Remove dust spots and power lines automatically

Use discount code GREGBENZ for $10 off Luminar Neo (valid through May 10).

Wouldn’t it be nice if you could get rid of distractions in your image with the click of a button? With the latest artificial intelligence (AI) software you can. But you need to know a couple simple tricks to ensure you don’t just swap one problem for another. In this tutorial, you’ll learn how to eliminate 80% of the tedious work required to clone out dust spots and power lines.

I previously demonstrated how you can use Skylum’s Luminar to add beams of light or other creative effects to your images (links below). Skylum has iterated the product quite a bit over the years and the latest version, Luminar Neo, contains a couple of tools that automatically remove dust spots and power lines. You’ll get substantial improvements with a single click. Combine that with a few simple manual cleanup steps and you can great results in a fraction of the time normally required.

Tools built with AI tend to provoke one of two immediate reactions: “amazing!” or “awful!”. That’s generally built on the assumption that the tool should do all the work. Perfection is not a realistic expectation for most AI and overlooks the real value it offers. There are very few cases (now or in the near future) where AI will produce perfect results every time. However, there are many cases where it can help you save time or produce better results if you combine the AI with some additional manual work. Even if your first impression was “amazing”, there’s a very good chance that you can get even better results by viewing AI as one step in a larger process – rather than as a crutch to do all the work for you.

So with that in mind, I’m going to outline a couple of workflows that I believe will help make Neo an incredibly powerful tool for avoiding a lot (but not all) of the drudgery of removing dust spots and power lines.


Removing dust, power lines, and other distractions:

The key workflow with Neo is:

  • Create a single layer to adjust. Ideally this is a Smart Object to work non-destructively. This will allow you to make changes to the underlying content at any time, without minimal to no need to redo the work you’ve done with Neo. For example, you can go back and change exposure or white balance and Neo will update its work automatically.
    • If you need to adjust across multiple layers, <shift>-click to select all of the layers, right click, and convert to a Smart Object.
    • If you don’t like working with Smart Object, create a stamp of your image (via <cmd/ctrl><alt/option><shift>-E) instead so that you preserve your original and can apply Neo’s changes selectively through a layer mask.
  • Select the layer to correct.
  • PS: Filter / Skylum Software / Luminar Neo **
  • Neo: Edit (at the top), Erase (blue icon at right), and click on “Remove Powerlines“, “Remove Dust Spots“, or both.
  • Click “Apply” (you do not have to wait for the preview to update to do this).

** Note that you can record this step as an action in Photoshop to apply more quickly or even as a batch operation. I’d include native support in Lumenzia to further automate the process, but Neo is not currently designed to allow 3rd-party developers to request that it remove dust and power lines. I’ve contacted them to request this, and hopefully we’ll see support in the future.


If you wish to use Neo to remove other types of objects, the top of the same “Erase” tool has an “erase” button to help do that. The basic workflow to use select/deselect to paint a red target area and then click “erase” to fix that red area. If you need to undo, make another red selection and click “restore” (or use the circular arrow at top-right to “reset tool” and undo all the changes).


Visualizing dust spots and Neo’s changes:

No AI is 100% perfect, so it helps to review what it has done so you can refine and perfect the results. Here are a few things to look out for:

  • The dust spot removal is incredible in skies and I would mask it into any sky area without much thought. However, it can cause slightly negative changes to other areas of detail. In particular, I’ve seen some loss of shadow detail. So be sure to review its impact when applying it to anything outside the sky or smooth areas.
  • The power line removal tool is incredible in sky areas as well. Occasionally it misses a line – especially if that line is very thin.
  • Shadows of power lines seem to be ignored, so you may need to manually clone them out.
  • Watch out for areas where power lines cross over buildings or other areas of clear detail. Use the manual refinement steps below to repair these areas as needed.


