The end goal for many photographers is the print. For many photographers, the print is the most satisfying way of experiencing an image. There is just something very gratifying about seeing one of your images as a beautiful print that everyone can enjoy.
Not only do many photographers enjoy making prints, they are making larger prints than ever before. When I first started shooting digitally, I made 8"x10" prints. I remember when I got my first printer that made 13"x19" prints. I thought that the prints were huge. Now, I make 24"x36" prints. Panorama prints can even be larger.
There can be no denying it. Large prints can be more effective in grabbing a viewer's attention and communicating the photographer's intent to the viewer than small prints. On the other hand, large prints also make weaknesses in an image much more obvious. Things such as a lack of sharpness, a poor mask, poor resolution, inferior editing technique, and a lack of printing skill that might go unnoticed in a small print will dramatically stand out in a large print.
Thus, as prints become larger, ones workflow and editing technique must dramatically improve. With that in mind, the purpose of this article is to cover those aspects of workflow that will maximize the quality of prints. It should be kept in mind that these methods often require more time than simpler methods. Thus, these methods are not necessarily for all prints. Rather, they are designed for those times when a very high quality print is desired. For instance, these methods would be appropriate for a large fine art print that is intended for display. They would be overkill for a 5"x7" print that will end up in the family photo album.
When a chef wants to create a gourmet meal, he starts with the best ingredients. It would be foolish for a chef to attempt to create an epicurean delight by starting with inferior ingredients. Similarly, any photographer that wishes to create the highest quality prints possible needs to start with the best data. That means that the image should be shot in raw.
Shooting raw has many advantages over shooting JPGs. However, one of the biggest advantages has to do with the number of bits in an image. JPG files are eight bit. That means that there are eight bits of information for each of the three color channels (i.e., red, green, and blue). Those eight bits can produce 256 shades for each of the three color channels. On the other hand, raw files contain more bits. Currently, most of the DSLRs being released are fourteen bit. Thus, there are fourteen bits of information for each of the three color channels in a raw file. Once the image is converted, that works out to 16,384 shades for each of the channels. So, what does this mean in practical terms? Actually, there are several ramifications.
First, in some cases, color transitions are noticeably smoother. Since there are a lot more shades of each color with images that start as raw, adjacent colors are closer together than with JPG images. Thus, with raw, the transitions from one color to another are smoother because it entails many diminutive steps instead of a smaller number of larger steps as is the case with JPGs. Now, notice that this paragraph starts off by stating, "in some cases". Whether or not the differences in color transitions between raw and JPG are noticeable or not depend on a number of factors such as the tonal region (e.g., quarter tones vs. three quarter tones), amount of editing, and size of print).
Second, there will be more shadow detail with images that start as raw files. This has to do with the fact that camera sensors are linear. Now, getting into a technical discussion of sensor linearity is beyond the scope of this article. However, in short, sensor linearity means that the sensor places most of the bits in the lighter shades and far fewer bits in the shadows. Thus, the shadows simply do not have nearly as many bits, and therefore shades, as the highlights. This is particularly a problem for JPG images. Since JPG images do not have a lot of bits to begin with, and more of the bits are allocated to the lighter shades, there are relatively few bits available for the shadows. Thus, there are few shades in the shadows, and detail suffers. With raw, the situation is significantly different. While raw files also have fewer bits in the shadows, there are so many bits that the shadow detail doesn't suffer as much. In general, a file that starts as a fourteen bit raw file will have sixty-four times as many shades in the shadows than an eight bit JPG file.
Third, images that start as raw files hold up to editing better than JPG files. When files are edited, some shades are spread farther apart. Other shades are lost due to quantization error. What this means is that the distance between one shade and the next can become larger due to the editing. If the distance between the shades becomes large enough, the human eye will be able to see the difference between adjacent shades. When this happens, posterization results. This is often seen as banding across areas of little detail (such as a featureless sky). Since the shades in images that started as raw files are much closer together than the shades in JPG files, posterization is much less likely to occur than with JPG images.
