Three kinds of Smart Objects in Photoshop

I make mistakes in Photoshop, all the time. I’m also improving all the time, and sometimes what used to look great is no longer my best work. I’m guessing you probably do too. All of these reasons to make a change to your image make “non-destructive” workflows true life savers. Layers, layer masks, and adjustment layers are all critical aspects of such a workflow, but they are incomplete. They won’t let you reprocess the RAW image, edit something you did before warping the image, change the filters you applied to the image or anything else that would directly change the pixels themselves. At least not without starting over. But “Smart Objects” can do all of that and more.

At the same time, Smart Objects can be a little tricky to understand and use sometimes. One of the reasons for that is that there are multiple different kinds of Smart Objects, and they all look the same at first glance. But now with the Basics panel in Lumenzia v9.1, you can easily tell which is which.

The three kinds of Smart Objects you are likely to use or run into are:

Camera RAW Smart Objects

This is perhaps one of the most incredible tools Adobe offers. RAW Smart Objects give you full access to the RAW data right inside Photoshop. You can go back and make changes anytime. This gives you the full power of LR/ACR’s ability to extract the most from your RAW image and combines it with Photoshop’s ability to make intricate selections and masks to reveal those changes locally. This is a powerful way to enable multi-processing of a single RAW file, improve the quality of your exposure blends, apply RAW corrections with a level of precision that is not possible with the radial/gradient/brush tools in LR/ACR, as well as make any adjustment in LR/ACR locally (HSL, color grading, tone curves, camera calibration, etc). If you aren’t using Camera RAW Smart Objects routinely, you probably aren’t getting the most out of Photoshop.

Regular Smart Objects

Any Smart Object that isn’t a RAW Smart Object. While the value of RAW Smart Objects is that you can change the RAW processing at anytime, these regular Smart Objects have more general benefits. Typically, you would use them when you want to work non-destructively when adding adding a filter (Gaussian blur, Nik Color Efex Pro, etc) or warping the image (transform, perspective warp, etc). You’d probably put multiple layers into one Smart Object, but there may be times when you just want to apply a non-destructive change to one layer.

Duplicate Smart Objects

This overlaps with the above categories, both RAW and regular Smart Objects can be either dependent or independent. A “dependent” Smart Object is 2 or more layers which contain the exact same contents. Not copies, but literally the same content. If you edit any one of them, then they will ALL update. You cannot change one without changing another. Why would you want to do that? Probably so that you could apply different filters to different parts of the same Smart Object, since you cannot create multiple Filter Masks on a single layer.

This isn’t something you are likely to use much, and possibly never, in photography. But it’s still important that you are aware of them, because you will almost certainly end up accidentally creating unwanted dependent Smart Objects at some point because the normal shortcut <ctrl/cmd>-J or Layer / New / Layer via Copy both create dependent Smart Objects. Whether you create them the right way or the wrong way, they will look exactly the same in Photoshop until you try to edit their inner contents (such as changing the sliders in a RAW Smart Object).

The way to get the preferred independent Smart Objects is by right-clicking and choosing “New Smart Object via Copy”. Or you can simply <shift>-click the “PreBlend” button in Lumenzia or use the “SmartObj” button in the Lumenzia Basics panel.

How to tell which kind of Smart Object you have?

With Lumenzia v9.1, just make the Smart Object active and then look at the color of the “RAW” button in the Basics panel. You’ll see one of the following:

  • “RAW” will be green if the active layer is an independent RAW Smart Object.
  • “RAW” will be yellow if the active layer is a dependent RAW Smart Object. This is generally unwanted and probably a sign that the active layer was not created correctly.
  • “RAW” will be red if the active layer is a regular Smart Object (not a RAW Smart Object, though it may still have one inside it).
    • This is normal if you intentionally put multiple layers into a Smart Object.
    • However, if your intention was to open the image as a RAW Smart Object to preserve the RAW data, this is probably a sign of trouble (ie, you may be seeing this because you opened the image and then converted it to a Smart Object, which does not preserve the RAW data).
  • The “RAW” button will not show any color if the active layer is not a Smart Object, there are multiple Smart Objects selected, or there are no layers selected.

If you don’t have Lumenzia, you can do the following tests:

  • If you double-click the Smart Object and it opens up like a new document, then it is a regular Smart Object.
  • If you double-click it and it opens the ACR interface, it is a RAW Smart Object.
  • Either one of these could be independent or dependent. The only way to know is to make a change and see what happens. A simple test is to make some dramatic change and close ACR or the newly opened tab (which saves the changes back to the parent document) and see if other layers changed or just the one. You can then undo this step and move forward now that you know what you are working with.

