jfrancho
Registered user
Global user
Registered: 01-2006
Location: Rochester, NY USA
Posts: 373
|
|
Reply | Quote
|
|
Making a Picture: Part 2 Convert, De-Noise, Sharpen, Resize, Convert to B&W
I left off with getting the image in the camera exactly the way I wanted to capture it. If you stumbled on to this thread, PART I can be found here. Now it is time to download the images from the card and see what we have in Photoshop CS2.
Generally speaking, I prefer Adobe Camera Raw (ACR), though not because I think it's better, but by sticking to Photoshop, I find the work flow is much more streamlined than using a "best in breed" approach, switching applications at certain portions of the image processing work flow.
I opened the image in ACR, and made my adjustments. I didn't really need to do too much to the image, since I worked so hard to set up the shot. In fact, with some custom parameters, I could have easily gotten away with shooting in JPEG mode. Before everyone starts howling about why raw is better, and I'm tossing out 12 bits of crucial color depth, know that this type of image is never going to be printed greater than maybe 3-4" or 600 px. on the web. To see the results of a JPEG work flow on similar shots, you can see the original bass bait thread here.
Below is a code snippet of the XMP sidecar that ACR uses to store the raw converter settings. This is another reason I like ACR, the settings are stored in an easy to understand, XML format that allows some opportunity for customization. That's for a different post though.
<crs:RawFileName>20070112-bassLure-07.CR2</crs:RawFileName>
<crs:WhiteBalance>Auto</crs:WhiteBalance>
<crs:Temperature>6100</crs:Temperature>
<crs:Tint>+5</crs:Tint>
<crs:Exposure>0.00</crs:Exposure>
<crs:Shadows>0</crs:Shadows>
<crs:Brightness>70</crs:Brightness>
<crs:Contrast>+30</crs:Contrast>
<crs:Saturation>+5</crs:Saturation>
<crs:Sharpness>0</crs:Sharpness>
<crs:LuminanceSmoothing>10</crs:LuminanceSmoothing>
<crs:ColorNoiseReduction>3</crs:ColorNoiseReduction>
<crs:ChromaticAberrationR>+5</crs:ChromaticAberrationR>
<crs:ChromaticAberrationB>0</crs:ChromaticAberrationB>
<crs:VignetteAmount>0</crs:VignetteAmount>
<crs:ShadowTint>0</crs:ShadowTint>
<crs:RedHue>0</crs:RedHue>
<crs:RedSaturation>+10</crs:RedSaturation>
<crs:GreenHue>+1</crs:GreenHue>
<crs:GreenSaturation>+10</crs:GreenSaturation>
<crs:BlueHue>0</crs:BlueHue>
<crs:BlueSaturation>0</crs:BlueSaturation>
<crs:ToneCurveName>Custom</crs:ToneCurveName>
<crs:ToneCurve>
<rdf:Seq>
<rdf:li>0, 0</rdf:li>
<rdf:li>64, 50</rdf:li>
<rdf:li>128, 128</rdf:li>
<rdf:li>192, 202</rdf:li>
<rdf:li>255, 255</rdf:li>
</rdf:Seq>
</crs:ToneCurve>
<crs:CameraProfile>ACR 2.4</crs:CameraProfile>
<crs:HasSettings>True</crs:HasSettings>
<crs:HasCrop>False</crs:HasCrop>
</rdf escription>
You should be able to translate that code pretty easily.
