How to Effectively Resize Images
As display and printing technology advances, the size requirements for high resolution images is getting higher as well. This is great, but what happens to those older pieces you created that were designed at smaller dimensions? If the work of art is traditional, this means you only need to take a new high resolution photo or scan of the piece, but if the work was created digitally it might not be so easy.
We wanted to share a few tips on resizing images, when possible, while retaining quality. For this tutorial we are going to use the latest version of Adobe Photoshop, however most current image editors will have similar settings.
The first thing you should be aware is the difference between resizing and resampling. When you resample an image, you are either adding new pixels or taking pixels away in order to achieve a new image size. Resizing an imagine without resampling, only changes the target print size and/or DPI and not the pixel dimensions. Take a look at the image below and you will see that with “Resample” unchecked, you cannot change the pixel size of the document. For this tutorial, we want this checked so we can take advantage of these resampling algorithms.
When resampling an image to a smaller size, pixels are removed to accommodate this new size. Due to this, tiny details can be lost. The image overall will become more pixelated and less ideal for printing at larger sizes. Normally, when you are resizing your images down this is so you can share them on the internet. If you are sending a print ready file to a printer, it’s best to send as large as a file as you can, or the exact pixel dimensions the printer has requested. The printer can always scale the image down using the software that is working in conjunction with their printing hardware. Otherwise, use “Bicubic Sharper” when resampling your images down. In this algorithm, Photoshop is paying more attention to pixels in close proximity, and will result in a sharper image.
Pro-tip: If you do need to scale images down, especially if the image is a layer in a larger composition, use smart objects. The smart object acts like a container, and while the image layer will be resized smaller, the smart object will still contain the original size file you added to the document. This means you can resize it back to the initial size it was as no pixels were removed, only the smart object was scaled down. This method also keeps your files non-destructive.
Resampling images up can be deceptive, as we often look at the newly upscaled image on our computer screens, and zoomed out so we can see the full image. We may think the image looks sharp and therefore ideal for printing, but without actually seeing the upscaled image printed, we can’t really know for sure. As stated above, when resampling images up Photoshop is adding new pixels to the document, and you can be sure it’s not being as critical as you were when creating the image.
For perspective, imagine having a document that is 1000 x 1000 pixels, and you resample it to 4000 x 4000 pixels. Photoshop is going to add 15 million new pixels to the document in order to achieve the new pixel dimensions you have chosen. That’s a lot of pixels, and what they are most likely going to do is make sharp edges blurry, degrade details, and add artifacts to your work. This may look ok on your screen, but once printed the image could very well be of low resolution.
So, in order to keep your fans wowed at the quality of your work, know when it’s possible to even scale an image up and how far you can go with it. As we touched on in our post on Image Resolution and Printing, you can scale an image up to billboard sizes and it will still look good, provided you look at it from the proper distance. Think about the purpose for resampling your image, and make your choice based on that.
When resampling images up, here are your choices:
Preserve Details: With this method the reduce noise slider is made active so you can smooth out noise as you upscale.
Bicubic Smoother: This algorithm is good for enlargements, but designed for smoother results.
Bicubic: This method is slower and produces smoother gradations than Bilinear and Nearest Neighbor.
Nearest Neighbor: This fast method reproduces the pixels in the image and preserves hard edges. It can create some jagged effects.
Bilinear: This algorithm adds in new pixels by averaging the color values of surrounding pixels. The results are middle ground at best.
Play around with each of these methods, and see in the preview panel what each of these algorithms will do to the details of of your design. While each of these methods make it easy to resample images up, there are often occasions when the image can simply not be scaled up to the size you need without a significant loss in quality.
- Check out the Alien Skin Plugin, Blowup. It can resample images up but the end result often looks as though some anisotropic diffusion has been added.
- Also, check out the Resize 10 plugin. a fractal based interpolation algorithm.
- If you want to read more about how you can push scaling up to new levels, read up on Stair Interpolation. In this method, the image is scaled up in 10% intervals. Due to this, the pixels that are added are done so in increments as opposed to one lump addition. The results are often very close to what the top two plugins can achieve.
- Know when to let an image stay at the approximate pixel dimensions it was created in. It’s worth it for you and your fans to not release an upscaled image that might look good on a monitor, but blurry and low quality in print.
Try these resizing methods out and share any tips you have in the comments below.