Photo Forum / Digital Photography / DSLR Cameras / February 2006
Technique suggestions
|
|
Thread rating:  |
G.T. - 01 Feb 2006 20:08 GMT A photo from the first batch I've ever shot at Yosemite:
http://www.photo.net/photodb/photo?photo_id=4080579&size=lg
There's a lot of noise in the shadows from using the Shadows/Highlights tool in Photoshop. Other than a grad neutral density filter or taking two shots and combining during post-processing is there anything else I could do to make the shadows brighter without the noise?
Thanks, Greg
Paul Furman - 01 Feb 2006 20:28 GMT > A photo from the first batch I've ever shot at Yosemite: > [quoted text clipped - 4 lines] > and combining during post-processing is there anything else I could do to > make the shadows brighter without the noise? That looks pretty clean to me, maybe some noise in the full size original? Shooting RAW will let you get more out of the shadows but there are limits to how realistic it looks when masking multiple exposures or raw conversions. There are lots of pure black shadows, if shooting jpeg, turn down the contrast in the menu.
The most natural technique is to shoot in softer lighting conditions at sunrise/sunset or even at night.
G.T. - 01 Feb 2006 21:12 GMT > > A photo from the first batch I've ever shot at Yosemite: > > [quoted text clipped - 10 lines] > exposures or raw conversions. There are lots of pure black shadows, if > shooting jpeg, turn down the contrast in the menu. I shot RAW, Rebel XT, and believe me, after pushing the shadows in Photoshop there is a lot of noise in the trees and water on the right. I can put up the RAW file which is a little underexposed in the first place if anyone wanted to look at it.
Greg
Paul Furman - 01 Feb 2006 21:22 GMT > "Paul Furman" <paul-@-edgehill.net> wrote in message > [quoted text clipped - 19 lines] > the RAW file which is a little underexposed in the first place if anyone > wanted to look at it. Sure, I'll take a look.
Scratch and Sniff - 01 Feb 2006 21:35 GMT : > > A photo from the first batch I've ever shot at Yosemite: : > > [quoted text clipped - 20 lines] : : Greg Noise ninja can help.
Jeremy Nixon - 01 Feb 2006 23:53 GMT > I shot RAW, Rebel XT, and believe me, after pushing the shadows in Photoshop > there is a lot of noise in the trees and water on the right. I can put up > the RAW file which is a little underexposed in the first place if anyone > wanted to look at it. Underexposure (even a little) will drastically increase shadow noise, as will pushing the shadows in Photoshop. That's the nature of the thing, given that the shadow areas are (by definition) already "underexposed" to begin with.
 Signature Jeremy | jeremy@exit109.com
G.T. - 02 Feb 2006 02:17 GMT > > I shot RAW, Rebel XT, and believe me, after pushing the shadows in Photoshop > > there is a lot of noise in the trees and water on the right. I can put up [quoted text clipped - 5 lines] > given that the shadow areas are (by definition) already "underexposed" to > begin with. Understood. But other than the two suggestions I've had so far of either using a graduated filter or taking two shots with different exposures and combining them is there anything else to do for shots like these?
Thanks, Greg
Roger N. Clark (change username to rnclark) - 02 Feb 2006 02:44 GMT > Understood. But other than the two suggestions I've had so far of either > using a graduated filter or taking two shots with different exposures and > combining them is there anything else to do for shots like these? What raw converter did you use? There is a surprising difference. See:
Digital Camera Raw Converter Shadow Detail and Image Editor Limitations: Factors in Getting Shadow Detail in Images http://www.clarkvision.com/imagedetail/raw.converter.shadow.detail
Roger
G.T. - 02 Feb 2006 03:11 GMT > > Understood. But other than the two suggestions I've had so far of either > > using a graduated filter or taking two shots with different exposures and [quoted text clipped - 6 lines] > Editor Limitations: Factors in Getting Shadow Detail in Images > http://www.clarkvision.com/imagedetail/raw.converter.shadow.detail Wow. I was using Photoshop CS and ACR 2.4.
Greg
JPS@no.komm - 02 Feb 2006 03:17 GMT >Wow. I was using Photoshop CS and ACR 2.4. Where was the "Shadows" slider? The slider clips shadows above about 3 or 4, and down at about 0, 1, and 2, it boosts the shadows. Your image is highly posterized in the shadows, but that could have happened at multiple points in the conversion/PP process.
 Signature
<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< G.T. - 02 Feb 2006 07:12 GMT >>Wow. I was using Photoshop CS and ACR 2.4. > > Where was the "Shadows" slider? The shadows slider was left at 5. Maybe I'll try again at 1 or 2.
