If you have only one exposure, your options are limited to upsampling or using software to create more pixels. While the information rendered by software is never as rich and sharp as information that’s optically captured, it nonetheless can be both pleasing and convincing. Upsampling is the best method for images containing moving objects, as other methods require multiple exposures and may produce motion artifacts.
Upsampling with Lightroom or Photoshop works well for most situations. Use the upsampling method of Bicubic Smoother (best for enlargement). It’s as simple as that. Upsampling in Lightroom can be performed manually during exporting or automatically during printing. If upsampling is substantial, I prefer doing it in Photoshop where follow-up sharpening can be better fine-tuned. In Photoshop, go to File > Image Size, and check Resample Image. Set your desired Pixel Dimensions or Document Size; they’re two different ways of asking Photoshop to do the same thing: resample an image. Pull down from Bicubic Automatic to Bicubic Smoother, which is best for enlargements. (You don’t actually have to do this step as Photoshop automatically chooses the best resampling method.) Click OK.
Much has been made of "stair-stepping" methods, where files are increased in size in small increments (sometimes 25%, sometimes 10%) multiple times. However, the increase in quality gained is minor, making this rarely worth the effort. The thing that makes the biggest difference in image quality is the sharpening applied to images after they have been upsampled. If you want to maximize quality, after upsampling, carefully sharpen images individually. Furthermore, if you’re upsampling an image dramatically, before upsampling turn off all prior sharpening and noise reduction, and after upsampling apply new sharpening and noise reduction. Upsampling routines will amplify noise reduction and sharpening artifacts, which may produce inferior or even objectionable final results.
If you’re performing extreme upsampling, consider onOne Software’s Perfect Resize. It can make a poster out of a postage stamp. Perfect Resize produces superior edge contrast that’s very useful when you’re aggressively upsampling. If you’re performing moderate upsampling, you won’t find the differences between Photoshop/Lightroom and Perfect Resize to be dramatic.
How far can you go? I could give you an overly simple answer: up to 300%. But the true answer is, it depends. Knowing what it depends on will help you choose a method, modify a routine and evaluate results to get optimal results for individual images. How far you can upsample an original depends on many factors found in the source, the destination and the statement you’re making.
When addressing the source, consider the frequency of detail represented and noise contained in the file. Noise is challenging to separate from detail. Strike an optimal balance between blurring to reduce noise and sharpening to increase signal. You can blur and upsample smooth low-frequency detail more aggressively than high-frequency detail. For images that contain both, limit how far you go based on the high-frequency detail in them. Don’t blur and/or upsample so much that the detail looks synthetically rendered; make it look optically captured.
When choosing a destination, consider the substrate. Hard-coated glossy surfaces in print or display will render finer detail, including artifacts, while soft, fibrous matte print surfaces will hide many artifacts, but not render detail as well.
No single solution will fit all visions. How you choose to process and present your images speaks volumes. While there are general parameters for what most people think is acceptable, or even well-crafted, the choices you make become an integral part of the statement you make. You can do anything and it will work if what you do effectively supports the statement you make. Statements make art; craft supports them.
Develop a critical eye for what to look for in images: contour, texture, noise and gradation. Then, using your best judgment, trust your eyes. If you create an effect that’s convincing to you, it’s quite likely to be convincing to others. In many cases, because of your expertise and insider knowledge, you may find that you’re more discerning and critical than others are.
You can increase the resolution of a file and improve the detail an image renders by making multiple exposures of the same composition and combining them into a single file. To do this, try PhotoAcute—super-resolution is its specialty, and using it’s as easy as 1-2-3.
1. Click Select/View images and load the images you want to combine. The benefits are marginal if you use fewer than four files; for best results, combine six or more files. Curiously, you’ll see slight benefits even if you use multiple copies of the same file instead of the preferred and recommended method of using multiple exposures.
2. Click Adjust processing options and check Increase resolution. PhotoAcute also offers options to improve noise and chromatic aberration, and increase dynamic range that are convenient. However, if the improvements that need to be made to these aspects of an image are substantial, there are other better ways to achieve these goals.
3. Click Process this group. How far can you go? A little more than you can with upsampling. Just as with upsampling, how far you can go depends on many factors, including source, destination and the statement being made. The overhead is high with this technique (merging six files takes a little time), but the results are good. Using this exposure and processing method isn’t something I would do with every image, but I would strongly consider it for images where resolving fine detail is particularly important and the technique is practical.
Another method for increasing the resolution of your image files is to break a scene into pieces with separate multiple exposures and then stitch them together using panoramic merge functions in today’s software. It’s a matter of simple addition—two files are better than one, three files are better than two, etc. With this method, detail is optically captured, though you also can choose to enhance it further with software. While you can consider dedicated panoramic software like Kolor Autopano Pro or PTgui for challenging images, it’s highly likely that Photoshop is all you’ll need.
Take these steps. In Bridge, select the files you wish to include in a photomerge and go to Tools > Photoshop > Photomerge. Alternately, in Ligh
troom, go to Photo > Edit In > Merge To Panorama In Photoshop. Either way, you’ll end up in Photoshop and encounter the Photomerge dialog. In most cases, you’ll want to use all three options it offers—Blend Images Together, Vignette Removal and Geometric Distortion Correction—but any of them can be unchecked before producing the final merged file. Photoshop will place separate exposures on separate layers, transform and align and mask them, then selectively adjust color to create seamless transitions between them.
When you make exposures for stitches, shoot a little loose, as it’s quite likely that the border of the final image produced will need to be cropped (or cloned) to be made rectangular. You also may encounter some spatial distortions in stitched images, either subtle or dramatic, depending on your exposure methods. These can be corrected, in part or in whole, after a merge with Photoshop’s powerful distortion tools: Free Transform, Puppet Warp and the filters Lens Correction, Adaptive Wide Angle and Liquify.
Moving objects can present challenges for stitches, and Photoshop will allow you to select their position based on a specific frame and automatically remove ghosts, but artifacts in fields of motion, such as water or clouds or foliage, are often inevitable, making this technique suitable for many, but not all, situations.
How far can you go? The lens is the limit. Theoretically, you can stitch an infinite number of images. The true limits lie in how much your lenses will allow you to zoom into a scene. Yet, the most important factor still remains: What’s practical in a given situation? Other questions arise. What’s better? Fewer exposures made with a higher-quality lens? Or, more exposures made with a lower-quality lens? The answer lies in how much the quality of one lens exceeds another. Compare manufacturers’ MTF charts (they’re readily available online) for useful objective data that will shed light on this.
So, when you want sharper, bigger digital images from your existing cameras, you have options. Remember, upsample, stack and stitch.
John Paul Caponigro, author of Adobe Photoshop Master Class and the DVD series R/Evolution, is an internationally renowned fine artist, an authority on digital printing, and a respected lecturer and workshop leader. Get his enews Insights with access to hundreds of lessons at www.johnpaulcaponigro.comand see more of his panorama tips at www.johnpaulcaponigro.com/downloads/technique/exposure.php.