This is a rather long article. It covers a lot of territory. You might want to scroll through first to skim the topics and find the ones that interest you most.
I received some inquiries about how the masks generated by the TK8 plugin compare with similar masks generated by the new masking capabilities in Lightroom and Camera Raw. I hadn’t investigated this as I don’t often use Lightroom/Camera Raw masks, but Dave Kelly gave me a few tips, and I was excited to see what I could learn.
At first I wasn’t sure how to proceed. In my mask experimentation in Camera Raw I was relying on the color overlay in to help me “see” the mask that was being generated. This usually looked good, but would not provide a usable comparison with TK8 masks, which are grayscale. Luckily, Dave showed me the “White on Black” option for displaying masks in Camera Raw.
This displays the Camera Raw mask in grayscale. With this option selected (and assuming it works like masks in Photoshop where white reveals, black conceals, and gray partially reveals/conceals) it would now be possible to make a more direct comparison between masks created in Lightroom/Camera Raw and those created via TK8. If you decide to compare masks on your own, be sure to choose this option.
The image I chose for testing is the one below. This is a raw file from my iPhone captured as a DNG file using the Lightroom Mobile app.
I wanted to test the Luminosity and Color masks in TK8 with the corresponding luminance and color range masks in Lightroom/Camera Raw. There is plenty of color in this image for color masks, and I made adjustments in Camera Raw to produce a histogram that displayed a wide-range of tonal values without clipping either the shadows of the highlights. The Camera Raw histogram is below.
In order compare the different masks side-by-side, I needed to screen capture of the mask generated in Camera Raw and then copy it into Photoshop. I have an Adobe RGB monitor and so dropped the screen captures into a document with the Adobe RGB color profile for cropping and resizing to fit this article.
For the TK8 masks, I simply opened the raw file into a document with the Adobe RGB color profile Photoshop. The TK8 panel could then be used to make the necessary masks. The “output to image” button on the panel would output the mask as a pixel layer. From there it was again resized to fit the article.
Luminosity masks: Lights-1
The most logical place to start when comparing these masks is the Lights-1 luminosity mask. This is the most basic of all luminosity masks and is easy to generate even without Lightroom/Camera Raw or the TK8 plugin. However, the image below shows the setting used to create the Lights-1 masks for this comparison.
The two different Lights-1 masks are shown side-by-side below, Lightroom/Camera Raw on the left and TK8 on the right. Drag the divider left and right to see the differences.
Based on these two mask, it’s possible to make a few general observations.
Lightroom/Camera Raw masks have more contrast.
The most striking difference between these two masks is the contrast. The Lr/CR mask is high-contrast and the TK8 mask is lower contrast. There are a more middle gray tones in the TK8 mask while the Lr/CR mask contains more textureless white and black areas. So, even though these masks are constructed to select the same tones, the end result is not the same.
Lightroom/Camera Raw masks have significantly more contrast than the original image.
Even though the original image does not appear to be particularly high-contrast, the Lights-1 mask created in Lightroom/Camera Raw is indeed high-contrast by comparison. The contrast in the TK8 mask more closely resembles the contrast of the original image, and this is borne out in the histograms of the original image and the masks shown below. Note how the Lr/CR Lights-1 histogram has significant clipping in both the shadows and the highlights.
TK8 masks have more flexible contrast adjustment options.
Seeing these histograms points out another distinct difference between the TK8 masks and Lightroom/Camera Raw masks and that is that TK8 masks can easily have their contrast adjusted higher OR lower. For Lightroom/Camera Raw mask, on the other hand, there is no way to decrease the mask’s contrast. After seeing the Lr/CR mask, the user might like to experiment with a lower-contrast version, but this is not possible. The amount of pure black and pure white in the mask can both be adjusted in Lightroom/Camera Raw, but only in a way that INCREASES overall contrast in the mask. There is no way to lower the Lights-1 mask contrast in Lightroom/Camera Raw.
The TK8 plugin has a modification section available as part of the the mask-making process. It’s relatively easy to perform a Levels adjustment using the red-outline button shown in the image below to adjust contrast .
Lightroom/Camera Raw masks appear slightly blurry (though this is likely just my image).
The Lr/CR mask appears slightly out-of-focus compared to the TK8 mask. However, I’m not convinced that this is an actual property of the mask. This was an image from my iPhone that I had readily available that had good color and contrast. I looked at other images from my dSLR after I started writing this, and they did not appear to have the same level of blur, so this looks like it might just be something associated the DNG files from the iPhone.
