The Online Photographer

Check out our new site at www.theonlinephotographer.com!

Wednesday, January 03, 2007

Post-Processing

Recently I seem to be getting a lot of questions from people about, and requests for information about, post-processing. (This probably came up because of my recent preoccupation with making and selling prints, and from posts such as "On Printmaking.") So I plan to do a few posts in the coming months on the subject, with some concrete examples.

The problem of post-processing is twofold: 1) What to do, and 2) How to do it. I tend to be a little too much interested in part 1 (my skills are a little too seat-of-the-pants), and most instructional books are skewed pretty radically toward part 2. But the two things really arise together. They also inform each other: what you know how to do influences what you do, and—similar but not the same thing at all—the effects you feel you need to achieve can drive you to go learn how to do it.

Illinois #20—Mike's version

I posted my picture "Illinois #20" for sale a couple of days ago (it's been selling like hotcakes*), and my e-friend Stephen Best of Macquarie Editions wrote from Australia to say he found it too "bleak," asking me if I wanted him to have a go at it. I happen to really like the print a lot—I'm sort of besotted with it right now, actually, in the way you get about the things that come along occasionally to "nourish your enthusiasm" (Ansel Adams's phrase). I keep wandering over to look at it. And one thing I like about it is its bleakness and apparent emptiness. But I'm never averse to learning—especially from people who know what they're doing—so I took Stephen up on his offer.

Illinois #20—Stephen's version

Here's what he came up with. He was working just with my posted JPEG, and he was also influenced somewhat by my interpretation, which of course he'd already seen. So it's possible—likely, even—that he would come up with something different if he were to start from the RAW file with no guidance from me and only his own taste to satisfy.

Stephen basically converted to LAB color and ran the procedure that Dan Margulis recommends in the first chapter of his new book on the subject. After converting to LAB, Stephen ran a Haze Reduction action to increase local contrast. Then, using a Curves layer, he first adjusted the L (luminance) curve to tweak tonality, and then steepened the a and b curves ("a lot," he says) to "bring out the latent colours in [the] flat original." Lastly, he added a gradient fill to the layer mask to stop the top of the image from blowing out. Once it was back on my desktop I reconverted it to RGB and, as Phil Davis never used to say, viola.

So what do you think? Here's what I think, provisionally: I think my version preserves the look and intensity of the light a little better—it was near dark and the light was murky—but Stephen's is actually more accurate in terms of color as I remember it. The sunset, behind hazy clouds, was behind my right shoulder, and you can see that in Stephen's interpretation.

So why "provisionally"? For the same reason that you only know what you think provisionally—I haven't seen prints side-by-side yet (and neither have you). To really decide, I'd have to try to replicate his effects on the RAW file and reprint it to his vision of it, and then live with the two for a bit.

There are two very basic lessons of post-processing that cannot be stressed enough: First, in interpreting a photograph as a print there is only bleh, fair, fine, better, better, better, and different: there's no best. (You can quote me on that. Please.) The interpretation an expert settles on will likely be better than the best interpretation a neophyte can come up with, but half a dozen experienced photographer/printmakers could conceivably come up with half a dozen different versions of the same file; it depends on how you think the particular picture works best, how you want it to communicate, and what aspects of it you like and wants to bring out. The second lesson is: You have to see it. Never skip the looking stage of the process. I think I like my version better than Stephen's, and I'll bet he believes the same thing in reverse. But unless and until we each actually see prints side-by-side, final judgment has to be reserved.

Posted by: MIKE JOHNSTON with STEPHEN BEST

*Have you ever bought a hotcake? Well me neither.**

**See William's comment and my response.

Featured Comment by advman: Sorry Mike, I couldn't resist. Here is a version I just did. You may hate it, it is almost certainly not what you've seen, but that's what I think would sell.

I made this with Lab as well, only not so decent, and added three curves adjustment layers to adjust contrast differently for the forest, the city, and finally the top of the sky.