Learn more about Lumenzia v9 here.

Learn more about Smart Objects here.

How to make dramatically smaller layered files

Luminosity masks and general layer masks are incredible tools for creating beautiful images and using non-destructive workflows. But like any tool, they can also be over-used. That can lead to unnecessarily large TIF files and some limits to your flexibility to make changes later. There are a number of situations where other approaches can yield the same or similar results as layer/luminosity masks, and give you other benefits when they are suitable.

Specifically, I’m referring to using BlendIf instead of luminosity masks, vector masks instead of simple layer masks, and combining group masks to eliminate extra masks. This can save substantial file space, allow you to save and re-open files more quickly, and enable more flexible workflows for non-destructive editing.

Lumenzia includes built-in tools to not only help you get the most from you luminosity masks, but also use alternatives when they are suitable. In this tutorial, you’ll learn about several of them in depth.

When to use BlendIf instead of luminosity masks

BlendIf is effectively a form of luminosity masking. It is quite limited in comparison to luminosity masks in general, but it can do a very good job of replicating generic light and dark masks, as well as midtones (to a lesser degree). This applies when you would directly apply a luminosity mask to an image, not when painting through a selection to create the mask (which offers substantial local control that BlendIf cannot come close to replicating). So this means that BlendIf is almost never a good alternative for exposure blending, but there are several other luminosity masking uses that it can replace quite well.

Color grading, protected vignettes, and more basic dodging and burning are all good candidates for using BlendIf. You can frequently get results which are just as good, but with some added benefits: smaller files and a more non-destructive workflow. While luminosity masks are grayscale images which can increase the size of your file by as much as 33%, BlendIf has absolutely no impact on the file size. Zero. That means less disk space used, faster saves, and faster re-opening of the image later. And the non-destructive benefits are also very useful. Whereas a luminosity mask does not adapt to future changes, BlendIf does. So if you use BlendIf to target highlights for color grading, you could later retouch the underlying image without having to worry about updating a layer mask.

Using BlendIf with Lumenzia is simple, but here are a few tips to get the most out of it:

  • Use BlendIf from the start where possible. Trying to convert a luminosity mask to a BlendIf later likely will require some other tweaks to maintain the same look. It’s much easier to do it once.
  • The fastest way to use BlendIf with Lumenzia is to <shift>-click the mask you wish to use, such as “L” or “L2”. Once you have created it, use the blue sliders in Lumenzia to customize (if the sliders are grey, you probably have an active layer mask on the layer, just deselect the mask or delete it so that the sliders are blue, which indicates that the BlendIf is the target for the sliders).
  • You can also create more advanced BlendIf by switching Lumenzia to the If:under or If:this modes via the dropdown at top-left. The “under” mode is the best choice most of the time (and is the only one you should use when working on an adjustment layer).
    • You can target color by clicking the swatches at the top to target the red, green, and blue channels (the others are combinations: yellow is really red and green).
    • The “not” BlendIf masks can be incredibly helpful for targeting color. Remember that targeting a color channel is not the same thing as targeting color. The highlights of the red channel include red, purple, yellow, and white. If you want to just target red colors, you should actually use “not L” green and “not L” blue (because red colors do not have significant amounts of green or blue). You could then target red highlights if you want to limit your reds to the brightest rests.

Do not use BlendIf for exposure blending, anytime you need to paint through a selection (most dodging and burning), when you need to customize the mask for precision, or any other time BlendIf yields inferior results. In general, you should be using luminosity selections to create luminosity masks for nearly all advanced work, and BlendIf is not a good substitute for any local work like that.

When to combine luminosity masks

Group masks (putting a masked layer into a group with its own separate mask) are a great way to help non-destructively reveal only portions of a luminosity mask without altering the luminosity mask in a permanent way. They are extremely useful for getting the perfect mask. Sometimes you need to keep them to be able to make refinements later, and sometimes you know you won’t need that capability any longer. If you don’t, then the extra mask is just consuming disk space unnecessarily. Lumenzia’s “Combine” button allows you to easily combine the grouped luminosity masks to save space.

To use it, just make the group layer active and click “Combine”. Lumenzia will do all the work for you and create a result which is identical. This will work even if you have several layers inside the group, which will save you additional file space. This is also a helpful tool to better understand how the grouped mask really affects the image, as you will now be able to see the exact mask being used.

