Please play with this picture

[Billy]

I'm looking to get a decent photo editing tool (Photoshop Elements maybe?). Looking for info on what people are using....pros and cons of what you're using....and please feel free to mess with the photo below......show what can be done with photo editing. Thanks.

edit: gotta learn how to attach a pic I guess...I used

around the url of the photo.
Until I learn how to attach a pic...here's the link http://walkinginthewoods.smugmug.com/gallery/7056019_poZDA#452041670_7NKcD

 

[grouseking]

Welcome to Views!

I don't have photoshop, though I know people love using it. I occasionally use a simple photo editing tool, called Windows Live Photo Gallery. You can dl it for free.

I downloaded your photo, and simply increased the brightness a little...I didn't want to do much. So here is what it looks like...originally



And here it is touched up a bit.



I put it in my webshots misc photo album. There is prob more you could do with this pic...maybe crop it a bit to eliminate some things and bring the mountains a little closer to you. But then again, this is much a gorgeous view, I don't know if it matters. Hope that helped! I'll let the pros talk now.

grouseking

 

[grouseking]

Ok, I couldn't help it, I played a little more with it, kind of cropped out a little of the rocks and part of the mountain scenery, and got this. It makes Bondcliff come in a bit closer.



Others might crop out some of the sky above, since it will make the mountains look even larger, and closer up. Sometimes I like to do that, but in this pic, the sky is such a beautiful color, I decided against it.

grouseking

 

[bikehikeskifish]

Here's my version.



Using the curves, I removed the last 1/2 stop from the right which lightened it up some. Then I increased the curve slightly at about 80% from the left. Finally I cropped some off the left mostly to bring the human closer to the lower right 1/3 intersection. It may be a bit too light, I wanted to get some of the texture on Bondcliff in the shadow area.

I need to get back to the Bonds...

Tim

 

[Tom Rankin]

I need to get back to the Bonds...

Come and join us! (Yes, sorry, thread jack!)

 

[Grumpy]

This is one of those high dynamic range scenes in which it is important to retain and exhibit detail in both the highlight areas (sky, illuminated foreground rocks, etc.) and shadows (especially in that large shaded area below the prominent ridge in middle distance). So it is first and foremost an exposure problem.

I think this image probably is underexposed, perhaps by two f/stops or more. Unfortunately, boosting the exposure to improve shadow detail would tend to make the important highlights block up and lose texture, so just boosting the exposure may be self-defeating. One way to cope with this is to (a) use a “low image contrast” setting, if your camera setup menu has such a thing, and (b) dial substantial “plus” into the “exposure compensation” control if your camera allows for that. Using a spot meter option, if your camera has it, to meter off a darker area in which you wish to bring out detail also may help.

What this does is help you replicate the old basic rule with black and white film: Expose for the deepest shadow in which you wish to retain image detail, and develop to preserve detail in the highlights. Essentially, it is a mandate to overexpose and underdevelop the film to compress the tonal scale (reduce image contrast).

In digital, it will help considerably to shoot in RAW format (if your photo processing software and camera system permits), which allows for considerable control in the electronic version of film developing. (Always save the unmanipulated original image, though, in case you want to do over.)

If you don’t have all these controls available, the result of Tim’s manipulations, shown above, is a solidly improved version of your original photo.

BTW, I rather like the photo itself. Nice composition and sense of third dimension (depth). In my view, the human figure adds scale and meaning.

G.

 

[DougPaul]

What this does is help you replicate the old basic rule with black and white film: Expose for the deepest shadow in which you wish to retain image detail, and develop to preserve detail in the highlights. Essentially, it is a mandate to overexpose and underdevelop the film to compress the tonal scale (reduce image contrast).

This works for film in part because it has a soft saturation. (ie a gradual transition between the "normal" and saturated exposures.)

BTW, the night scenes in movies used to be shot using the opposite of the above technique: shoot underexposed in daylight and overdevelop because it increased the contrast and made the shadowed zones disappear.

In digital, it will help considerably to shoot in RAW format (if your photo processing software and camera system permits), which allows for considerable control in the electronic version of film developing. (Always save the unmanipulated original image, though, in case you want to do over.)

Digital, unlike film, has a hard saturation--there is no transition zone between the normal and saturated exposures. (Each color saturates individually so first the color will change and then the saturated zone will become full white with absolutely no detail.) So saturation of any part of a digital image can be fatal to the image. (Exception: saturation of a bit of something that is already bright white can be ok, such as clouds.) The general advice that I have seen is that one should set digital exposures to just below saturation. Much of the shadow detail will still be there (but may be noisy) and can be brightened in post-processing, particularly from a RAW file. (RAW files can have more dynamic range than JPEGs.)

In terms of exposure histograms*, this means moving the histogram as far to the right as possible without saturating (which shows as a spike on the right edge of the histogram). Some cameras (and post-processing software) can be set to indicate saturated pixels by flashing them on and off.

* Color histograms are much more useful than "intensity" histograms. (Intensity histograms are often just a green histogram and thus you can saturate the red or blue without it showing on an intensity histogram.)