From there, I save the image as a PSD file, using 16 bpc and Adobe RGB color. Now I have the image open in Photoshop proper, and the first thing I want to do is eliminate noise. I don't want to give the the image a plastic look, or lose any of the details, especially in the highlights. I'll control where the noise reduction plug-in - I use Noise Ninja - works with a layer mask. I like to use the TLR Pro Mask Toolkit to build the mask for me. I fired up the script, and chose surface for the mask type, narrow for the width, shadows for the tone range, all for the color range, and checked the box for enhanced mask. I only wanted to reduce the noise in the shadows, but I also wanted to be sure I protected any detailed edges that may occur in that tone range. The goal of this image is to show as much detail as possible, without giving it that crispy-crunchy, over sharpened look. I want realism. Once the script runs, it looks like nothing happened. But is you go to the Channels panel, you'll see there is a new Alpha Channel named "Shadows Enhanced Surface Mask." I clicked on the channel to inspect the mask, and make sure it was protecting the areas I was concerned about. Sometimes the mask needs to manually edited. Remember black conceals and white reveals, so the mask will protect all the black areas from the noise filter. Here is what the mask looks like:
Fig. 1
I went back to the Layers panel, used Ctrl-J to duplicate the layer. Then, I Ctrl-click the new alpha channel to make the selection, and click the Add Layer Mask icon. Now I have a duplicate of the background layer with a layer mask that protects the highlights and mid tones from the noise filter. Next, I click on the layer thumbnail to so the noise filter works on the layer, and not the mask, and go to Filter > PictureCode > Noise Ninja... to bring up the NN dialog.
Fig. 2
I have my Noise Ninja preferences set to automatically profile the image for noise rather than use the downloaded ISO based profiles. It takes a little longer to run, but I prefer to have it this way. It probably doesn't make that much of a difference, but if I can add small bits of quality at each step, then surely the final output will be better. That's the idea anyway. You'll notice that I reduced the settings to 6 from the default 10 in Fig.2. Using the preview window, I decided that even with using the mask, the effect was too strong - remember those details I was trying to protect? It's worth noting that Noise Ninja will not not show you the effects of the mask in the preview window, so you are "flying blind" but after some experience, you be able to determine what settings give you the desired effect. I usually use one step higher than I like it when using the noise filter with a mask. Remember, the mask is going to protect the details, so you can be more aggressive with settings.
Now I can move on to sharpening. I won't go into a prolonged dissertation as to why I use a three phased approach to sharpening, other than I have bought in to the "sharpen early and sharpen often" methodology recommended by the late Bruce Fraser. I again will Dr. Glenn "Mitch" Mitchell's scripts for sharpening. He has a great paper on sharpening in general, and his TLR Pro Sharpening Toolkit is one of the best tools I've found. I ran the Capture Sharpening script, and chose narrow for the width, USM for the method, 2-2-1 for light, dark, and surface, and checked the box for enhanced masks. Here is the result:
Fig. 3
All this script really does, if used conservatively, is just demoirize the image. This is often necessary with raw images where no sharpening is applied in the raw converter. You can often skip this step if you add 25-30 sharpening in ACR or +1 in camera sharpening if you are using JPEG mode.
When inspecting the image at this point, I discovered that the capture sharpening script did such a good job protecting the edges that I was probably too conservative with the surface sharpening. No problem, I will be running the Creative Sharpening script next, and I can add a bit of surface sharpening here. All I want to do is bring out the texture of the lures sides and enhance the other textures in the image. The lure is made of wood, plastic metal and paint, and it is my goal to show off these texture as well as I can. Certainly lighting played a big part in this earlier, but I want to continue to enhance these features.
Fig. 4
You'll notice that I also ran the script with Anti Haze USM setting. This applies localized micro contrast to the image globally. Look up "localized contrast control with USM" to see how it works. The surface sharpening isn't actually applied to the image by the script. You have to "paint" it in using the brush tool.
Fig. 5
I used a soft, medium opacity brush to paint in where I wanted the surface sharpening. You must remember to click on the mask thumbnail, or you'll make a mess out of the image. If you don't like what you did, switch from white (reveal) to black (conceal) to reverse the effect. This is truly non destructive editing. Not only can the background layer be protected by using additional layers, but the I am also controlling where, and how much of the effect or filter down to the pixel - with 100 levels of opacity to boot! Amazing! Fig. 6 shows just the mask I "hand painted."
Fig. 6
My next step is to resize the image, using Image > Resize... This easier said than done. Most users leave the default resampling method to the default bicubic method. There are actually several resampling methods, but the bicubic family is probably the best for digital pictures without resorting to the more complex and expensive fractals methods. I definitely do not feel this is a one size fits all situation, as many authors have suggested. I'm going to try two of the methods, and compare the differences. I won't be using bicubic smoother because my experience tells me this probably won't work well for htis image. It tends to obscure the details when down sizing. I don't want this effect at all.