Do not use “combine” when you will likely need to revise the group mask later. The flexibility is well worth the extra file size when needed.

When to use a vector mask instead of a layer mask

Unlike BlendIf, vector masks do not support any luminosity targeting. They should never be used to replace luminosity masks. But they are excellent replacements for simple masks created from lasso or marquee selections. This includes such masks which are subsequently feathered. Just like BlendIf, vector masks take up absolutely no space and can therefore save substantial space compared to a layer mask. If you are saving an uncompressed file, these potential gains are just as great. If you are saving compressed files, these simple layer masks do compress much better and the savings won’t be as great – but there is still much to be gained.

In addition to saving space, vector masks are paths which can be easily revised in ways that layer masks cannot. So if you are comfortable with paths, the pen tool, or direct selection tool – you can easily make non-destructive changes to vector masks.

Vector masks can also reduce clutter, as you can place both a vector mask and layer mask on the same layer. So instead of needing a group mask, you may be able to do everything on one layer (with a layer mask or BlendIf to target by luminosity and a vector mask to localize to a general area of the image).

Vector mask support is built into Lumenzia and here are a few tips to get the most out of it:

  • Anytime you are creating a mask from a lasso/marquee selection or using a vignette, consider creating a vector mask. These are all excellent times to use them.
  • When you use any of the buttons to create a mask in in Lumenzia (such as “Mask” or “Vignette”) and have an active selection or path, Lumenzia will ask if you would like to create a layer mask or vector mask. Just choose vector. If you previously choose layer and to remember that choice permanently, you can instead <shift>-click Mask to see all the options again or go to the menu (three bars icon at top-right) and reset popup notifications to be prompted again.
  • Lumenzia will feather vector masks by default. You can always change this later via the slider in Lumenzia. Make sure the vector mask is targeted (has white corners around it), as the slider will target a layer mask or BlendIf if they are on the same layer and the vector mask is not active.
  • To refine a vector mask’s shape, use Photoshop’s “direct selection” tool. This is the white arrow, which you can activate by clicking <shift>-<A> to toggle between the selection tools. **
  • Never try to create a vector mask from an active luminosity selection. This will not create useful results and can take a while to process while Lumenzia tries to make sense of the complexity of the luminosity selection.

Do not use vector masks instead of luminosity masks or when you need custom brushing, this is just a replacement for simple lasso/marquee selections.

** Note: Photoshop treats feathering of layer masks and vector masks fundamentally differently at the edges of the image. As a result, a feathered vector mask (unlike the otherwise identical layer mask) will cease to affect the edges, which would create problems for vignetting (as the edges would suddenly get light again instead of showing the expected result). The fix is to move edge points further outside the image canvas, and Lumenzia will do this for you AUTOMATICALLY when you create a vignette. There is nothing you need to do, but in case you wish to revise your vignette vector mask (path) later, you should just be aware that this is done on purpose to ensure a proper vignette so that you can refine it properly. The reason for this is that vector masks can extend beyond the edges of the visible image, while layer masks cannot – so Photoshop treats them differently at the edges.

 

Making smaller layer masks in compressed files

While I did not demonstrate this in the video above, there is yet another way to save on file size. When you save your image as a compressed TIF (or PSD / PSB, which are compressed by default), the amount of detail in the layer mask matters. So painting a solid black or white color on unused portions of the mask will help your image compress to a smaller size. This is a great option when you might be tempted to use “Combine”, but still want to retain some flexibility. In this case, you just remove areas of detail you would definitely not use (rather than removing everything you aren’t using right now).

There are a couple of quick and easy ways to do this:

  • In Lumenzia, just draw a rough lasso selection around the area you would like to keep, make your masked layer active, click the * button at the top of the panel, and then click “Mask”. The * button will intersect the selection and mask, which means it will only keep what’s inside the lasso selection. It is best to not feather the selection here when asked.
  • Alternatively, you can just manually brush on the layer mask. Be sure to paint all the way to black or white. Leaving a very dark or very bright area with detail will still consume a lot of space. I would recommend using the first method with Lumenzia for simplicity and guaranteed results.

Luminosity masking is easier than ever with Lumenzia v9

Version 9 of the Lumenzia luminosity masking panel for Photoshop is now available as a free upgrade for all customers. It’s never been easier to create the perfect luminosity mask or selection to make beautiful photos. And the updates have been designed to avoid changing any existing workflows, so you can jump right in. Be sure to see the highlights and initial demo videos below (more to come in the months ahead).

 

Hear me discuss new v9 features with Jeff Harmon on the Master Photography podcast.