> The slider clips shadows above about 3 > or 4, and down at about 0, 1, and 2, it boosts the shadows. Your image > is highly posterized in the shadows, but that could have happened at > multiple points in the conversion/PP process. A lot of that posterization came in PS when I used the Shadows/Highlights function. I'll try to remember to put up the RAW file in the morning in case you or Paul F want to look at it.
As to Jeremy's last suggestion how much higher could I have exposed without blowing out details on the light colored granite? I guess I should be pushing my exposure more.
Greg
 Signature "All my time I spent in heaven Revelries of dance and wine Waking to the sound of laughter Up I'd rise and kiss the sky" - The Mekons
Paul Furman - 02 Feb 2006 18:13 GMT >>> Wow. I was using Photoshop CS and ACR 2.4. >> [quoted text clipped - 6 lines] >> is highly posterized in the shadows, but that could have happened at >> multiple points in the conversion/PP process. If I'm going to do heavy processing, I tend to do the raw conversion with as little contrast as possible, save to 16 bit and add the contrast with curves.
> A lot of that posterization came in PS when I used the > Shadows/Highlights function. I'll try to remember to put up the RAW > file in the morning in case you or Paul F want to look at it. Hold down the alt key while working the exposure & shadows sliders in ACR to see where clipping occurs. The brightness slider is less destructive than boosting exposure in my experience but the recommendation is to start at the top controls first & work your way down.
> As to Jeremy's last suggestion how much higher could I have exposed > without blowing out details on the light colored granite? I guess I > should be pushing my exposure more. JPS@no.komm - 02 Feb 2006 23:31 GMT >As to Jeremy's last suggestion how much higher could I have exposed >without blowing out details on the light colored granite? I guess I >should be pushing my exposure more. That's what the 12 bits are there for. You'd be surprised how little of the RAW data's dynamic range is actually used in many "0 EC" images.
The problem is, the cameras themselves are totally dumb regarding RAW exposure, so your pretty much on your own figuring out where the clipping level is in the RAW data (with imprecise hints from the histogram). It would probably take only a few hours of programming to put RAW RGB histograms onto the LCD review, but the manufacturers couldn't care less, because no one is holding back their credit cards in protest.
 Signature
<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< Jeremy Nixon - 03 Feb 2006 00:35 GMT > The problem is, the cameras themselves are totally dumb regarding RAW > exposure, so your pretty much on your own figuring out where the > clipping level is in the RAW data (with imprecise hints from the > histogram). Yes, experimentation is the only way to do it.
> It would probably take only a few hours of programming to put RAW RGB > histograms onto the LCD review, It would take more, but I don't think it would be prohibitive, to add a metering mode for "expose as high as you can without clipping", perhaps allowing for spectral highlights or whatever. I'd rather that as a starting point for EC than the current state of the art, even though you can learn to deal with current methods well enough through experience.
> but the manufacturers couldn't care less, because no one is holding > back their credit cards in protest. The problem is that holding back your credit card in protest over that means you don't have a camera at all, which is further than most of us are willing to go.
 Signature Jeremy | jeremy@exit109.com
JPS@no.komm - 03 Feb 2006 01:37 GMT >It would take more, but I don't think it would be prohibitive, to add a >metering mode for "expose as high as you can without clipping", perhaps >allowing for spectral highlights or whatever. I'd rather that as a >starting point for EC than the current state of the art, even though >you can learn to deal with current methods well enough through experience. My idea is an "escape" control, rather than EC, for something like that. It would be an exponential scale of the amount of highlights that you are willing to allow to be clipped:
* * * * * * * * * * * * * * * *
| | | | | | 0.001 0.01 0.1 1 10 100
Of course, this requires a hi-res metering sensor.
 Signature
<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< C J Southern - 03 Feb 2006 03:56 GMT Out of interest, I'm wondering if it would work if you could set a 'clipping tolerance' (per John's suggestion) -- then the camera takes 2 exposures - uses the first to work things out (and then discard the image after calculating the result) - then take a 2nd based on the calculations for the first.
Would that work?
Paul Furman - 03 Feb 2006 04:26 GMT > Out of interest, I'm wondering if it would work if you could set a 'clipping > tolerance' (per John's suggestion) -- then the camera takes 2 exposures - [quoted text clipped - 3 lines] > > Would that work? It seems to me they already have something like this in the metering only it's set too sensitive and meters the RGB average, ignoring the potential in the raw file. If I shoot aperture priority at night with bright lights (or bright reflections in the day) the camera will underexpose to preserve those highlights. Or I guess the biggest difference is that when those highlights are absent, the meter won't push the exposure up but will seek middle gray. But the basic algorithm seems to be in place.