Doug

 

[Billy]

...Using the curves, I removed the last 1/2 stop from the right which lightened it up some. Then I increased the curve slightly at about 80% from the left...

Tim

Thanks all for the touch-ups and technical info.

Tim, you say curves, I think Beyonce'. What is curves? I assume that's part of your photo editing s/w? What are you using? Like it / hate it? Thanks.

 

[bikehikeskifish]

Curves are a way to manipulate the image's histogram. You can do the same with brightness and/or contrast controls for the most part. I use GIMP, mainly 'cuz it is free and does most of the same things that Photoshop does. Plus it runs on Linux if that kind of thing is important to you.

Tim

 

[kmorgan]

my edit...

In Photoshop:

Used Highlights/Shadows adjustment set to 50% highlights. Converted to LAB mode and steepened the A and B channel 10%, then adjusted the Lightness channel curve for contrast. Converted back to RGB and this is what I got:

I couldn't download the original file from Smugmug so I just copied it from Grouseking's post.

Kevin

 

[DougPaul]

Curves are a way to manipulate the image's histogram. You can do the same with brightness and/or contrast controls for the most part. I use GIMP, mainly 'cuz it is free and does most of the same things that Photoshop does. Plus it runs on Linux if that kind of thing is important to you.

The Gimp, unfortunately, is limited to 8-bit processing. Serious photo processing requires at least 16 bits and the ability to import RAW images (which are generally >8-bit). Cinepaint is a fork which includes >=16-bit processing, but it isn't ready for general use yet.

Doug

 

[NeoAkela]

Nice shot and good retouches.

This is one of those light situations that would work well with a tripod, three exposures (around -2, 0, +2), and a HDR program like Photomatix or something similar to put all the correctly exposed areas of the three versions together into one shot. This has worked well for me on a few occasions, such as this past winter on top of Mt Clay when the sun and shadows caused some extreme contrasts. However, it is easy to get a bit carried away with these things, especially with HDR, and create an unrealistic looking photo instead of staying true to what the eye sees.

The first touchup by Grouseking is my favorite - just brings out a bit of the color from the shadows.

 

[bikehikeskifish]

The Gimp, unfortunately, is limited to 8-bit processing. Serious photo processing requires at least 16 bits and the ability to import RAW images (which are generally >8-bit). Cinepaint is a fork which includes >=16-bit processing, but it isn't ready for general use yet.

Doug

GIMP itself is limited to 8 bits - but GIMP has a 16-bit plug-in for Adobe DNG work (ufraw, it runs standalone as well), which I use for my CHDK-augmented Canon A570IS. Since most P&Ses only record 8-bit JPEGs anyway, GIMP is fine for most purposes. And 8 bits is plenty for photos being shared on web sites.

Tim

 

[DougPaul]

GIMP itself is limited to 8 bits - but GIMP has a 16-bit plug-in for Adobe DNG work (ufraw, it runs standalone as well), which I use for my CHDK-augmented Canon A570IS. Since most P&Ses only record 8-bit JPEGs anyway, GIMP is fine for most purposes. And 8 bits is plenty for photos being shared on web sites.

8 bits is fine for driving a computer screen or printer. JPEGs are ok here. But transparencies (positive or negative) can require more bits to match their capabilities.

8 bits is not adequate for many originals (RAWs). Any camera generated JPEG has thrown away information that may have been useful in any additional processing.

8 bits is not adequate for processing (even for 8-bit images). Processing can require greater dynamic range or can be sensitive to quantization noise at intermediate stages.

Professional quality processing typically starts with a RAW (>=12 bits) image and will do all processing in >=16 bits until a final conversion to 8 bits (eg a JPEG) for printing or screen display. (Camera generated JPEGs can also keep the processing at >8 bits inside the camera until final output of the JPEG if the designers so choose.)


Using ufraw (I also have it here) to convert RAW to an 8-bit format (such as JPEG) for input to Gimp is a half-way measure. Ufraw allows one to do white balancing, some forms of gray-scale manipulation, and some other fairly simple manipulations. If these manipulations are adequate, then there is no problem converting to JPEG. But if you need to feed the output to an 8-bit processor (such as GIMP) then you are performing the latter part of your processing at 8 bits which may be inadequate. (This still may be better than starting with an 8-bit RAW or the camera JPEG.)

The amount of damage from using 8-bit processing depends upon the image, the processing, the internal representations, the output device, and how critical the observer is.

Doug

 

[DougPaul]

Cinepaint

In an earlier post, I commented that cinepaint wasn't ready for general use. I just tried fooling around with version 0.22-1* and it looks like it might be in good enough shape to be useful. (For raw images, use ufraw or dcraw to generate a 16-bit TIFF file to feed to cinepaint.) Not sure it will work on MS, but it will work on Linux and Macs. Its capabilities are intended to be similar to GIMP with >8 bit processing and movie capability. (It has been used on a number of commercial movies.)

http://www.cinepaint.org/
http://ufraw.sourceforge.net/
http://www.cybercom.net/~dcoffin/dcraw/
http://www.gimp.org/
(All are free and opensource.)

* version 0.25 is available from CVS, but I couldn't get it to compile.