Creating matching masks
After seeing the high contrast of the Lr/CR mask, I thought it might be interesting to see if I could replicate it by adjusting the TK8 mask to have more contrast. I just eyeballed it, and the image below compares my results with the Lr/CR mask.
It’s not quite perfect, but it’s close. The Levels adjustment on the original TK8 Lights-1 mask needed to achieve this result is shown below.
I have tinted the clipped highlights red and the clipped shadows blue to show what was “lost” as part of this adjustment, and this is perhaps a good time to discuss what this clipping means in terms of mask performance. One of the most important benefits of using masks generated directly from pixel-level data (like luminosity, color, zone and “range” masks), is how well they can help adjustments through the mask blend into the image. There’s no need to blur the mask to facilitate blending because it is already perfectly feathered at the pixel level to seamlessly match the image. However, clipped highlights and shadows in these masks means the pixel-level feathering is degraded. The dark gray tones go to pure black and the delicate light gray tones go to pure white, and this was clearly seen in the histogram for the Lightroom/Camera Raw Lights-1 mask above. Having the dark gray tones end up black in the mask is not a huge problem. The corresponding tones in the image were barely being revealed in the mask and so wouldn’t receive much adjustment anyway once this mask was deployed.
However, having the light gray tones in the mask clipped to pure white is a different situation. These pixels are already significantly revealed by the mask, and pushing them to pure white means they are now 100% revealed in the mask. For the Levels adjustment shown above, all tonal gradation in the highlights is lost above a pixel value of 171. The light gray highlights present in the original TK8 Lights-1 mask no longer match their pixel-level value in the image and therefore are no longer able to blend adjustments through this mask proportionally to their corresponding pixel-level value, in this case pixel luminance. All pixels above a luminance of 171 will be adjusted as if they have a luminance of 255. This means that there is essentially no mask in place for pixels brighter than 171. They all receive the same 100% adjustment through the clipped mask even though they actually have different tonal values in the image.
Another area of concern with high-contrast masks is the potential for more-obvious edges as the amount of adjustment increases. As gray is removed from the mask, there is less of a transition between the pure black and pure white parts of the mask. This transition zone, which is important for smooth blending at the edges, shrinks. Some types of strong adjustments, like Exposure and Hue changes, can have more obvious edges when this occurs.
Other luminosity masks
All the luminance range masks in Lightroom/Camera Raw have similar properties to the Lights-1 mask described above.
- They’re relatively high-contrast and have more pure black and pure white than their TK8 counterparts.
- The histogram of the mask contains clipped highlights and shadows.
- Contrast can be increased in the Lightroom/Camera Raw masks, but it cannot be decreased to match TK8 masks.
Below are side-by-side comparisons of additional Lightroom/Camera Raw masks with TK8 masks
The image below shows how a Darks-1 mask is created both in Lightroom/Camera Raw and TK8. Darks-1 preferentially selects the darkest tones in the image and tapers off into the lighter tones, selecting lesser and lesser light tones along the way. Tones that are pure white in the image are 0% selected in the Darks-1 mask.
The resulting masks are shown below. Again, pull the divider left and right to see the difference.
A Midtones-3 mask selects the midtones in the image and excludes the lightest highlights and the darkest shadows. The image below shows how these masks are created in Lightroom/Camera Raw and TK8.
Below are the Midtones-3 masks from the two different sources. They both exclude the highlights and shadows, but the Lightroom/Camera Raw version clips additional shadows and highlights even though the sliders for tonal feathering are set to their extreme dark and light positions (0 and 100, respectively). The TK8 version has dramatically less contrast because it has no added clipping. The TK8 masks will therefore allow for adjustments to be revealed in a wider tonal range in the image.
Zone masks are another type of “luminosity mask” that target specific tonal values between pure black and pure white. The targeted value can vary, and the user usually clicks on the image to choose a tone. The chosen tone is lightest in the mask and then tapers off into the surrounding lighter AND darker tones.
It could be argued that nearly all the luminance range masks in Lightroom/Camera Raw meet this definition for a zone mask as this is how these masks are generated. The Lights-1, Darks-1, and Midtones-3 masks discussed above were just specific variations of the technique for creating luminance range masks in order to have masks to compare with their TK8 counterparts.
To make this zone mask comparison, I switched it around. I first made a luminance range mask using Lightroom/Camera Raw by clicking on a tone in the image to create a mask. Then I used the numbers from that mask to create a corresponding mask in TK8 using the Multi-Mask module’s Zone mask function. The image below shows the setup for both masks.
