yearn for their reform. Thank goodness they are a minority.

 

<p><i>how do you manage to comment when you haven't read the conversation... I never said anything pro or con about Ben doing color digital to bw conversions...</i>

 

<p>With my apologies to you for my being unclear, I did read the conversation, and I do not believe you have said anything pro or con about Ben doing such conversions.

 

<p><i>I took exception with the statement, "B&W prints full of exciting tonal values", sorry, I don't think that is very accurate.</i>

 

<p>The <a href="http://dictionary.reference.com/search?q=accurate">dictionary definition of accurate</a> is "conforming exactly to fact," among other similar things. But what tonal values one finds "exciting" must be dependent upon personal preference and subjective value judgements, don't you think? (Or is there now a standard for art appreciation?) So it would seem that you can not take issue with that statement on the grounds of accuracy. You can only take issue with it insofar as you, personally, do not find some tonal values (which presumably digital b&w images are full of) exciting.

 

<p>Which, I guess, is no big deal, unless Ben has set out very specifically to please <i>you</i> with his conversions. But this is a strained hypothesis that I cannot support on the evidence available.

 

<p><i>"you get a one-trick pony of a sort that is disadvantaged even more than a film camera loaded with B&W film."

 

<p>further evidence of your bias, which makes your opinions irrelevant.</i>

 

<p>I'm glad you made this comment. Who is possessed of more, or more damaging, biases: Those who acknowledge that there are potential disadvantages to a camera loaded with B&W film? Or those who cannot acknowledge that there are such disadvantages, and brand those who think there might be disadvantages as unacceptably biased?

 

<p>Most reasonable people would agree that a camera loaded with B&W film has a big disadvantage if you want to shoot color digital photos and later convert them to B&W. And this example is relevant precisely because the OP's original question was about doing just this!

 

<p>As regards biases more generally.... If their presence really rendered opinions irrelevant, none of us would have an opinion worth listening to, since we all have biases. The claim that biases render opinions irrelevant is preposterous, of course. I have a lot of biases both in favor of, and against, the idea of digital photography, and of doing black and white conversions of color digital photos. But I am able to overcome these biases to provide some meaningful advice and information about doing those conversions. Can you overcome your biases, whatever they are, to similarly make a constructive and utilitarian contribution?

[jeff medkeff anchorage, a]

<p>Suffering from the same sort of laziness that Ben describes, I've always been negligent about evaluating my digital shots as black and white images. The main reason has been that I haven't liked the tedium of playing with channel mixer sliders in Photoshop.

 

<p>As promised (or threatened), I've created an action to partially address this issue. It is described and is downloadable from <a href="http://jeff.medkeff.com/photo/actions/">here</a>. It is really simplistic, basic, stupid, and brute-force, and has probably been done before many times over, but what the heck, it didn't cost me much. Please let me know if you have any problems.