I have no idea how bad the effect on noise would be. That is always the problem with so drastic contrast adjustments, but you might well get away with it. It is only crucial that you don't use any noise reduction on the forest. Micro-detail in the forest and noise reduction, they don't go well with each other.

25 Comments:

Blogger Guy Batey said...

I like your original a lot Mike -it's my favorite of the photos you've offered so far. But then I have a thing for apparently empty geometric bleakness.

The 2nd version just looks a little over-cooked to me.

1:40 PM  
Blogger Gordon said...

This all presupposes that a print is the final output (which I know it is in this case, because you are selling those hotcakes). But it doesn't have to be - in fact, I suspect that many more images end up as digital files than prints these days.

Someone will no doubt open a gallery exhibition with HDTVs or LCD monitors on the wall, instead of prints soon (or may already have done that). The resolution might even be there to better prints at some point in the nearish future.

So - is the digital file just a step in the process, or the best representation of the image (given that prints are not always - or ever - going to maintain the luminosity and colour of the edited end product, on the screen.

After all, you edit the digital image to the best of your vision - or do you edit the digital image with an eye to how those changes will impact the end print ? I'm sure different people approach editing in different ways, some with an eye to it looking as best as possible on the screen, others with the constraints of their print media in mind.

2:13 PM  
Blogger William said...

"*Have you ever bought a hotcake? Well me neither.

Hundreds of times. Usual order ones with blueberrys when I have to eat breakfest while on the road. In some parts of the world they are called pancakes.

2:32 PM  
Blogger Mike Johnston said...

Gordon,
I certainly "post-process" images for the web all the time (most every day, in fact, for this blog). The steps and procedures might be a little different, but the principles are the same. Namely, you can do a little or a lot; you can do it badly or well; and a lot depends on how you want it to look. I don't see that there's any fundamental difference regardless of the ends.

--Mike

3:03 PM  
Blogger Mike Johnston said...

I'll be darned--learn something new every day. I never knew that, William. I guess I HAVE bought a hotcake then.

And in that case, the prints are NOT selling like hotcakes. But I'm going to have a half price sale on Friday. Maybe they'll sell like hotcakes then. (s)

--Mike

3:06 PM  
Blogger eolake said...

I think part of the whole raison d'etre of the picture is the flatness and subtlety.

3:18 PM  
Blogger Jeff Kott said...

I always post process for prints, not the web. My favorite prints (and the prints I get the most oohs and aahs over) seem to be the ones that I do the least post processing on (with the exception of exposure comp., white balance correction and sharpening). I experiment a lot, but tend to go back to a lesser processed version. It's probably just that I'm not that accomplished in PP, but I'd be interested in knowing if others feel the same way.

3:19 PM  
Blogger matt said...

On the subject of "interpretive printing" I remember when I had a class in fine black and white printing, and our instructuor gave us all identical negatives (I think he shot a roll of each scene to have enough). Some of the prints that we had at the next crit were so different that it was hard to tell that they came from the same scene. And until I worked in as a custom printer in a lab (color RA-4) I never knew how far that medium could be pushed. I'm consider myself a streight printer, using minimal controls, but after that job I find myself tweaking areas of the image more, using 7-8 steps on some (wet) prints. In the computer it's the same.

FWIW, I prefer the last version of the skyline image... but my opinion might be different if it were printed large (for some reason I keep thinking of a Richard Misrach type of thing) instead of a web sized jpeg.

4:53 PM  
Blogger neumero4te said...

Mike, how well you can predict how the final print will look by looking at the image on your monitor?

5:00 PM  
Blogger DaveP said...

This discussion brings to mind a recurring article I enjoy in Black and White Photography magazine (GMC Publications) titled “the printer’s art”. Two photographers each produce a print from the same negative; the article documents the approach each takes and shows images of the work in progress plus the final version. In some cases the results are similar, in others quite different; neither result is right or wrong, they simply represent the individual interpretation of the image recorded on the negative. A similar approach done with RAW files would be interesting.