The resulting masks are shown below.
No surprise, this Lightroom/Camera Raw zone mask has the highest contrast of all the masks compared so far. That’s because the clipping was already built into the mask in the “Select Luminance” settings (shown above) when it was created. The “15” and “85” numbers create tighter clipping. Of course, the user does not have to accept these settings. In the same way that I removed as much clipping as possible for the Lights-1, Darks-1, and Midtones-3 masks when creating them in Lightroom/Camera Raw, the user could move the outer sliders for this zone mask to the outer extremes (0 for shadows and 100 for highlights), to lower the mask’s overall contrast. However, there will still be significant clipping of the mask’s shadow and highlight values because, as should be clear by now, that’s just the way the luminance range masks work.
NOTE: Clipping the shadows and highlights in the mask is NOT the same as clipping the shadows and highlights in the actual image. Clipping these values in the mask just means that the MASK’s darker gray values are pushed to black, lighter gray values are pushed to white, and there are fewer midtone grays.
Color range masks in Lightroom/Camera Raw offer a bit more control over contrast than the luminance range masks because the “Refine” slider can be used to shrink or expand the range of selected colors. To get less contrast, move the “Refine” slider to the right to expand the color range. Below are the settings I used to create the color mask comparison.
In the Lr/CR mask below, the “blue pie” icon that shows the color that was sampled for both the Lr/CR mask and the TK8 mask. Slide the divider left and right to see the differences in these masks.
Again, the Lr/CR mask is still higher contrast than the TK8 mask. The histograms for these masks (below) again show the higher contrast in the Lr/CR mask compared to the TK8 mask. The Lr/CR mask has clipping in both the highlights and shadows. TK8 is only clipped in the shadows, but not clipped in the highlights, and, as mentioned above, clipping the shadows is less problematic than clipping the highlights.
NOTE: With TK8 color masks it is possible with saturated colors to clip the mask in the highlights by pushing the brightness slider to a higher value. However, the user has complete control over this and can easily avoid clipping highlights in the TK8 color mask by adjusting the brightness slider to prevent this. With Lightroom/Camera Raw color masks there is no way to avoid clipping both highlights and shadows.
There are many factors that will influence how well a particular mask based on pixel-level data performs. There are many reasons for using such a mask and many different types of images where it can be applied. A low-contrast mask of this type will always provide smoother blending of an adjustment into the image compared to a high-contrast mask. However, a low-contrast mask also spreads the adjustment more broadly across the image and might “bleed” the adjustment into areas where it’s not wanted. So, it’s highly desirable to create a mask that has the “right” contrast that reveals the parts of the image that need adjustment, conceals the parts that should not be adjusted, and provides a smooth transition between these two extremes.
For small adjustments, especially those that can differentiate between light and dark pixels (like Curves in Photoshop and Highlights and Shadows in Lightroom), masks with clipped highlights and/or shadows probably aren’t an issue. The inherent ability of the adjustment to differentiate between lighter and darker pixels in the image means the adjustment can still be applied proportionally even in areas 100% revealed by the mask. For adjustments like Hue and Exposure that don’t inherently differentiate pixel brightness, it’s more important that the mask be able to do this especially if a significant adjustment is necessary.
In my somewhat limited testing in Camera Raw, I’ve found that the following sliders generally produce natural results with good blending for most Lightroom/Camera Raw luminance and color range masks:
The following adjustments, especially if the adjustment was anything beyond minimal, quickly led to results with obvious edges that did not blend in well through Lightroom/Camera Raw luminance range and color range masks
Intermediate results were obtained with other Lightroom/Camera Raw sliders. Small to moderate adjustments looked OK, but bigger adjustments sometimes not so good.
Just to be clear, I do not consider myself a “user” of the Lightroom/Camera Raw luminance and color masks, and the above assessments were based on testing only a few images. I would fully expect frequent-flyers with Lightroom/Camera Raw masks will have a different opinion. Please feel free to leave a comment about your experience using the luminance and color range masks in Lightroom/Camera Raw.
The one thing that should be quite clear from this article is that luminance range and color range masks created in Lightroom/Camera Raw are indeed different from TK8 masks. The masks in Lightroom/Camera Raw are always higher contrast than the corresponding masks made with TK8. Additionally the TK8 plugin’s ability to modify masks means it can make high-contrast masks that are similar to those in Lightroom/Camera Raw, but the opposite is not true. Lightroom and Camera Raw are not capable of making low-contrast luminance and color masks that match TK8.
