What's Needed?

[FrankMi]

I've been wrestling with this image, trying to get it right for my son. Here is the SOOC and after Post Processing. Any suggestions on what it might need.

Shot with the Sony DSC H1, f2.8, 1/40 sec.



I've changes so many things and restarted from scratch trying to get it right I'm not even sure I can remember what it did!

[MiniChris]

Much too hard to do a rework on anything but the original image. The biggest problem I see with the image is a lack of setting the black point. In failing to hit that mark, your image tended toward a red/blue overcast. I fiddled for a while before deciding that at this point, with a much tighter crop, it made a pretty decent B&W.

[jiro]

Hello, Frank. I think your White Balance is still a little bit off. Here is my suggestion for you:

1. Copy the background layer.

2. Apply your first level adjustment layer. Set the values of the RGB input data to:
R= 0, 1.15, 255
G = no change
B= 0, 0.66, 255 This will set your middle gray adjustment.

3. Apply another level adjustment layer and adjust the RGB input values to:
R = 9, 1.03, 251
G= 5, 1.02, 253
B = no change. This will set your adjustment on the black and white levels.

4. At the top most layer, combine all layers by pressing CTRL+ALT+SHIFT+E. This new layer will apply the two level adjustment layers into one combined layer.

5. Using this new combined layer, you could use my vivid light detail enhancer tutorial to add extra detail on the image.

6. Adjust the contrast by applying a very mild S-curve to add contrast to the image.

7. Darken the top background a little bit to increase perceived depth on the image.

8. Apply a little vignette to center the attention of the viewers to your kid.

9. Sharpen the image.

Here is the sample of the output when all the recommended adjustments are made:



and here is the photoshop file link should in case you want to further tweak the image yourself: http://www.mediafire.com/?5n9mgc840mlas38

Hope this could be of help. Good luck my friend.

[jjbacoomba]

Wow! Great job Jiro!

[MiniChris]

Hello, Frank. I think your White Balance is still a little bit off. Here is my suggestion for you:

1. Copy the background layer.

2. Apply your first level adjustment layer. Set the values of the RGB input data to:
R= 0, 1.15, 255
G = no change
B= 0, 0.66, 255 This will set your middle gray adjustment.

3. Apply another level adjustment layer and adjust the RGB input values to:
R = 9, 1.03, 251
G= 5, 1.02, 253
B = no change. This will set your adjustment on the black and white levels.

4. At the top most layer, combine all layers by pressing CTRL+ALT+SHIFT+E. This new layer will apply the two level adjustment layers into one combined layer.

5. Using this new combined layer, you could use my vivid light detail enhancer tutorial to add extra detail on the image.

6. Adjust the contrast by applying a very mild S-curve to add contrast to the image.

7. Darken the top background a little bit to increase perceived depth on the image.

8. Apply a little vignette to center the attention of the viewers to your kid.

9. Sharpen the image.

Here is the sample of the output when all the recommended adjustments are made:



and here is the photoshop file link should in case you want to further tweak the image yourself: http://www.mediafire.com/?5n9mgc840mlas38

Hope this could be of help. Good luck my friend.

I got to the same place as this, but couldn't rid myself of that reddish-bluish tint in the skin tones...not sure you did any better in that aspect but it is nice to know we went to the same place and in the same manner (and almost exactly so).

[Colin Southern]

Hi Frank,

Any better (be sure to click for bigger view)?

[FrankMi]

Much too hard to do a rework on anything but the original image. The biggest problem I see with the image is a lack of setting the black point. In failing to hit that mark, your image tended toward a red/blue overcast. I fiddled for a while before deciding that at this point, with a much tighter crop, it made a pretty decent B&W.

Hi Chris, thank you for taking the time to help me unravel this mystery.

I realized when I went back to the EFIX data that this was taken with the older Sony DSC-42 Cybershot and not the H1 as I originally thought. The SOOC image is the original JPG produced by the camera so I didn't have the benefit of the RAW on this one.