In practice, I’ve found that Neo is so good with dust spots in areas of clear sky that it’s safe to simply mask it into the sky without needing to confirm. For areas of detail outside the sky, it’s best to review the changes the software is making and use a mask to selectively reveal it only where it is helpful. There are a few ways to do this:

  • You can use the “isolate” (eyeball icon) button in the Lumenzia‘s Basics panel to quickly review what changes a filter is making on a Smart Object. Once you’ve applied Neo to a Smart Object, just select the Smart Object and <ctrl/cmd>-click the isolate button. This will create a visualization showing all the areas which have changes that have not been applied (darker areas either did not change or the changes have been applied because the filter mask is white in that area). Just paint white on the filter mask anywhere you’d like to accept the changes from Neo. When you’re done, click “isolate” again and your edited filter mask will be loaded back to the original layer and the visualization removed (you can alternatively manually delete the layers or click “X” if you wish to discard the preview and any changes).
  • If you don’t have Lumenzia and you’re removing dust, you can use the “visualize spots” feature in Adobe Camera RAW (ACR) to see any remaining dust spots (keyboard shortcut: cmd/ctrl-A to invoke the filter, <B> to switch to the healing panel and then <Y> to visualize). This is a more iterative process, since you cannot edit the filter mask while using this visualization, but it is helpful.
  • A more generalized manual workflow (for power lines and such) is to create a duplicate of the layer, change it to difference blend mode (which highlights changes), and you’ll probably want to add a curves or levels adjustment above that difference layer to make the differences more obvious. Make sure you set the layer or filter mask on your working layer to black so that you are comparing the changes (otherwise you’re comparing two versions of the adjusted image, which is identical and will show pure black with the difference blend mode). When you’re done, delete the extra layers (the duplicate in difference mode and any levels/curves you added).


Manual refinement:

Once you’ve reviewed Neo’s changes, you may still need to do some manual cloning in areas the AI missed or where you don’t feel it did a good job (such as possible artifacts when removing power lines in front of detailed buildings). At this point, you should use your normal cloning techniques. I prefer to create a new blank layer and use the clone and healing tools set to “current and below” or “sample all layers”. This approach allows you to work non-destructively by preserving your Smart Object and allowing you to easily undo or update the cloning later.

For power lines, you can save a lot of time by taking advantage of the fact that power lines are typically nearly-straight lines. With the spot healing brush, just click on one end of a straight section and then <shift>-click on the other end of that straight section.*** This will cause the spot healing brush to be applied along the line connecting those two points. If you’ve used a soft brush large enough to cover the line, it should be removed. You might need to touch up a few points, but you’re now doing that detailed work in only a small portion of the image.

*** There are a few things to watch for if you’re using a Wacom pen with the spot healing brush. There are 2 different settings which may cause the spot healing brush to be much smaller than you desire. Both of these set the size proportional to the pressure you use with the pen, but it becomes very unpredictable when clicking on a point (nearly impossible to get the full size). The first one to check is the icon with concentric circles in the toolbar is off / not dark (this has a tooltip which says “always use pressure for size…”). The second is a very hidden dropdown inside the brush options (same place where you have sliders for hardness and spacing). At the bottom of those popout options, “size” should be set to off (if set to “pen pressure”, this has the same effect as the icon button). Alternatively, you can just use a mouse for these steps to avoid issues with pressure-sensitivity.


Other great tools in Luminar Neo

I previously demonstrated a few other features from Luminar 4 which are part of Luminar Neo (in a slightly different locations, but they work the same):

  • The “Sunrays”  filter is amazing for adding beams of light to your images (jump to 9:20 in the video for Luminar).
  • The Accent AI slider can add general “pop” and interst to your image. This  feature is now found under “Enhance” in Neo. The tutorial also shows how you can use luminosity masks to help get the best results with Luminar or any 3rd-party software.

Neo has several other powerful features. My favorites include Atmosphere AI to add fog, Toning for sky color, Landscape (golden hour) for sky color, Sky AI for quick sky replacements, and Develop’s white balance sliders or Color Harmony Color Contrast / Split Color for correcting non-RAW color.


Use discount code GREGBENZ for $10 off Luminar Neo (valid through May 10).

[Disclosure:  This post contains affiliate links.  If you purchase through these links, a small percentage of the sale will be used to help fund the content on this site, but the price you pay remains the same.  Please see my ethics statement if you have any questions.]