New features in Lumenzia v9 include:

  • Automatically create subtracted selections just by double-clicking the “-“ button. This helps quickly and easily get better shadow detail in your photos.
  • Create special selections, channels, or layers based on HSB, HSL, LAB, CMYK, or RGB values. This includes a preview interface for trying different blend modes when creating layers.
  • Out of Gamut masks. Take more control over your conversion to sRGB, prints, or other strong colors that may need adjustment.
  • Reload the last orange preview layers. This includes any customizations, such as tweaks to the sliders, levels layers, color targeting, etc.
  • Automatically crop or fill transparent edges when aligning exposures via PreBlend to ensure great-looking edges in your blends.
  • Fade slider in the Basics panel. Quickly and easily dial in just the right amount of the last brush stroke, etc. Click the slider to set the exact amount you need.*
  • “Sky” button offers enhance Sky selection and replacement (Basics panel):*
    • Control whether the results are based on the active layer or sample from all layers.
    • Choose on-off sky replacements without having to import them to the PS library.
    • Select foregrounds
    • Guidance to avoid common issues (with hidden layers, adjustment layers, etc).
  • “Vignette” results have been substantially improved.
    • <shift>-clicking for BlendIf now automatically helps protect shadows.
    • When using vector masks, edges are now expanded so that the vignette is darkened all the way to the edges (to get the same as when using layer masks, but with more flexibility and smaller file sizes).
  • Enhanced “SmartObj” in the Basics panel:
    • Improved organization and clarity with named smart objects.
    • Copy channels to the new Smart Object.
  • Improved optimization, including customized menus, toolbar, and workspace. Available via “Optimize Photoshop” in the Utilities menu (top-right of the panel or ctrl-click the “Tutorials” button in CS6).*
  • And much more. These are just some of the highlights. There are 170 new features, updates, and bug fixes compared to v8.5.1. See the release notes for full details.

* Note for CS6 users: While Lumenzia v9 includes hundreds of enhancements for CS6, some updates are limited to Photoshop CC or work somewhat differently in CS6. This is unfortunately due to limitations of the >8-year old CS6 platform. Differences are primarily in appearance or workflow. Functional capabilities generally remain nearly the same. And of course, you automatically have access to the CC panel if you ever update to Photoshop CC.

Buy Lumenzia v9 now.

Existing customers can download any time via the links on this page (which is also linked from the bottom of all my newsletters).

Note: Lumenzia v9 is compatible with Photoshop CS6 and CC (including PS 2021), Windows, and MacOS (including Big Sur and M1 / Apple Silicon).

How to Optimize Color Settings in Photoshop

I’m not going to sugar coat it, color management is intimidating. I read 10+ books on the inner workings of ICC profiles last year and I still wouldn’t claim to fully get it. But if you want your images to look great and consistent on your monitor, the computer you buy next year, printed on your wall, and anywhere you show it on the internet – it’s something you need to get right. One of the foundations for that in Photoshop is the “Color Settings” dialog, and in this tutorial you’ll learn what the various settings are and how to set them to help ensure your photos look the way you intend.

Hear Jeff Harmon and I discuss this tutorial and the basics of color management on the Master Photography podcast.

What are working spaces and profiles?

First, a quick background on ICC profiles. The pixels in your images are saved as red, green, and blue (RGB) values from 0-255. But 255 what? Even your computer does not know. It would be like trying to bake a cake with half a recipe that called for 2 sugars. 2 cubes? 2 teaspoons? 2 cups? ICC profiles are meant to help resolve that ambiguity so that a pixel with an RGB value of 255, 130, 194 would look the same on your computer as it would on someone else’s phone when they see your image on Facebook.

We could spend weeks discussing how ICC profiles work, how to make them, and so on. But there are really only a handful of critical things every photographer needs to know:

  • For color management to work, you need an ICC profile embedded in your image. This ensures that your file accurately describes the color in the image. This is typically sRGB, Adobe RGB, or ProPhoto RGB. However, there are other good general purpose “working spaces” (which means they are a standard, but not based on any specific device). If it is missing (“untagged”), the computer has to guess (we’ll get to this below). When working with an untagged image, Photoshop will add a “#” indicator next to the bit depth in the file name (unless you have set the color management policy below to “off”).
  • You should open your image the first time with your preferred color profile and then avoid converting the profile in your working file. Any conversion may cause a loss of quality (and converting from a small gamut to a large gamut won’t bring back lost color). There are only a few cases where I convert profiles. The first is when outputting an image to the web or for print. I always do this on a duplicate copy, never my original working file. The other (which is rare and I try to avoid) is when combining images that were created with different RGB profiles, as all layers in an image must ultimately use the same profile (more on pasting with different profiles below). You can get deep in the details to consider all the options, but Adobe RGB is a safe and very good choice to use from start to finish in your working files. The settings below will help you avoid unwanted conversions.
  • You also need an ICC profile for your monitor or printer. This ensures that your output device accurately displays the color described in the file. You’ll need a profiling device such as the X-Rite i1Studio to do this. This is a very deep topic that we won’t cover here, but it needs to be done in addition to the other color management choices discussed here. Using the analogy above, skipping this would be like knowing the recipe calls for 2 teaspons of sugar, but you just use some random spoon in the kitchen to get close enough.

The color settings dialog (Edit / Color Settings) in Photoshop refers to ICC profiles in a few different ways:

  • The “color management” section describes what to do about “embedded profiles”, which refers to saving the ICC profile with your document. Without an embedded profile, you are back to the “2 sugars” scenario above and Photoshop will just guess what to do when it opens the file, which will likely lead to some disastrous results. Photographs should always include an embedded profile.
  • “Working spaces” are the ICC profiles that Photoshop will assume when there is ambiguity, such as not having a profile saved with the image.
  • “Conversion options” tells Photoshop how to manage conversion from one profile to another. While you should avoid converting the profile in your working document, you will likely need to convert copies of it for output to the print or web, so these are important.
  • The “advanced controls” section includes a bunch of random options that you should leave alone, but we’ll cover them below as well.

 

Optimal “Color Settings” values in Photoshop

If you are using Lumenzia, just use the “optimize” utility and it will take care of most of these settings for you automatically. Just go to the flyout menu (top-right three bars icon) and click on Utilities / Optimize. (Note: CS6 users should <ctrl/cmd>-click the Tutorials button to get to the utilities menu).

The settings dropdown at top allows you to choose from some standard defaults. Leave this alone to set things as recommended below.

Working spaces:

  • RGB Working Space:
    • This setting is only used when there is ambiguity. If you open your image in Adobe RGB, there is no ambiguity and this working space setting is ignored. So assuming you are using embedded profiles, there are only a couple of places where this setting matters.
    • One is when opening untagged images (which is common when working with images that came from a scanner or the internet). If you check the option below to warn when opening images with missing profiles, you won’t have any problems and will be prompted with options when you open the image.
    • The other is if you use the Image / Mode menu to convert between RGB and LAB, because this menu option does not specify which RGB profile to use. I recommend you never use that and instead use Edit / Convert to Profile so that you can choose your preferred RGB profile.
    • If you use LAB a lot and have a habit of using the Image / Mode command, you may wish to set the working RGB to match the embedded profile you use when opening your images (ie probably Adobe RGB or ProPhoto). This would avoid accidental loss of color that could occur by converting to sRGB. This is also safe with untagged images if you enable the warning for missing profiles as recommended below.
    • If you use a scanner which does not embed profiles and you have created one, you may wish to set it as your working profile to make it easy to assign it to these untagged images as you open them.
    • Otherwise, setting this to sRGB is probably ideal. That will give you faster access to choose it when opening an untagged image (rather than scrolling through a list if it wasn’t the last one you used). And if you work with a lot of untagged images which you assume will always be in sRGB, you could then turn off the missing profiles warning below to have these images automatically treated as sRGB and avoid getting prompted over and over.
    • Do not set this to “Monitor RGB” (this will disable color management). You should also not set it to any custom profile you have created, these are not good choices for your working files. If you need to use other spaces for output, duplicate your file and use Edit / Convert to Profile to convert that one-off file.
  • CMYK working space: Unless you work on files in the CMYK workspace, this probably has no effect on your work at all. Regardless, leave it at the default unless you have a good reason to change it.
  • Gray working space:
    • As I described in a previous tutorial, this setting can significantly affect the quality of luminosity masks. However, there is no good general setting that you can just set and forget. It should be matched based upon the active (embedded) RGB profile in each document. Lumenzia automatically optimizes this for you on the fly for each image. If you are not using Lumenzia, see that older tutorial for suggestions on the best alternative approach.
    • Beyond luminosity masks, I strongly recommend you do never use the grayscale working space for photography. The only benefit is smaller TIF files. RGB mode is equally capable of producing the same black and white images. More importantly, there are numerous tools and filters which are only available in RGB mode (including non-destructive options to control the conversion from color to black and white). Additionally, great black and white images often have a slight color tint added to them.
  • Spot working space: It is very unlikely that you work on files in the spot workspace. Leave it at the default unless you have a good reason to change it.