I'm not sure what you mean by failing to set the Black Point. I took the image into ACR and set the Black Point to 0 as anything higher clipped the shadows in the histogram. Did I use the wrong procedure for this situation? What is the relationship between having the Black Point at 0 and the red/blue overcast? Although I'm not overly fond of B&W images, this certainly addresses the skin-tone and other color tone issues I was having problems with.

[FrankMi]

Hi Willie! Great to see you again!

I think I need to clear up a bit of miss-communication on my part. Brandon, pictured here, is my grand-son. I was trying to get Brandon's image ready to send to his dad (my son Rob's family) to share. With 8 kids and 15 grand-kids, it can get a little confusing, particularly around Christmas!

Hello, Frank. I think your White Balance is still a little bit off.

Agreed. It was difficult to find a valid setting, without which, skin and other color tones were difficult to adjust properly!

Here is my suggestion for you:
1. Copy the background layer.
2. Apply your first level adjustment layer. Set the values of the RGB input data to:
R= 0, 1.15, 255
G = no change
B= 0, 0.66, 255 This will set your middle gray adjustment.

This is the first time I’ve tried setting the individual RGB values in the Levels Adjustment Layer. I used the Gray Eyedropper to sample the location you had targeted as #1. I got values all over the place. Is there a procedure for getting the correct values?

3. Apply another level adjustment layer and adjust the RGB input values to:
R = 9, 1.03, 251
G= 5, 1.02, 253
B = no change. This will set your adjustment on the black and white levels.

I see two Eyedroppers for white and black and I see where you have targeted a white and a black location on the image, but I don’t know the process for getting the values.

I suspect that you are following a process that I have not learned as yet. Rather than pester you with countless questions, is there a tutorial you could point me to that will get me on the same page as you are on?

I should be able to follow the rest of the steps as they look familiar to me.

4. At the top most layer, combine all layers by pressing CTRL+ALT+SHIFT+E. This new layer will apply the two level adjustment layers into one combined layer.

5. Using this new combined layer, you could use my vivid light detail enhancer tutorial to add extra detail on the image.

I believe you are referring to this one? High Contrast and Detail Effect using dodge and burn technique.

6. Adjust the contrast by applying a very mild S-curve to add contrast to the image.

7. Darken the top background a little bit to increase perceived depth on the image.

8. Apply a little vignette to center the attention of the viewers to your kid.

9. Sharpen the image.

Here is the sample of the output when all the recommended adjustments are made:



and here is the photoshop file link should in case you want to further tweak the image yourself: http://www.mediafire.com/?5n9mgc840mlas38

Hope this could be of help. Good luck my friend.

Thank you so much, Willie, for the pointers and sample file!

It seems like every time a get a bit cocky and think I know something, I run into an image like this and quickly realize just how little I understand and how much I still have to learn!

[jiro]

Hello again, Frank.

A little explanation on my process in finding a neutral gray on an image and setting the proper black and white levels using numbers.

The thing that can trick the WB correction function of any camera is what we call color cast. It happens all the time. That's the reason why it's much easier to control and make a white balance correction with a studio shot because you have full control of everything. The only problem that you will encounter with the WB issue on a studio setting is if you don't know exactly what you are doing. With indoor or outdoor shots using the sun as your main source of lighting naturally its rays will be reflected on the grass, on the nearby building, on the next person sitting beside you, etc. The colors of these elements will determine the color cast that would influence the skin tone of the person you are trying to take the photograph. This is the reason why professional photographers invest on some device to help them correct the white balance on their images. One of them would be an 18% gray card and others use the more versatile Xrite Color Checker Passport. It would be nice if we all have one of these WB guide with us. Unfortunately, most of us don't (including me).

Using an old trick in photoshop that I know, I will try to explain how to color correct and image using level adjustment layers BY THE NUMBERS. The first thing you need to do is to find where exactly in the image is the most neutral gray area to point the eyedropper tool. You have to do this first before you adjust the white and black levels so your neutral gray will still be reliable during the adjustments.