JPS@no.komm - 03 Feb 2006 21:31 GMT >Out of interest, I'm wondering if it would work if you could set a 'clipping >tolerance' (per John's suggestion) -- then the camera takes 2 exposures - >uses the first to work things out (and then discard the image after >calculating the result) - then take a 2nd based on the calculations for the >first.
>Would that work? Sure. All you need is a RAW histogram from the first exposure to compensate the second one. It wouldn't be useful for timing-critical shooting, of course.
 Signature
<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< Paul Furman - 03 Feb 2006 00:37 GMT >>As to Jeremy's last suggestion how much higher could I have exposed >>without blowing out details on the light colored granite? I guess I [quoted text clipped - 7 lines] > clipping level is in the RAW data (with imprecise hints from the > histogram). "Just a little bit of blinking in the highlights." -but you can blow out the blue in a sky without any blinking because it's pure blue and that doesn't count as much in the histogram and blinker thing.
> It would probably take only a few hours of programming to > put RAW RGB histograms onto the LCD review, but the manufacturers > couldn't care less, because no one is holding back their credit cards in > protest. JPS@no.komm - 03 Feb 2006 01:57 GMT >>>As to Jeremy's last suggestion how much higher could I have exposed >>>without blowing out details on the light colored granite? I guess I [quoted text clipped - 12 lines] >it's pure blue and that doesn't count as much in the histogram and >blinker thing. The only time I see blue skies blow out in the RAW highlights is when the blue sky is only a small portion of the image, not influencing the metering very much, and is the brightest thing (this implies shaded subject matter). It is more often the hazy, white sky near the horizon that blows out (*really* blue sky isn't really all that bright) Blue sky is not as saturated, or as blue in the RAW data as it is in the JPEG output. Conversion with normal saturation and color tone settings can boost the red levels in saturated red subjects and blue in blue subjects by a couple of stops, and clip well below the RAW clipping point. Fact is, it is usually the green channel that blows out first in blue skies that are over-exposed.
It is *very* easy to be tricked into under-exposure by clipped red and blue in the converted output. One image I was looking at a few months ago had a red flower blowing out in the JPEG (with -2 contrast setting, no less!), and it was a at least 2 stops away from clipping in the red channel of the RAW data. Had someone looked at the JPEG, and took the shot again with -1 EC, peak red would be 3.5 stops from clipping in the RAW data! One way to better handle this would be to expose at +2 EC, instead, bring it down in the converter, and then take the red channel of the RAW and use it in a layer to increase the contrast of luminance in the petals (as most converters are not very good at some types of red flowers). The individual RAW color channels will often have high greyscale contrast where the color conversion is totally lacking in contrast. This is why color -> greyscale conversions are usually so dull; luminance contrast is only strong with unsaturated colors in converted RGB color.
 Signature
<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< Paul Furman - 03 Feb 2006 04:32 GMT >>>>As to Jeremy's last suggestion how much higher could I have exposed >>>>without blowing out details on the light colored granite? I guess I [quoted text clipped - 18 lines] > subject matter). It is more often the hazy, white sky near the horizon > that blows out OK, yes, thanks for pointing out that distinction. It's when trying to eek out some blue from a hazy sky that you have to be careful. When it's right on the edge of being white. I'll often see the blue evaporate leaving a blotchy aqua on white.
> (*really* blue sky isn't really all that bright) Blue > sky is not as saturated, or as blue in the RAW data as it is in the JPEG [quoted text clipped - 19 lines] > dull; luminance contrast is only strong with unsaturated colors in > converted RGB color.
 Signature Paul Furman http://www.edgehill.net/1 Bay Natives http://www.baynatives.com
G.T. - 03 Feb 2006 04:57 GMT If JPS or Paul or anyone else that does DNG files is interested in looking at, and possibly processing, my file it's at:
http://homepage.mac.com/getosx/yosemite.dng
6.6 MB
Thanks, Greg
 Signature "All my time I spent in heaven Revelries of dance and wine Waking to the sound of laughter Up I'd rise and kiss the sky" - The Mekons
Jeremy Nixon - 03 Feb 2006 05:52 GMT > If JPS or Paul or anyone else that does DNG files is interested in > looking at, and possibly processing, my file it's at: > > http://homepage.mac.com/getosx/yosemite.dng I just loaded it up. My observations:
You've got an extreme range from light to dark, but the exposure is perfect. You couldn't have exposed much higher without blowing some detail.