Color Management Policies:

  • RGB, CMYK, and Gray should all be set to “preserve embedded profiles“.  Using “off” creates those disastrous scenarios covered above.
    • Specifically, it causes new documents to have no profile and strips the profile when opening a file that has a profile different from the working space (it will leave an embedded profile alone if it happens to match the working space).
    • And we should avoid conversions in general, so the third option to convert to the working space should also be avoided.
    • Note that if you are unable to to change this value, make sure your RGB working space is not set to “Monitor RGB” as this automatically forces the RGB policy to off.
  • Profile Mismatches / Ask When Opening” should be left unchecked. This will warn you when opening an image with an embedded profile which is not the same as the working RGB. Since we wish to avoid conversions, just stick with the embedded profile.
  • Profile Mismatches / Ask When Pasting” should probably be left unchecked. This serves as a helpful reminder that you haven’t been consistently using the same profile, but the right answer is almost always to convert and that’s what will be done if this is left unchecked.
  • Missing Profiles / Ask When Opening” should be checked. Missing profiles are a serious issue and checking this box will both warn you and give you a chance to fix the problem.

Conversion Options:

  • Engine should be left as Adobe (ACE). This is an excellent choice and consistent between Mac and PC.
  • Intent should be set to “relative colorimetric“. This is most often the best choice, and you can use “perceptual” as needed by using Edit / Convert to Profile to control the process when you need that instead. (Note: photographers should generally not use absolute colorimetric other than for some advanced hard proofing scenarios, and I cannot think of a good reason to use saturation for photography).
  • Use Black Point Compensation” should be left checked for best results (to avoid light/muddy shadows).
  • “Use Dither” should probably be checked. This adds a slight bit of noise when converting to 8-bits to help disguise any possible banding. You can also control this on the fly by using Edit / Convert to Profile if you need to make a different choice once in a while.
  • “Compensate for Scene-referred Profiles” should be left checked. It is intended for those using Photoshop as part of their video work, so it probably does affect you.

Advanced Controls:

  • Desaturate monitor colors” should be left unchecked as it is deliberately causing your monitor to deviate from an accurate profile. The potential benefit here is to help visualize colors so strong that they are outside the gamut of your monitor. This is not what I would consider a precise nor highly useful tool.
  • Blend RGB Colors Using Gamma” should be left unchecked. Setting this to 1 actually creates better blends of color. For example, paint red at 50% opacity over solid green. The default behavior will be dark where red and green mix, whereas a gamma of 1 will show the preferred yellow. While I would like to use this more “colorimetricly” correct approach, it only makes sense to do so as a setting in the document, not across Photoshop in general. If you change this behavior, then you are likely to see some potentially significant changes in the appearance of your layered documents. And if you share files with others or do not consistently use this setting going forward, your images may not appear as intended. In the end, this doesn’t provide a lot of value for the kinds of colors we actually mix in photography so leaving it off is fine and preferable since changing it may cause unexpected changes.
  • Blend Text Colors Using Gamma” should be left checked and set to the default 1.45. The idea here is similar to the previous setting, but just affects text. If you change it, you’ll likely see changes in the edge detail of your text layers. The default is fine, and the concerns for unexpected changes are the same.

 

 

 

The best software for beautiful large prints?

Note: Topaz is running a sale through Jan 1, 2021. Use discount code gbenz15 to stack discounts for a total savings of 40% off Gigapixel‘s regular $99 price.
I print a lot of images both for myself and clients. Most of the time, these are large prints which require enlargement. For example, I frequently print 40×60″ prints from my Nikon D850. Without interpolation, those 46-megapixel RAW files would only be 137 dpi at that size. My goal is 300dpi, which would be a 218% enlargement of the linear dimensions or equivalent to shooting on a 172-megapixel camera. Safe to say I won’t own one of those anytime soon. And that’s not the largest size I might print. Unless you only make small or medium-sized prints, it’s important to know how to enlarge your images.
In this tutorial, you’ll learn the pros and cons of the most common options and how to get the best results from each. I use a variety of tools, including Photoshop’s “Preserve Details“, Topaz Gigapixel AI, and ON1 Resize. There are numerous alternatives, but I recommend Gigapixel. (Note that if you decide to purchase Gigapixel, use discount code gbenz15 for 15% off.*)

Note that I also generally recommend ON1 Resize for those printing on canvas, as it has an excellent tool for created reflected edges for gallery-wraps. However, in my experience with the latest 2021 version, there are some problematic bugs and so I skipped it for this video and recommend holding off purchasing or upgrading for now. ON1 is a great company and I’m sure they will address these bugs. I have reached out to them for support and will update this article when they have a release which addresses these issues.