FINDING THE NEUTRAL GRAY AREA IN AN IMAGE:

1. Add an empty layer on top of the background layer by pressing SHIFT+CTRL+N. When prompted, use the default layer name and just press ENTER. Fill this layer with 50% gray by pressing EDIT > FILL > and then on the drop down options press 50% GRAY. Set the blending mode to normal and the opacity to 100% and then press OK. Now you'll notice that the layer is covered with gray and you can't see the background layer below it. We will use this 50% gray layer to find the neutral gray area. If you only use your eyes to scan your image (we will use your grandson's image as an example) you will logically choose any of the gray areas on the toy car as the spot to click the middle gray eyedropper tool. If you do so, you'll be surprised to find out that by randomly doing it, it will also give you different color cast based on the spot you chose. Why? Because each different part of the toy car is receiving different intensities of color reflection (or color cast) from the surroundings.

2. Now, by selecting this gray layer (make it active) change the blending mode to DIFFERENCE. The way the DIFFERENCE BLENDING MODE works is that it is comparing the top layer against the bottom layer and finding the difference between them. Anything that is different will show as different colors. Anything that does not change in color will appear as BLACK. What will do now, since we set the top layer to 50% NEUTRAL GRAY is to find where in the image has black areas based on the difference of the two layers. This is how your image should look like.



3. Apply a threshold adjustment layer. This will give you the location where you can find the neutral gray area. By default, if you would look at the adjustment layer window, you'll see that the image became totally black and the THRESHOLD LEVEL was set to 128. Why 128? That is the middle value between 0 (pure white) and 255 (pure black). There are 256 steps from 0 to 255 so don't get confused why we did not use 255 as the max value. We are talking to gradation steps and not just numbers.

Drag the threshold adjustment slider all the way to the left. Now, your image will appear as pure white. That is normal. Now, Try to drag the slider slowly towards to the right until you see black areas appearing. Don't click on the first area where the black spots appear. try to select on the area where a considerable size can be located. Also, a more logical way to know what area to select is by comparing that area from the original background image. Here, by adjusting the threshold slider to 14, noticed that I can select the black area between the two tires on the left side of the image. If you will hide the eye icon of the threshold and the 50% gray layer, you would see that this gray area actually looks gray from the original image.



4. Click the eyedropper tool and select the COLOR SAMPLER TOOL from the available menu. Use the Color Sampler Tool and click on the area we selected as our neutral gray. We will need the numerical RGB values of this area later. Now that we had selected where the neutral gray area is located in the image, we will not use the gray layer and the threshold layer anymore. You can send these 2 layers to the trash bin to delete them.

5. This is the common way to adjust the image when you've found the neutral 50% gray area: Apply a new LEVEL ADJUSTMENT LAYER to the image. Select the gray eyedropper tool (the one in the middle) inside the level adjustment layer window and click the area we selected on the image as our 50% gray and you're done.



As you can see, the image has this yellow orange cast coming from the late afternoon sun. The color is normal, but if you want to further tweak it this is where the NUMERICAL ADJUSTMENT PROCEDURE would come in. I will continue the tutorial on a new thread since this one is a bit too long. The process is very simple. The explanation became too long for me so you won't be lost in the tutorial. Once you practiced this trick even if you don't have a gray card with you, you can use this trick to bail you out on some tricky images. Hope this helps, Frank. Good luck.

[jiro]

CONTINUATION: CORRECTING COLOR CAST USING NUMERICAL VALUES IN LEVEL ADJUSTMENT LAYER.



As you can see on the top image. I pulled out the INFO window so we could see the RGB values of the gray area after we clicked the eyedropper tool on it. The original RGB reading values prior to using the gray eyedropper tool was:

R = 116, G = 125, and B = 162.

As you can see from the RGB values, the color cast came from the Red and Blue color spectrum. When we used the GRAY EYEDROPPER TOOL, Photoshop recommended a correction value of:

R = 137, G = 134, and B = 134.

Not bad, but the value 134 is a bit higher compared to our target value of 126. Now, here is the interesting part. From the original reading, GREEN = 125! Since we are dealing with 50% gray, what we need to do now is to adjust the RED and the BLUE middle sliders to coincide with the 125 values of GREEN. This is now adjusting or making color correction using numbers!