How to Process Milky Way Photos

The Milky Way is one of the most awe-inspiring sights you’ll ever photograph, but it can be tricky to post-process the RAW images into something that truly captures the experience. In this tutorial, you’ll learn how to use RAW processing and luminosity masks to create gorgeous night skies. Be sure to see the list below with several more tutorials to help improve your night sky images using Photoshop and Lumenzia.

Key steps for making the Milky Way pop in your images:

  • Proper white balance in RAW. The colors are subtle and can easily get lost if you don’t spend the time to get the right balance.
  • Use the camera calibration tab to boost color.
  • Lighten the Milky Way using something like a +brightness adjustment revealed through a lights luminosity mask. In Lumenzia: create the adjustment, alt/option-click mask for a black mask, then L for a lights preview, slide or customize if needed, “Sel” to convert the preview into a selection, and then paint white through the selection onto the mask to brighten the light parts of the Milky Way.
  • Darken the surrounding sky using something like a -brightness adjustment revealed through a darks luminosity mask.

Other related tutorials for shooting and editing images of the night sky:

Lumenzia LITE: free luminosity masking panel for Photoshop

I’m happy to announce the release of a FREE luminosity masking panel for Photoshop: Lumenzia LITE. This panel allows you to quickly create 16-bit luminosity masks, channels, and selections. It’s completely new and a major improvement over my previous free panel (including native support for Apple Silicon / M1). In the video below, you’ll learn how the LITE panel works, how you can use it to create some gorgeous images, and some of the ways it differs from the full version of Lumenzia. You can download and install anytime via the footer of this and future newsletters.

Learn more about Lumenzia LITE here (including how it compares to the premium version of Lumenzia).

What’s new in Web Sharp Pro v3.7?

Web Sharp Pro makes it quick and easy to create gorgeous exports of your images for sharing. It offers much more than just sharpening, with custom templates, blurred borders, custom watermarks, advanced crops, film grain, batch processing, and much more. The best way to understand it is with the short demonstration below, and you can also learn more about the rest of its extensive capabilities here.

Web Sharp Pro v3 has advanced steadily with a series of free updates since its release last year and now v3.7 builds on these great features with secondary borders (under Settings / Quick Export), webP exports, and more. See the release notes for details on the 137 new features, updates, and bug fixes that have come out just since Photoshop 2022 was released in October.

Get Web Sharp Pro here.

Exporting webP files from Photoshop

Photoshop v23.2 just added support for a new file format: webP. In this tutorial, you’ll learn what webP is, why you should consider using it, and how you can easily take advantage of it with Web Sharp Pro.

What is webP and when should you use it?

The webP format is an open standard licensed for free from Google. It is designed to offer smaller file sizes on the web at similar or higher quality. Unlike JPG, it offers a completely lossless format to completely avoid any artifacts (albeit with a larger file size than a high-quality JPG). And unlike JPG, it offers support for transparency, which is critical when you want to allow the background to show around your subject (such as a product being sold on a website). If you’ve been exporting images as PNG to support transparency, webP offers massive reductions in file size (about 30% smaller lossless and >80% when saving in a lossy format comparable with high quality).

It is worth noting that there is an important difference between the lossless format and 100% quality lossy. Very small areas of color detail (such as a red light in a cityscape) may look significantly desatured compared to the lossless version. In general, you’ll probably need to look extremely close to notice the difference, but high-frequency color is one use case where I think the extra size of the lossless format may make sense. I don’t know if this behavior is inherent to the webP format or Adobe’s implementation of it, but I’ve seen differences in several test images. Aside from this narrow difference, I generally find the high-quality lossy version would look identical to the lossless version for most viewers at a normal viewing distance. You only see the differences when you’re pixel peeping pretty closely. And keep in mind that all JPGs are lossy, so we’ve been dealing with this for a long time.

This file format has already been around for a decade, so there is already widespread support including every major browser (Internet Explorer does not support it, but IE will be discontinued in June). So webP is an outstanding choice for sharing images on your website when you need transparency, and also offers useful file size reduction compared to JPG. This is a great way to improve page load times and save bandwidth costs for your website.