So I resize the image to 750x500 px. using bicubic as the resampling method, and create a snapshot.
Fig. 7
Then I rollback in the history, and resize using bicubic sharper as the method, and create another snapshot. Now I can easily toggle back and forth to compare the two methods, and decide which is best. Fig. 7 shows a split screen for you to see what I was looking at. Bicubic is on the right, bicubic sharper on the left.
Fig. 8
I don't like the white halos around the hooks, and the texture of the painted lure looks a little too sharpened for my taste using the bicubic sharper method. But the overall edges look pretty good, and at actual size, putting aside my nitpicks above, I prefer the overall look of the bicubic sharper method. There aren't very many pixels left after the resize, so masking and sharpening aren't really good choices. I've discovered that using the High Pass Filter is an excellent way to sharpen up images as the last step before saving in you destination output. Luckily, The TLR Pro Sharpening scripts have already worked out the settings for me, based on viewing device, format, and image size. Does it get any easier?
Fig. 9
Now all that is left is to flatten that puppy, and save it for the web. Well, not exactly - remember what I said about infusing quality to every step of the work flow? That applies here too. First thing, I need to convert the image to the destination color space. Since this is for the web, I want to use sRGB. I want to do this now, before I switch to 8 bits per channel. This important to preserve as much of the tonality as possible achieved by using a wide gamut working color space like Adobe RGB. Going to Edit > Convert to Profile... brings up the following dialog box:
Fig. 10
Then I can go to Image > Mode > 8 bits/channel and then File > Save As... and choose JPEG, quality - 12. And there you have it: a finished image, ready to upload to your image hosting site to share with the world.
You want to see the picture don't you? If you've made it this far through both tutorials, you deserve it 
Nice in color, isn't it? Well, I wanted to do this in B&W as well, so I just used a simple channel mixer adjustment layer. Well, actually, simpler than that: it was a TLR Black &White action, yellow-green filter I believe. It's basically just a bunch of preset channel mixer layers; very easy to use. And, I think you've all seen the result, but I'll post it again.
|
|
8/Feb/07, 2:14 am
|
Send Email to jfrancho
Send PM to jfrancho
|
mjbanak
Registered user
Global user
Registered: 01-2006
Location: Harrogate, UK
Posts: 989
|
|
Reply | Quote
|
|
Re: Making a Picture: Part 2 Convert, De-Noise, Sharpen, Resize, Convert to B&W
Very thorough John, and a welcomed glimpse into your noggin. I don't really have much to say about the overall process' you have described. I think others may have different tools they utilize, but the rest was very instructive. Did you attempt any other "lure posing?" I was thinking maybe the lure sort of angled towards the lens face first, as opposed to being parallel with the rod. Also, do you do a lot of editing after you convert to B&W?
By the way...with your location, we expect a tutorial on snow photography!!
Last edited by mjbanak, 8/Feb/07, 3:15 pm
---
Mike
----------------------------
"Photography is more than a medium for factual communication of ideas. It is a creative art."
-Ansel Adams
----------------------------
Canon 40D, 17-85mm, 70-200 2.8L IS
|
|
8/Feb/07, 3:08 pm
|
Send Email to mjbanak
Send PM to mjbanak
Blog
|
crimbo
Administrator
Global user
Registered: 02-2006
Location: Shetland, UK
Posts: 3729
|
|
Reply | Quote
|
|
Re: Making a Picture: Part 2 Convert, De-Noise, Sharpen, Resize, Convert to B&W
I'll be honest...I have yet to read it all but thanks in advance for the shed load of info...I am sure I will end up using some of it...