Doug

 

[Grumpy]

My guess is that the version of this photo produced by Kevin in post #10 comes closest to replicating what Billy (the photographer) saw in the scene, with decent detail in both the deeper shadows and brighter highlights. Our brains are marvelous information processors in this respect. The challenge is to reproduce the brain-processed image using various “capture” media.

G.

 

[bikehikeskifish]

GIMP has been fine for me and my hiking photos for the last year+. I generally capture as 8-bit JPEGs in the camera and host on Webshots at a greatly-reduced resolution anyway. For particularly difficult exposures, I'm experimenting with CHDK and raw capture using DNG format. I generally use GIMP for cropping and fixing the exposure. Clearly I am not a professional user. Since Billy thought of Beyonce when I said "curves", it is probably an acceptable starting point for him given it is free.

I'm curious to see how much better the 16-bit conversion process works on a low-resolution web photo such as Billy's. Care to submit your version, Doug?

Tim

 

[DougPaul]

Clearly I am not a professional user. Since Billy thought of Beyonce when I said "curves", it is probably an acceptable starting point for him given it is free.

It is quite possible that he knew the concept, but not the name.

I'm curious to see how much better the 16-bit conversion process works on a low-resolution web photo such as Billy's. Care to submit your version, Doug?

I don't usually spend much time fooling around with other people's images, but just this once for you...

I was able to download the full size original http://walkinginthewoods.smugmug.com/photos/452041670_7NKcD-O.jpg and the grouseking's small "before"* http://inlinethumb16.webshots.com/24655/2453352120049158222S600x600Q85.jpg. Both show spikes near the right edge of the histogram as if there was clipping (saturation) followed by some processing (perhaps the creation of the JPEG) which brought the saturated pixels to a lower value. (The brightest part of the image is the sky in the upper left corner--much brighter than anything in the foreground.)

* Grouseking's "before" image is only 48KB (1.6 bits/pixel)--thus it is likely to be damaged by the JPEG conversion. (The original is 3.1 bits/pixel--saved in Fine mode. Fine is the highest quality JPEG available from the XT.) If one expects to post-process a JPEG, one should save it at the highest quality possible (or a quality of 95 or so). Better yet, shoot RAW...

I can easily replicate grouseking's "after" with xv (http://en.wikipedia.org/wiki/Xv) by simply using a straight-line "curve" from 0,0 to ~220,255 (ie a small brightening of the entire image). One can get a similar brightening of the dark part of Bondcliff with minimal effect on the bright zones by increasing the gamma from 1.0 to 1.1. One can also get a combined effect by pushing the one-quarter to mid-point up a little (ie a 2 straight-line segement "curve"). I can also do essentially the same thing with GIMP, CinePaint in 8 or 16-bit mode, and LightZone (a 16-bit commercial image processor aimed at the professional/semi-pro market http://www.lightcrafts.com/).

It is clear from the sensitivity of the image to the gamma setting in xv that the appearance of this image is highly sensitive to one's display and its settings. (This is probably true for most high dynamic range images.) I suspect that variation in our displays (most of which are likely uncalbrated) plays a significant factor in our individual opinions and preferences of the images.

After fooling with the image for a bit, I don't think it is a very good one for showing the limitations of 8-bit processing. My website space is very limited, so I didn't bother to save or post my resulting images.

The original image was shot with a Canon XT. If Billy saved a RAW version of the image, it might be more interesting to play with. (But as I stated above, there are probably better images for exploring this issue.)

Doug

 

[bikehikeskifish]

I will mention, at the risk of being corrected, that one of the reasons I installed CHDK on my Canon A570IS was because it has live histograms (scroll down about 1/4-1/3 of the page), including the RGB model (among others), which helps a lot with getting the exposure correct before capturing the image (as opposed to after which you can look at the luminance histogram only and decide if you want / need to retake it.) The other good reason is raw support.

Tim
p.s. Doug I've got a ton of online storage so feel free to e-mail me your converted image(s) and I will host them for this thread.

 

[DougPaul]

I will mention, at the risk of being corrected, that one of the reasons I installed CHDK on my Canon A570IS was because it has live histograms (scroll down about 1/4-1/3 of the page), including the RGB model (among others), which helps a lot with getting the exposure correct before capturing the image (as opposed to after which you can look at the luminance histogram only and decide if you want / need to retake it.) The other good reason is raw support.

Do you know how many bits you get in the raw files? (Most DSLRs give 12-14 bits/channel.)

p.s. Doug I've got a ton of online storage so feel free to e-mail me your converted image(s) and I will host them for this thread.

Thanks--I didn't come up with anything noteworthy or bother to save them. I simply observed that I could replicate what grouseking had done with a variety of tools and I could see no reason to prefer my versions over his. I could have come up with something slightly different, but as far as I was concerned, both grouseking's and your processings were fine. As stated in my previous post, I don't think this is a very good image for comparing the inner working of the processing programs.

FWIW, the original image (or at least a full-screen display of it) is a lot nicer than the small one. A nice picture and an even nicer image with just a light touch of brightening and/or gamma increase (at least on my monitor).

Doug