There is chroma noise in the deep shadow areas in the left bottom corner and on the right side in the trees and mountain; I assume that's what you were talking about initially. I don't think there was any way to avoid it in a single exposure given how dark those areas are -- it's just the noise characteristic of the camera. However, taking "color noise reduction" up to 5 in Adobe Camera Raw eliminates it with no negative effect on the picture. Non-issue.
There is chromatic aberration in the high-contrast edges, in particular on the right side near the sky and on the lower left in the sunlit snow. The red/cyan CA slider in ACR takes care of it; I swooped it over to about -40 and that looks good, at least as a start.
There is posterization in the shadows, noticeable in for example the rocks on the lower left. That's the nature of digital shadows, and could only be avoided with more exposure, which you couldn't really do in a single exposure without a graduated ND filter (which would have been tricky given the composition). Just let it go; while it was slightly noticeable at screen resolution, I tweaked the curve a tiny bit on the shadow end to darken the shadows a bit and you'd never notice. You can still see the artifacts of it at 100%, but really, no big deal.
Don't bother trying to boost the shadows. It will bring out the problems and I don't think it's necessary -- it reduces the overall visual contrast and (in my opinion, we're getting into aesthetics here so take it for what it's worth) detracts from the picture. In fact, I think darkening the shadows a bit as I described above improves the impact.
To get really good detail in those shadows, even with a "better" camera, would require more exposure, possibly using multiple exposures and the Merge to HDR function in Photoshop. Yes, a "better" camera would have had less chroma noise there, but the noise you have is really just not anything to worry about.
In short -- I think there is nothing technically wrong with this picture that isn't quickly solved with simple processing. In addition to what I've described, if it were mine, I'd probably take it into Photoshop proper and try to increase local contrast in the sunlit mountain, but all in all, I think you might be falling into the pixel-peeing trap.
If you shot this (and similar pictures) on a tripod, consider using auto- bracketing to take a series of exposures, starting with where you were at and then overexposing by a couple of stops, and playing with Merge to HDR in Photoshop. That takes a bit of learning, but if you want better detail in those shadows that's how you'll get it.
 Signature Jeremy | jeremy@exit109.com
G.T. - 03 Feb 2006 06:18 GMT >>If JPS or Paul or anyone else that does DNG files is interested in >>looking at, and possibly processing, my file it's at: [quoted text clipped - 5 lines] > You've got an extreme range from light to dark, but the exposure is perfect. > You couldn't have exposed much higher without blowing some detail. That was the biggest thing I was wondering.
> There is chroma noise in the deep shadow areas in the left bottom corner > and on the right side in the trees and mountain; I assume that's what you [quoted text clipped - 8 lines] > red/cyan CA slider in ACR takes care of it; I swooped it over to about -40 > and that looks good, at least as a start. I'm going to run through it again with this advice.
> There is posterization in the shadows, noticeable in for example the rocks > on the lower left. That's the nature of digital shadows, and could only [quoted text clipped - 4 lines] > darken the shadows a bit and you'd never notice. You can still see the > artifacts of it at 100%, but really, no big deal. I was wondering about the graduated filter on such a hilly subject. And I need to learn more about curves. I barely have an idea of how they work.
> Don't bother trying to boost the shadows. It will bring out the problems > and I don't think it's necessary -- it reduces the overall visual contrast [quoted text clipped - 13 lines] > try to increase local contrast in the sunlit mountain, but all in all, I > think you might be falling into the pixel-peeing trap. My Mom has done just that along with eliminating the sticks and snow at the bottom. She is good at the aesthetic side of the things but is still learning the tech side of digital photography such as noise and sharpening. My Dad got her one of the Pentax *ist models for Christmas.
> If you shot this (and similar pictures) on a tripod, consider using auto- > bracketing to take a series of exposures, starting with where you were at > and then overexposing by a couple of stops, and playing with Merge to HDR > in Photoshop. That takes a bit of learning, but if you want better detail > in those shadows that's how you'll get it. It was so bright out that day that I didn't bother bringing my tripod. In hindsight it's obvious that I should always bring it.
Mucho gracias. I've learned alot from you, Paul, and JPS.
Thanks, Greg
 Signature "All my time I spent in heaven Revelries of dance and wine Waking to the sound of laughter Up I'd rise and kiss the sky" - The Mekons
Paul Furman - 03 Feb 2006 16:24 GMT Here's my attempt at bringing out shadows, if that's the desired look the noise probably won't show in a print though it does look bad while pixel peeping. Your first version is the first in the series and the second is my default ACR settings. http://www.edgehill.net/1/?SC=go.php&DIR=Misc/temp&PG=1&PIC=3 -scroll down for a mini-psd file that shows masking with a broad brush eraser on 2 versions of the raw conversion & a curve. I fiddled around so can't recall what I ended up doing exactly.