Does this matter? Well, at least theoretically, yes it does. Pushing mask shadow detail to pure black and mask highlight detail to pure white means one of the primary advantages of these masks based on pixel-level luminance or color is compromised. These masks are self-feathering precisely because some pixel-level value in the image (luminance or color) is matched pixel-for-pixlel in the mask. So, not only is the fine separation of tones and colors lost in the shadows and highlights when these values are clipped in the mask, but the remaining tones now have exaggerated contrast that will be less well matched to the actual contrast in the image.
Once pixel-level data is incorporated into masks, the potential for better blending of adjustments through the mask increases considerably.
In reality, though, any mask based on pixel-level data, even a high-contrast one, will likely produce better results than a hard edge selection that selects pixels based on their physical proximity to each other, like with Photoshop’s Lasso tool. Once pixel-level data is incorporated into masks, the potential for better blending of adjustments through the mask increases considerably. Additionally, there are times when a high-contrast, sharper-edged mask is desirable in order to prevent an adjustment from bleeding into tones and colors where it’s not wanted. So, while high-contrast masks would not be my preferred starting point for making masks based on pixel-level data, having access to these masks, even high-contrast versions, is better than having just radial and linear gradients or brushes from which to make masks.
NOTE: If you are using a color overlay to view your masks in Lightroom/Camera Raw, you might want to try the “White on Black” option discussed at the beginning of this article. It potentially gives a better visualization the mask that’s being created and can always be toggled off to view the actual image if that’s necessary.
One final point after making and comparing these different masks (and it’s purely personal), is that I like making masks using TK8 better than using Camera Raw. I like the ability to simply click on a button to get a Lights-1, Darks-1, or Midtones-3 mask instead of having to adjust four luminance range sliders. I also like starting with a low-contrast mask and using the modify options in TK8 to adjust contrast to get it looking exactly the way I want before using it. Lightroom and Camera Raw have the advantage of being able to directly make adjustments through the mask as soon as it’s created whereas TK8 masks need to be output to an adjustment layer as a layer mask or as a selection in order to be used, but first and foremost, I want to create the best mask I can before using it, and TK8 lets me do that. Finally, I think the mask calculator in TK8 is more intuitive for combining masks than the methods in Lightroom/Camera Raw, but have a feeling that those more familiar with Lightroom/Camera Raw probably prefer it. In the end, masks from TK8 and Lightroom/Camera Raw have both provided new possibilities for using pixel-level data for creating masks, and users now have options in both Lightroom/Camera Raw and Photoshop to incorporate these masks into their workflow.
Sean Bagshaw has a good video that provides a broader perspective comparing Lightroom/Camera Raw masks with TK8 and Photoshop masks. I’ve linked to this before in another article, but it also fits well with the topics discussed here.
If you have any thoughts or additional information to add to this topic of comparing masks, please leave a comment. I’d love to hear about your experience.
17 thoughts on “Comparing Masks: Lightroom/Camera Raw vs. TK8”
Thanks, Tony, for this article that analyzes LRc and PS TK8 masking. I found the contrast issue with LRc masks most helpful. I am more practiced in LRc than PS, but use the TK8 panel to enhance images beyond what LRc can do. About luminance range masks in LRc, I find them cumbersome and difficult to adjust, if I ever get the adjustment as I want it. The lights and darks masks in TK8 are much easier to use and to get at what I vision for my images.
Question, when should one group TK8 masks into groups and how do I determine whether to use a white vs. a black group? Does grouping masks in TK8 save memory? Perhaps you could discuss this question with Dave Kelly as well. I think many users would benefit from a tutorial on mask grouping.
As always, many thanks. Larry Greenbaum
Thanks for that information, Larry.
WRT grouping masks, I’ve been doing that less and less now. If you watch Dave Kelly on Friday you’ll see that he uses different methods like painting the mask onto an adjustment layer through a selection, or modifying the original mask with the mask-the-mask modify option, or getting a very customized mask by painting black on it or adjusting it with a Levels adjustment. He also uses the mask calculator frequently, which is also very effective at doing what a group might have been used for previously. Once you’ve created a mask that targets exactly what you want in the image, then groups becomes less important, I think. Each layer has it’s own masks for it’s own purpose. Not sure if that really answers your question, so let me know if you have additional thoughts.
Hi Tony. Thankyou for this summary, very much appreciated. After I started using your TK8 panel in PS I stopped using LR and PS masking because, as you illustrate here, the pixel-based approach is a huge improvement. I can refine the masks, in a subtle manner, very easily to suit any particular image. So thankyou for your work with this. Cheers.