In addition to use on websites, webP is a great way to send smaller files for email or otherwise sharing with others. MacOS has native support via its image preview software and there are free tools for Windows. However, as that may require installing or configuring for support on Windows, I probably wouldn’t use webP to email everyone in your list just yet. However, this can be a nice way to hold down the size of your email account, save on wireless data costs, and send images over slow connections. Support for this standard keeps growing, and I expect we’ll see more and more JPG usage migrate to webP over time.

The latest version of Adobe Bridge supports webP, but Lightroom does not at this time. However, as this is a format for export, that shouldn’t matter unless you manage a set of pre-prepared images to share on demand.

This leaves the question of when you should keep using JPG or PNG:

  • Keep using JPG whenever you just want to know something will work. No need to test if you can upload to social media (though sites like Facebook support it) or know if your friend’s computer will be able to open the image. The file size savings are nice, but probably not worth a lot of hassle for most people.
  • Keep using JPG if you see a loss of saturation/detail in small vibrant colors, such as vibrant city lights at night. You can use the lossless webP format to avoid the issue, but the resulting file is larger than a JPG. Using 100% quality with webP will not prevent the issue, all lossy values are at risk. This is unlikely to the vast majority of images (especially when viewed at a normal distance), but it is an edge case to keep in mind. If you can’t see any image at 100%, it’s safe to use even if you can see something when you pixel peep further.
  • Keep using PNG when if you run into the extremely rare scenario where there is visible banding in your output and you need 16-bits to avoid it. You pay a large size penalty for this, but there may be times when it is worth it. Don’t do it because 16-bits sounds good (while 16-bits is critical in your source and working files, it rarely makes a difference for a finished image).

My quick summary would be:

  • Use webP when you have transparency. Even the lossless format is much smaller than PNG and JPG doesn’t work with transparency.
  • Use JPG if you don’t care about optimizing file size and bandwidth or don’t want to have to think about edge cases (with high-frequency color or recipients who may not be able to handle webP). This is the simplest solution for uploading to social media or emailing images to people with an unknown ability to handle webP.
  • Use webP when you want to optimize for fast transfer sizes and minimal storage, except potentially for a few rare cases where high-frequency color may be impacted (such as cityscapes at night). This is most beneficial when you’re hosting the images on your own website or trying to keep your email sizes down.


Exporting “webP” with Web Sharp Pro v3.7

Web Sharp Pro makes it quick and easy to create gorgeous exports of your images for sharing. It offers much more than just sharpening, with custom templates, blurred borders, custom watermarks, advanced crops, film grain, batch processing, and much more. The best way to understand it is with the short demonstration below.

Web Sharp Pro v3.7 just added a range of new features, including support for webP. See this quick demo below on how to take advantage of it. If you need to export an image with transparency, this is automatically handled for you. There’s nothing to do (other than checking that you aren’t adding a border, since that wouldn’t be very useful if you want transparent edges.

When exporting from Web Sharp Pro as webP:

  • “ultimate” quality setting will export with the lossless format. This will completely avoid any JPG artifacts, though the final file size will be somewhat larger than a JPG (about double the size). And it will be much smaller than a PNG (about a quarter of the size), though it does not offer 16-bits. This is a great option when quality is your priority and you have areas of very small color detail or want to completely avoid artifacts.
  • “high” is designed to offer quality comparable to similar settings for JPG, but with somewhat smaller final file sizes. This is a great option when quality is your priority. Differences from ultimate are generally very hard to detect without zooming in well past normal viewing sizes.
  • “good” is also designed to offer size savings over a comparable quality JPG. This is a great option when small files are your priority (and the loss of quality at this level will be minimal, usually very hard to detect).


Exporting “webP” directly from Photoshop

If you don’t use Web Sharp Pro, you can simply use File / Save a Copy in Photoshop and select webP. I recommend the following settings:

  • Always leave the option to embed the profile checked in the first dialog (before the webP options)
  • Lossless if you need an exact copy of your image. Otherwise 75-90% lossy is a great range to consider. As you get closer to 50%, there is a noticeable loss of quality, but it may not matter for your application and this can really shrink the file size.
  • Embed XMP if you want to keep keywords and other such non-camera data.
  • Embed EXIF if you want to keep camera data (shutter speed, lens, etc).
  • Embed extras if you want to keep paths, guides, and print settings.
Greg Benz Photography