1. Go to the LEVEL ADJUSTMENT WINDOW and select the RED color on the drop down menu. Slide the middle slider to the right until the value of the RED is equal to 125.

2. Do the same with the GREEN and BLUE levels. Go to the LEVEL ADJUSTMENT WINDOW and select the GREEN and BLUE color (one by one) on the drop down menu. Slide the middle slider to the right until the values are equal to 125.

Here is the end result of the numerical adjustment:



Now that we're done with the gray area, we will now proceed with the BLACK and WHITE CORRECTION process.

1. PRESS CTRL+ ALT + SHIFT + E to combine all the layers again into one new layer. This will apply the neutral gray adjustments that we did before on a new layer on top of the background layer.

2. On top of this new combined layer, apply a new THRESHOLD ADJUSTMENT LAYER by pressing LAYER > NEW ADJUSTMENT LAYER > THRESHOLD. This is how it should look like:



Slide the threshold slider to the left. The image should look totally white.

3. Now, slowly slide the slider to the right. At a value of 8, I saw a good area from the bottom rear tire (very near our previous neutral gray area). Use the color sampler eyedropper tool again and click on this area. This will now be our TRUE BLACK area. We are now dealing with the black and white level adjustments here and not about gray areas anymore. Upon clicking, the RGB value reads:

R = 7, G = 5, and B = 1.



4. Continue sliding the threshold slider to the right until it goes to a value of 255. The image will go almost completely black. Now slide the slider to the left until you see white areas coming out. I selected the area of the left sleeve of the boy when my threshold slider level reached a value of 235. This will be my WHITE area guide. Upon clicking it using the color sampler eyedropper tool, the RGB value reads:

R = 239, G = 240, and B = 237.



The threshold adjustment layer has already done its function so we don't need it anymore. You can send it to the trash bin. We will now use a new LEVEL ADJUSTMENT LAYER to numerically correct the black and white level values.

5. Go to LAYERS > NEW ADJUSTMENTS LAYERS > LEVELS and press OK.

Select the RED Color and adjust the value of the Black input data to 6. This will adjust the numerical value of RED from 7 to 1. Why 1? because 1 is the lowest value on the RGB levels from our initial findings at procedure 3: R = 7, G = 5, and B = 1.

since we are dealing with black areas, we need to select which among the RGB values is the lowest. In this case it is the Blue value of 1. We will now try to adjust the RED and the GREEN to match its lowest values to 1.

You know the drill, so adjust the GREEN value accordingly.

6. One point of importance here is that WHILE ADJUSTING THE SLIDERS, you will notice that THE VALUES OF THE MIDDLE SLIDER IS ALSO CHANGING. To correct this, you will need to re-adjust the middle slider until your final values reaches the values of 1 as the lowest for the black sliders, 125 for middle sliders, and 240 for the white sliders.

This is how the values and the final image should look like:



7. By adding the extra processes, you would closely arrive at the end result based on your personal preference on the contrast, sharpness, and vignette that you will add to the image.

Hope this could be additional help to you, Frank.

[FrankMi]

Hi Frank,

Any better (be sure to click for bigger view)?

A little different. I'm seeing a pale blue reflection in the buggy and to be honest, I've been looking at this image for two days and I'm starting to loose track of what I though it should look like! Part of me says it looks great and part of me says that something's not quite right but I can't put my finger on what it is that doesn't feel right. Maybe it's just me aging eyes. <sigh>

[FrankMi]

Wow Willie! You never fail to impress me with both your knowledge and your ability to present complex material in a way that is relatively easy to follow!

It will take me some time to absorb all that you have presented here and get my copy of the image to respond accordingly. I'll post the results when I can get all my ducks in a row! Thank you so much for taking the time and effort to make this thorny issue clearer. I certainly hope that others notice this thread as I'm sure there are others that have run into this issue.

[Kris V]

Jiro, Thanks for this post. Saved for further reference!

[FrankMi]

Wow, what a difference Willie's White Balance process makes! I couldn't figure out what was wrong with the image until I could see the before and after side by side. The original has a blue color case that wasn't obvious against the foliage and other colors in the image. OK! Learned something new! Thank you Willie!

Here is how it turned out.