5:04 PM  
Blogger Max said...

I agree with Eolake's opinion. Also, Mike has some unique data that for some will work as a "prejudice", he was the only one that witnessed the original scene, so he had an idea of what he wanted to capture. So he's less "free" in what he considers "fair" post-processing.
Besides that, there's another point that I find very important in this kind of decisions, the size of the print. If the noise, definition, etc, permitted a big print size, the subtle version will look a lot better in my opinion. Radius of all these corrections produce results that are perceived differently at different sizes. What we feel brings out information in the web version will obviously be overkill in a big print. The bigger the print, the more the subtle detail comes out without needing too much local contrast and color enhancement. Does this make any sense to you?

6:37 PM  
Blogger Thomas Passin said...

OK, I had to join the fun. Here's my effort:

http://www.tompassin.net/pix/tbp_illinois20small.jpg

I pulled the trees and nearby buildings out of the haze a bit, darkened the sky a bit, and lightened the trees a bit. The Chicago skyline seems more ghostly to me, more - perhaps - painted and so a bit more eerie. 'Course, it's hard to tell, what with all the jpeg artifacts in Mike's first version.

To me, I like the more ghostly skyline and the dark but less hazy foreground.

I worked in Photoshop elements. I used mainly local contrast enhancement, levels, and the lighting adjust dialog where I darkened to bright tones, increased midrange contrast, and lightened the trees a tad. No masking, no gradients, no channels. All adjustments were small.

7:21 PM  
Blogger thoughtdujour said...

love you site. my feeling on processing is that I make changes that i feel are appropriate to ensure that the print version is powerful and emotive. But that's just what I like...

7:40 PM  
Blogger NIMBY said...

davep raised a point I have been thinking would be a great idea for a while - and that is extending the "Pritnter's Art" feature in B&W Magazine to include a print made of the same image using a digital workflow. I think that would be very interesting and would also be very relevant in these times. I realise B&W Magazine tread a fine line on "going digital" and God forbid they go the way of many other magazines, but I do believe that 2 wet print interpretations and a scanned/photoshop/inkjet interpretation would be very informative.

Mike, any chance you could suggest that?

Jules.

12:55 AM  
Blogger Dierk Haasis said...

While there can be a wrong interpretation I am not quite sure - still not after all these years - if there can be better ones among those that are not wrong. The main purpose of communication is to find other perspectives, new subtleties or facts, learn from others. Thus any interpretation that is not wrong is valid and different.

Regardless of the publication method for ones photos, they almost always need a bit of [interpretative] post-processing for the very simple reason that our memory is not very reliable. The moment we look at a scene we have an impression, get an impetus to take a picture. Unfortunately we will not find the same when we open the image sometime later on a computer monitor.

Lately I've come to post-process images with the notion of 'immediacy' - they shall evoke a similar impression as I originally had.Makes looking for ones own intentions much easier. And it leads me to the most important question, the one you, Mike, may have subsumed under the What:

Why do I want to post-process?

The formulation is a bit clumsy since I am not asking 'Why at all' but specifically 'Why do I do this now'. Answering this question is not as easy as one might think but it leads to question ones motives [no pun intended but applicable]; we think about the elements in a scene and photo, too. Eventually we decide with every interpretation we post-process a photo to, what we find interesting and the theme of the image - that should be emphasised. Bringing out the focus, so to pun ...

2:20 AM  
Blogger Player said...

Gordon sparked some thoughts with his forward thinking ideas about final output. I can imagine a gallery exhibit, with LCDs in picture frames, powered by light. These could be offered in various sizes, essentially a borderless picture browser, with interchangable frames, and drawing power from ambient light, similar to calculators. Also, with the ability to recognize and utilize high gamut color spaces. The museum of the future?