(oh, I forgot to change to sRGB from Adobe RGB)
>>> If JPS or Paul or anyone else that does DNG files is interested in >>> looking at, and possibly processing, my file it's at: [quoted text clipped - 84 lines] > Thanks, > Greg
 Signature Paul Furman http://www.edgehill.net/1 Bay Natives http://www.baynatives.com
G.T. - 03 Feb 2006 17:26 GMT > Here's my attempt at bringing out shadows, if that's the desired look > the noise probably won't show in a print though it does look bad while [quoted text clipped - 4 lines] > eraser on 2 versions of the raw conversion & a curve. I fiddled around > so can't recall what I ended up doing exactly. Cool. In my attempt last night I dropped the Shadows to 1 in ACR and dropped the contrast, I think, to 0. Just doing that and boosting contrast in PS CS helped tremendously.
Thanks, Greg
 Signature "All my time I spent in heaven Revelries of dance and wine Waking to the sound of laughter Up I'd rise and kiss the sky" - The Mekons
Paul Furman - 03 Feb 2006 17:55 GMT >> Here's my attempt at bringing out shadows, if that's the desired look >> the noise probably won't show in a print though it does look bad while [quoted text clipped - 8 lines] > dropped the contrast, I think, to 0. Just doing that and boosting > contrast in PS CS helped tremendously. To use curves, wave your cursor over the area you want more contrast noting where it is located on the curve and make the slope steeper in that area. Add more control points to fine tune the shadow & highlight areas if needed. Usually a s shape is the answer for increasing contrast of middle tones.
JPS@no.komm - 04 Feb 2006 23:35 GMT >If JPS or Paul or anyone else that does DNG files is interested in >looking at, and possibly processing, my file it's at: > >http://homepage.mac.com/getosx/yosemite.dng That scene, as other have already said, has tremendous dynamic range. Your mountain-top snow in the sun is slightly clipped in the RAW data (most of it is clipped in the green channel, less in the blue, and even less in the red, as you'd expect with snow in sunlight). Besides the snowcaps, the next brightest thing in the RAW file is over a stop darker than it, so, even though you have achieved an "exposure to the right" with very little clipping, the only thing you would have lost by exposing a stop higher would be a little detail at the darker edges of the sunlit snow. Since the snow is such a smaller feature of the overall image, I would have let it go clipped, for the shadows' sake, if I wanted the shadows.
 Signature <>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<> John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>>< Jeremy Nixon - 02 Feb 2006 03:44 GMT > Understood. But other than the two suggestions I've had so far of either > using a graduated filter or taking two shots with different exposures and > combining them is there anything else to do for shots like these? Well, the other approach would be not to underexpose. I'm not familiar with your camera to know how much it can take in the highlights, but you can often push them further than you think you can -- though it depends on the camera. With digital capture the best exposure is the highest exposure you can get without losing the highlights you want, even if that might look grossly overexposed on the LCD when you do it (for a low-contrast scene in particular). The bottom stop or two of shadows is the weakest point in current digital cameras.
 Signature Jeremy | jeremy@exit109.com
John A. Stovall - 01 Feb 2006 23:44 GMT >A photo from the first batch I've ever shot at Yosemite: > [quoted text clipped - 4 lines] >and combining during post-processing is there anything else I could do to >make the shadows brighter without the noise? Noise Ninja.
http://www.picturecode.com/
*********************************************************
"It looked like the sort of book described in library catalogues as "slightly foxed", although it would be more honest to admit that it looked as though it had been badgered, wolved and possibly beared as well."
_Light Fantastic_ Terry Pratchett
zeitgeist - 04 Feb 2006 04:49 GMT > A photo from the first batch I've ever shot at Yosemite: > [quoted text clipped - 4 lines] > and combining during post-processing is there anything else I could do to > make the shadows brighter without the noise? Full sun and full shade is 4 to 5 stops variance. Consider what that would have looked like if shot with film.
A. use a tripod, shoot one exposed for the sun, one exposed for the shade.
B. take your raw file and slide the exposure over to get best of the hightlight, then make a second file and slide the exposure over for the shadow detail. Now mask the areas to reveal the best of each.
C. get a graduated ND filter to cut the exposure on the top half.
D. shoot in the dawn and dusk hours when the lighting is more even, smooth.
E. don't shoot in such extreme situations.
|
|
|