Our ability to print fantastic, high resolution imagery is ultimately dependent on the quality and type of the digital files that we receive or create. Occasionally (but fortunately rarely), customers want to use an image pulled from a web site that might be 200-300 pixels wide, and cannot understand why the printer will not deliver a clear, crisp image on a poster. Obviously, it has nothing to do with our printer's output, and everything to do with the original art file. As the saying goes - “garbage in, garbage out.”
There are two basic types of files- Raster (bitmaps) and Vector (outlines). For example, a photo is usually a Raster Image, and a font is usually a Vector or outline.
Vector file formats (scalable to any size UNLESS they contain bitmaps in addition to vector images)
Raster/bitmap file formats (may not support being scaled to large sizes)
Be careful of GIF files as these are most commonly used for websites and are usually not high enough resolution to be enlarged significantly.
There are a couple of additional considerations to keep in mind. While EPS, PS, and AI files commonly contain vector data, they can also include bitmap data as well. The opposite is not true. TIF’s, JPG’s, and GIF’s only carry raster data and do not carry any vector data. In addition, when working with JPG files, they use a compression scheme to reduce file size. (TIF’s may or may not use an optional file compression.) Resizing a JPG file compresses the image which means reduced quality. Bitmap files are best as uncompressed TIF’s. The files are large but there is no loss of quality when changes are saved.
The biggest difference between Raster and Vector file formats is that a vector image can be infinitely enlarged since the image is created from a series of mathematical formulas defining the paths of curves, lines, colors, and other components. On the other hand, the amount a Raster (bitmap) image can be enlarged depends on the original resolution, and acceptable loss of clarity relative to the typical viewing distance. Raster images are commonly used for photographic images, but they are also a bit more complicated to set up and produce a quality large format print.
A Raster image (bitmap) is made up of tiny dots, and each dot can be one of many thousands, if not millions of colors. There are also two color modes or pallets – CYMK (Cyan, Yellow, Magenta, and Black) or RGB (Red, Green, and Blue.) Generally, the RGB color space is larger, and many RIP (Raster Image Processor) programs like this format to produce the most vibrant colors. This might seem odd, as the printer is a CYMK device and eventually all files will be converted to CYMK by the RIP program when it outputs to the printer. Nonetheless, starting with RGB files and letting the RIP make this conversion will generally produce the most vibrant images.
Raster images also have limitations on how much it can be enlarged depending on the original file’s resolution. We recommend keeping the final output resolution between 100 and 200 DPI (dots per inch) at FULL SIZE. Below 100 DPI and you will start to see a sacrifice of quality. Over 200 DPI, you are probably wasting hard drive space and processing time. (Unless you are printing 4 point type and the print is to be viewed from a very close distance.) This file resolution is not to be confused with print resolution. Even though our input files are 100-200 DPI, we typically print at least 360 DPI and more often 720 DPI. This ensures the smallest droplet and lack of graininess.
One exception might be on a billboard or a tractor-trailer. These areas are so large, that a 200 DPI input file will be too large for most computers to deal with. When working in Photoshop, these files can swell over 1 Gigabyte.
Here is an example. When following the 100 DPI, full-size guideline; if you take a file that starts out at 600 pixels wide, the largest you can print, keeping to the 100 DPI rule is 6 inches. (Divide 600 pixels by 100 = 6”). Of course, it is possible to enlarge the file to a much larger size in Photoshop, but this just multiplies the pixels around each other, and you create a fuzzy mess. Photoshop will just take each pixel and create a new pixel next to it, halfway in color to the next, and so on. This is why most website images are not useable for large format output. There is just not enough pixel data. Website graphics are created at as small a resolution as possible to create a reasonable image to view on a monitor and reduce bandwidth. This makes the website load as quickly as possible.
A high-end camera in today’s market may be as high as 12-Megapixels or more. (1 Megapixel = 1 million pixels). A 12-Megapixel camera, for example, may have a resolution of 4288 pixels wide x 2848 pixels high. In keeping our file within our 100 DPI guideline, the maximum size we can output with minimal distortion is 42.88” wide (4288 pixels wide divided by 100). The larger the original image and the more pixel data you have, the more leeway you have to work with. With this much pixel data, we can go a reasonable amount larger than 42.88” and still achieve good results. Keep in mind, the larger we go from the original, the fuzzier and pixilated the final output will be.
The next consideration is how much fuzziness is acceptable. It depends on how the image is viewed. Just as you would not be viewing an 8 1/2” x 11” piece of paper from 40 feet, you would also not expect to view a tractor-trailer from 6”. Billboards would be another extreme and their resolution might be as low as 2 to 20 DPI. However, the billboard might be 40 feet in the air, and viewed from many hundreds of feet away, thus the resolution can be much, much less.
There are also filters that can do a nice job of smoothing out the “jaggies” of less than optimal files. One popular program is Photozoom, which we have available. Its proprietary algorithms filter less than optimal artwork to result in a more acceptable image than Photoshop when enlarging artwork, as long as there is sufficient pixel data there to start with. It cannot perform miracles, but it can make a significant improvement.