I use all of these options because each has its own strengths. A direct comparison is difficult to do, for a few reasons. First, the controls are different and you really have to spend time to find optimal results to compare from Resize and Gigapixel. And second, all enlargement interacts with sharpening and noise. For example, some of these algorithms may add a lot of sharpening, others very little. The immediate result isn’t a fair comparison until you have finished the image. The best algorithm is really the one that gets you to the best final print, including consideration of how you process noise and sharpness in the image both before and after using one of these enlargement tools. With those caveats out of the way, I think there are some clear differences between the various options.
So before we get to the comparison further below, let’s take a look at how to get best results from each of them. Note that for each of them, make sure you start from a flattened copy of your work (just duplicate the image and flatten all layers). Resizing layers and layer masks is possible, but may easily lead to poor results. And the other tools may throw some cryptic error (especially with Smart Objects and Adjustment layers) about “command not available” or “GatherImageDataFilter”.

How to get great enlargements with Photoshop “Preserve Details”:

The options here are limited and pretty straight-forward.
  1. Go to Image / Image Size.
  2. Make sure “resample” is checked and the dropdown next to it is set to “Preserve Details (enlargement)“. While there is a “2.0” option, the original is generally best.
  3. Enter the desired output size.
  4. The “reduce noise” slider is helpful when working with noisy images, such as ISO 6400 images of the night sky. You should leave reduce noise to 0 for images shot at low ISO, but consider something like 25% for noisy images.

How to get great enlargements with Gigapixel AI:

  1. Go to File / Automate / Topaz Gigapixel AI to start the plugin from within Photoshop.
  2. Choose your preferred settings
    • You can choose size by scaling factor, but picking a height or width is typically simplest. Click the dimension units to switch between pixels, inches, or cm and choose the desired resolution (300 pixels per inch is a good choice if you don’t know what to use). The full output dimensions are shown below the image.
    • Choose Standard for most images or Architectural for cityscapes or other highly detailed images with sharp edges. The Compressed option is meant for when working with low quality source files and Art is for graphics, so you probably won’t use either of them.
    • Clicking “auto” is a very good option. While you can often do a little better in most cases, the difference is unlikely to be noticeable in most prints. But the greater the enlargement, the greater the benefit to spending time on the next few options.
    • “Suppress noise” helps improve detail significantly in high ISO images. For most (clean) images, just leave this at 0-20%. For high ISO images (such as ISO 6400), try up to 80 or even 100%.
    • “Remove blur” is a confusing way of basically saying: sharpen the image image. It can be used for handheld images or other scenarios where there is some subtle blurring of your subject, but it can also simply be used to add sharpness to any image. This slider is a significant part of the apparent sharpness of Gigapixel at default settings and the best choice here depends on your image, personal preference, and whether / how you plan to sharpen the image separately after enlarging it. 50% is a good general-purpose setting here, but this is the one setting where it really pays off to try different values and to fine the best preview.
    • “Reduce color bleed” isn’t quite what you would expect. I find that it adds a bit of detail, with a risk of some artifact. Toggle it on and off if you’d like to see if the extra detail is helpful in key areas. I generally leave it off.
    • Try “face refinement” if you are enlarging and image with people.
    • In my testing, none of the options other than size affect the processing time, so just pick what looks best in the preview.
  3. Click “Apply” when done.
The final output includes both the result from Gigapixel AI, as well as another version which is identical to Photoshop’s “bilinear” resampling method. In most cases, you can probably just delete the lower layer, but it is helpful if you wish to blend the results using lower opacity on the top layer or paint black on a layer mask.
Note that I have found the time to process the image is not affected by any of the options, so you should just choose the settings that look best in the preview. If you need to save time, you should use one of the other approaches for enlargement.