I have been using Lr since its first public beta (around 2004 I think), it quickly became my goto application. When Adobe went to the subscription route, I suddenly had PsCC to work with and began some cautious experiments with Luminosity masks. My experiments were painful as I was not familiar with the Ps interface at all. Somewhere along the way, I discovered Sean Bagshaw’s channel and purchased my first of the TK products (TK6) and Sean’s education package for the plug-in. Now I use LrC only for library management, applying Linear Profiles and the most basic raw level adjustments, everything else is done in PsCC. I have looked at the new masking features in Lr/CR found the loss of fine control I’ve become accustomed to in your products a non-flier.
Thanks for making Ps so approachable and effective. Best of the Happy Holidays to you and yours.
Cheers, Gord Roberts
LikeLiked by 1 person
Thanks for that perspective and feedback, Gord. Be sure to give yourself some credit too. You’ve certainly made quite a bit of effort if you’re effectively using these tools in Ps, especially after using Lr for so long. I think we all tend to gravitate to those tools that work for us, and there are plenty of devoted Lr users still. The fact that Adobe added luminance and color range masks to Lr (finally) is an indication that they came to realize how valuable these masks were to photographers, and lots of Lr users seemed very happy to have them. I’ve never gotten into using the masks in Camera Raw very much, but comparing them did point out some potential advantages to TK8 masks. However, I think Lr users will still get significant benefit from the range masks that they can now access.
Timely piece. Reference to this comparison will go into my upcoming book which will of course, also reference Tony’s tools.
Matt Kloskowski in his Portraits tutorial also points out that even with the LR masking panel the masks are different if you select People vs selecting Subject especially if there is wispy hair. LR masks are many times approximate & PS are more specific. Depends on your need – if your mask is too hard for people then it looks like they were stuck on the background. I think you & all the LM gurus are wonderful instructors & make us aware of how we use your & Adobe masking tools.
Hi Tony, thank you for explaining why TK8 panel always produce better masks . I always use TK8 panel for creating masks,as they not only produce better masks but it has tools to modify masks like the mask calculator, masking the masks and even using luminosity masks and color masks togather, besides using levels and curves .
Thanks for your thoughts, Shailesh. The bottom line remains that the “right” mask is the one that does what you want it to do, and Lr/CR masks are very useful to many photographers also. I do agree that the ability to take better control of mask modification with TK8 is a plus when it comes to making the “right” mask. I also like the ability to make Lights, Darks, and Midtones masks very quickly. In the end I always support using whichever tools that work best for the individual photographer.
Thanks for detailling that out.
In fact I also recognized in LR that artefacts will appear when picture adjustments based on applied masks were done. This leads very fast to artificial looks of the picture. Altough I have sometimes to be careful with TK8 but it gives me a much more sensitiv possibility for adjustments.
All the best from Bavaria/Germany
That is always something to watch for. Luminosity and color masks generally blend way better than their hard-edge counterparts, like the Lasso or Marquee tool, but if you push hard enough with an adjustment, especially with a high contrast mask, you’ll still have edge issues. My goal when making masks is to try a retain detail an texture in the mask as much as possible as this is what helps make smoother blending when the adjustment is applied.
The only way I would give up my TK8 was if it was pried out of my cold dead hand 😉
Very interesting article, and thanks so much for increasing my knowledge of the subject. Also, thanks so much for your continuing work with the TK8 product. I learn so much from both yourself, and from Dave Kelly every Friday.
Excellent article Tony and the point made by Sean in the video resonates with me in that if you make initial edits in LR/ACR and then open as a smart object in PS then this gives the best of both worlds
Tony thanks for the review. I have a question, the LR/CR masks is being done on the raw file, the TK8 mask is being done on a PSD file that has had some adjustments made to it. Could that effect the results and is there any pros and cons of doing the masking in the raw vs the psd file?
I can’t really say how Lightroom/Camera Raw makes the range masks it generates. Perhaps it is limited to making high-contrast masks for some reason related to working with the raw file instead with a file that’s been converted to pixels. However, if Lr/CR can adjust contrast of the on-screen image (which it can), then it would also seem feasible that these apps could adjust contrast of the range masks as well. So, I still think the level of mask contrast is more likely an engineering decision than some type of actual limitation resulting from working directly with raw files. The pros and cons of the two masks types are basically which type of mask gives you the results you’re looking for and which is easier for you to use.