I've been mulling over these different versions of the Chicago skyline, and without seeing prints, obviously, I really like advman's version the best. It's simply more interesting and beautiful to look at, but the tradeoff might be that it doesn't make as strong a statement.

8:18 AM  
Blogger Mike Johnston said...

"Gordon sparked some thoughts with his forward thinking ideas about final output. I can imagine a gallery exhibit, with LCDs in picture frames, powered by light"

Player,
I don't know about the powered by light part, but otherwise it sounds very similar to what Bill Gates has in his home in Washington State.

--Mike

8:38 AM  
Blogger Vesa Loikas said...

Wow! That last image is special, no question the most interesting one.

That said, the 'bleak' original image is like a poem that slowly gets to you.

sincerely,
Vesa from Turku, Finland

9:48 AM  
Blogger Adam McAnaney said...

I'm going to have to pile on the advman bandwagon. But I realize its a very different interpretation from what you were going for, and I agree that I would need to see prints to know which I really like better. Maybe you should consider a double-print offer (one copy of yours, one copy of advman's), instead of going for a straight 50% off deal...that would be an interesting buy.

Best,
Adam

10:57 AM  
Blogger Mike Johnston said...

I like advman's too...as a jpg. I'm afraid the file wouldn't stand up to that level of abuse in a print.

That said, vesa lolkas summed up nicely what I'm going for with the picture...and I really think it works in the print. Not so much as an onscreen jpg.

--Mike

11:09 AM  
Blogger Peter Smith said...

This is the kind of image for which LightZone was created. Both ToneMapper and an easily masked ZoneMapper working on the foreground would do wonders.

10:08 PM  
Blogger Dan Kanagy said...

Between Mike's and Steven's interpretations, my preference is for Mike's. Here’s why. I think the strength of Mike's interpretation (MI) comes from the gap between what you see on first and second glance. It seems to me that Steven would prefer that more be revealed in the first glance, but by doing so he diminishes the surprise that comes from looking more closely and discovering how much you missed at first. It's this ambiguity that gives MI its power. So you could say that Mike's version is more patient than Steven's. MI is willing to assume the risk that the viewer will take the time necessary to see the print for what it is. Not all viewers will. But for me, it's that generosity of spirit, a willingness to assume the best of the viewer, that makes MI the great image that it is.

1:02 AM  
Blogger advman said...

Adam,

Glad you like my version. Basically it is a complete reinterpretation, guided only by what I saw in the image, uninfluenced by any real experience.

I can't be sure without having the RAW file, but I fear Mike is right. It most probably won't stand up to the expectations for the quality of a big print. It would be fun to try, though.

I have experience with only two RAW converters, Adobe Camera RAW and Nikon Capture NX. Both work well and sometimes I think extreme changes come out with less noise from Capture NX. Well, maybe Nikon knows a bit about their cameras :)

As regards LightZone, I know the concepts, the program looks interesting and I think I'll try it some time. I know, I try them all, sooner or later. But the fact is, that none of these programs can pull details out of a RAW file that are not there. I admit that LightZone may be more intuitive to some, though. On the other hand, what I did in Photoshop was not at all hard. I guess it depends on where you come from and what you've worked with.

5:19 AM  
Blogger shooter said...

Mike this is my first visit and post, I discovered the site via B&W photog and your article, I decided to download it and have a go myself, I would forward it as an attachment but obviously have no idea where to send it. Mine is a B&W tritoned version, which is in keeping methinks with your original image. I would echo what has been said re this issue though. This is no different from giving a variety of folk a negative and asking them to print it, each will hopefully produce a quality print, but there will and should be differences. This after all is the power of an image to stir the soul of the viewer and therefore provide variety in the final image. I will definitely pop back again...

5:35 PM  
Blogger roentarre said...

This one really reminds me how important post processing is all about.

5:04 PM  

Post a Comment

<< Home