How to get great enlargements with ON1 Resize:

The interface is a bit complicated if you want to dive into all the options. I recommend just using a few to keep things simple. Here’s the workflow I recomend:
  1. Go to File / Automate / ON1 Resize to start the plugin from within Photoshop.
  2. Under Document Size: Set the final dimensions and resolution you need.
  3. Settings is the key set of controls for enlargement. Choose the best preset under “image type”. Leave the method on “Geniune Fractals” and the sliders as set by the preset. You can optimize texture, threshold, and smoothness; but I don’t think it’s worth it for most users. The controls are confusing and there isn’t really a live preview. The defaults are fine. If you want to tweak details for best results, I recommend using Gigapixel instead.
  4. Skip “Sharpening” and use options in Photoshop instead.
  5. Skip “Film grain” for most use, though it’s worth trying various options for portrait work.
  6. Use the “Gallery wrap” for mirrored edges if printing on canvas. Set the type to “reflect“, thickness for the depth of your canvas frame, and leave the opacity at 0 to avoid adding an overlay color to the edges.

Photoshop vs Gigapixel AI vs ON1 Resize

Here’s a quick rundown of the pros and cons:

Photoshop (Preserve Details 1.0):

  • Lightning quick. In my baseline test with an images from a Nikon D850, Preserve Details 1.0 took 6-8 seconds. By comparison, ON1 Resize took 50-120 seconds (roughly 8-15x longer) and Gigapixel AI took 18 – 42 minutes (up to 400x longer).
  • Batch processing. If you want to use actions to process multiple images, Photoshop’s built-in method is the only one that can be fully automated. You can specify the outputs and use an interactive mode. While Resize and Gigapixel have plugins which can be launched via actions, you can’t request specific values to use and will have to manually interact with each image.
  • Good-enough quality. While it comes in third place here, it still provide very good and usable results. For less demanding scenarios such as modest enlargements or images that will be viewed from a distance, the quality benefit of the other options isn’t worth the time.
  • No additional cost. However, if you print more than very occasionally, I wouldn’t say cost savings is a benefit here, because I think it’s worth buying either or both of the other options for at least some printing.
  • Best for: Small enlargements, tight budgets, and saving time.

 

Gigapixel AI:

  • Excellent quality! Gigapixel AI flat out wins on quality for most, if not all, images I have tried.
  • However, that quality requires patience. Be prepared to wait a long time for each image. And the heavy CPU use will quickly crush a laptop battery, so you’ll probably want to do this work while you’re plugged into the wall.
  • Best for: Highest quality.

Note: Topaz offers a free trial, and you can save 15% off when purchasing Gigapixel by using discount code gbenz15. You can also save even more by purchasing it as part of the Topaz Suite with all their major products.

 

ON1 Resize:

  • Canvas: Where Resize really shines is for printing on canvas. It has a built-in option to create gallery-wrapped edges. And with the lower resolution of canvas media, the quality benefits of Gigapixel are much less pronounced and won’t be something you’d see in the finished canvas in many cases.
  • Great quality. I would say that Resize is definitely better than Photoshop’s Preserve Details, but Gigapixel is clearly the best for large enlargements on high-quality media.
  • Best for: Printing on canvas.

 

And while I don’t use it myself, I do want to at least mention one other very good free option…

RAW Therapee:

  • Great quality for free. RAW Therapee is designed as a RAW processing program, but is free and happens to include an enlargement algorithm which produces better results than Photoshop. I don’t personally think it is as good as Gigapixel, but it’s definitely worth a look if cost is a concern.
  • The major downside here is ease of use. There is no Photoshop plugin and the interface is much more complicated than Resize or Gigapixel. Just choosing new dimensions and saving is a confusing task.
  • To resize: open the image, go to the transform tab (looks like scissors and a triangular ruler), check resize and allow upscaling, enter the desired dimension (leave method on Lanczos), then click the save option towards the bottom-left of your image.
  • If you need to address noise, you can do that in RAW Therapee as well by looking under the Detail tab (icon just right of the +/- exposure icon) and then clicking the dot by Noise Reduction to turn it on and then adjusting the settings below. However, you will probably find it easier to do noise reduction with other software first, given the complexity of the options here.
  • Best for: Lots of control for free. This is probably best for those looking to use a completely open-source software workflow.

 

 

Enlargements are just one piece of the puzzle when making prints. If you’d like to see more tutorials on printing, please let me know in the comments below.

 

[* Disclosure: This post contains affiliate links. I rarely endorse other products and only do when I think you would thoroughly enjoy them. By purchasing through my on this post, you are helping to support the creation of my tutorials at no cost to you.]


See the store page for Lumenzia and course info.
"Lumenzia" and "Greg Benz Photography" are registered trademarks of Greg Benz Photography LLC.
See licensing for Commercial and Creative Commons (Non-Commercial, Attribution) Licensing terms.
Join my affiliate program.
See my ethics and privacy statement.