Anybody else have a different work flow that they commonly use (we obviously have Martins)
---
Chris
60N
1W
http://www.paddle.shetland.co.uk
http://www.paddle.shetland.co.uk/my%20piccys/index.htm
|
|
8/Feb/07, 6:54 pm
|
Send Email to crimbo
Send PM to crimbo
|
jfrancho
Registered user
Global user
Registered: 01-2006
Location: Rochester, NY USA
Posts: 373
|
|
Reply | Quote
|
|
Re: Making a Picture: Part 2 Convert, De-Noise, Sharpen, Resize, Convert to B&W
quote:
mjbanak wrote:
Very thorough John, and a welcomed glimpse into your noggin. I don't really have much to say about the overall process' you have described. I think others may have different tools they utilize, but the rest was very instructive. Did you attempt any other "lure posing?" I was thinking maybe the lure sort of angled towards the lens face first, as opposed to being parallel with the rod. Also, do you do a lot of editing after you convert to B&W?
By the way...with your location, we expect a tutorial on snow photography!!
Thanks Mike.
I'd love to see where others' are using different tools, for instance many prefer DPP, Lightroom, C1, Bibble, Silkypix, DxO, etc. for their raw conversion. Perhaps a different raw tool would eliminate an aspect that I address later in the work flow. Or, perhaps someone has a different/better tool for sharpening. The idea behind this was to get the ideas flowing. There isn't much happening in this topic here, so anyone can feel free to add their thoughts and screen shots to this. If there are questions, that's cool too.
I did consider other poses long before getting to this point. I went so far as to sketch out a layup before even setting up the shot, a detail I neglected to share in Part 1.
Chiefly, I wanted the rod and reel "layered," behind the bait. To do this and still fill the frame with the bait, I needed to shoot across the table. If you look at the initial, available light images in Part 1, you'll see that at f/2.8, DOF is a serious issue. Even at f/11 it is pretty shallow at close to 1:1. I knew from the outset that light was going to be an issue, and in reality, I had just enough light to complete the image at f/8.
Bottom line, yes, I want to do exactly as you suggested. But I'll need more powerful lighting to do it right. Here is an earlier example of an angle I'd like to pursue: Rapala DT-10.
I have some other ideas that I'll be trying using more interesting substrate, and possible a little more clutter.
When I created the B&W image, I went back to the 16 bit/channel PSD file that I archive before resizing for the web - basically the image above just after Fig. 4. From there I added the channel mixer layer via the TLR action and made some tweaks to the individual CM settings to keep the exposure balanced. Then I resized it to 750x500 px. and ran the output sharpening script, and made the JPEG from that.
This image was really easy. The "Two Content People" picture I posted in the Portraits section was considerable more difficult, requiring two CM adjustment layers, some dodge and burn, and more selective sharpening on grandpa's eyes. The whole monoConvo thing really is case by case business, as I'm sure you know.
Ahh yes the snow. The snow doesn't bother me that much, but it's been dreadfully cold and windy lately, creating wind chill factors that are unbearable for me. It's supposed to get into the low 20's °F next week, so perhaps I can get to the lake for some shots.
|
|
8/Feb/07, 10:54 pm
|
Send Email to jfrancho
Send PM to jfrancho
|
jfrancho
Registered user
Global user
Registered: 01-2006
Location: Rochester, NY USA
Posts: 373
|
|
Reply | Quote
|
|
Re: Making a Picture: Part 2 Convert, De-Noise, Sharpen, Resize, Convert to B&W
quote:
crimbo wrote:
I'll be honest...I have yet to read it all but thanks in advance for the shed load of info...I am sure I will end up using some of it...
Anybody else have a different work flow that they commonly use (we obviously have Martins)
You're welcome, Chris. Martin's stuff focuses mainly on what he does to get the image to black & white, and I didn't think it was as necessary to regurgitate his well documented process. He also has a tremendous amount of knowledge about creating and using certain effects to either embellish or enhance details that already in the image. Truthfully, I really admire how a seemingly simple and usually somewhat dark capture is transformed into such art by his hands. This is more of a "straight" shot, and didn't need too much embellishment. Most of the work was done in Part 1 through lighting and careful setup.
|
|
8/Feb/07, 11:00 pm
|
Send Email to jfrancho
Send PM to jfrancho
|
Add a reply
Link to us
- Blogs
- Hall of Honour
- Chat
|
You are not logged in (login)
Board's time is: 30/Nov/09, 6:57 am
|
|
|