Fourth, images that start as raw files are much more suitable for large color spaces than JPG images. Large color spaces include more colors. The shades in an image must be spread across the colors in a color space. The larger the color space, the farther apart the shades must be to cover the entire color space. This is not a problem for fourteen bit raw files. There are so many shades that the shades easily cover the larger color spaces. However, JPG files have a problem. To cover all of the color space with only eight bits, the shades must be spread farther apart. This makes posterization much more likely -- especially if a significant amount of image editing is used (remember, image editing can spread shades even farther apart).
There are other advantages to shooting raw. For those that are interested, the advantages of raw are covered in detail in my three part article series Why Raw.
Most photographers convert their raw files with Camera Raw (see Figure 1). Most likely, this is because Camera Raw comes with Photoshop (the most popular image editing software).
The point should be clear, choosing a raw converter based on the requirements of an image produces superior image quality. To determine which raw converter is best for an image, the image can be converted with each raw converter. Then, an examination of the conversions will reveal which converter will work best.
As much of the editing as possible should be done in the raw converter. This is because edits in a raw converter do not degrade image detail. Conversely, edits in image editing programs, like Photoshop, do degrade image detail. The reason for this is that raw converters work with analogue data. On the other hand, tools within image editing programs work with digital data. The issue here is that edits with analogue data do not degrade image detail while edits with digital data do degrade detail due to quantization error (for a more detailed explanation, see Nondestructive Editing -- Part III).
This can be seen in Figures 5 and 6. Figure 5 shows an image histogram after the image was edited in a raw converter. The important point here is that the histogram is very smooth. This indicates that there has been no image degradation. Figure 6 shows the histogram after a similar edit was performed, on the same image, in Photoshop. The gaps and spikes clearly show that the image has been degraded after the edit was performed.
There are two types of edits:
For those edits that can not be performed in a raw converter, nondestructive editing should be used.
There are many types of nondestructive edits. Some of the more common are:
Nondestructive edits have three main advantages over destructive edits:
For this article, the most important advantage is image quality. Now, there are a number of ways that nondestructive editing improves image quality. While the details of how nondestructive editing creates better image quality are beyond the scope of this article, the end result is that images that are edited with nondestructive edits will have less loss of image detail than images that are edited with destructive edits (for a more detailed investigation of nondestructive editing, see Nondestructive Editing).
Now, there is one point that needs to be addressed with respect to nondestructive editing. Many people think that nondestructive edits do not degrade image detail at all. This is incorrect. With the exception of raw converters, nondestructive edits do cause some image degradation. To better understand this, it is important to remember the definition of nondestructive edits: Nondestructive edits do not change the original data. The definition does not state that nondestructive edits do not degrade image detail. This is clearly demonstrated by referring back to Figure 6. This figure shows a histogram after an image was edited with an adjustment layer -- which is a nondestructive edit. The gaps and spikes in the histogram clearly show that some image degradation has occurred even thought the edit was nondestructive.
So, why use nondestructive edits if they produce some image degradation? The reason is that, while nondestructive edits do produce some image degradation, they produce less image degradation than destructive edits.
So, when editing images, these rules should be followed in order to produce the highest quality images:
One of the last steps before printing is to interpolate the image to its final size. Now, there are many interpolation methods available. Photoshop has several methods. In addition, there are interpolation software packages that can be purchased that have additional interpolation algorithms (for a detailed evaluation of several interpolation methods, see Interpolation).
The main conclusion that I have drawn is that, for linear interpolations of around 200% or smaller, there is not a clear winner between Bicubic Smoother (a Photoshop method), stairstep interpolation (a method done in Photoshop using Bicubic), and the commercial interpolation methods that I have tested. What is obvious is that, for any specific image, one interpolation method may produce better results than another interpolation. Thus, in order to produce the best print quality for a specific image, it is necessary to select the interpolation method that best suits the requirements of that image. This is demonstrated in Figures 7 and 8.
Figure 7 shows an image that was originally interpolated using stairstep interpolation. However, this interpolation method resulted in halos around the very dark edges of the plant. Switching to Bicubic Smoother interpolation eliminated these halos and resulted in a significantly superior quality image.
On the other hand, Figure 8 shows an image where the stairstep interpolation method produced superior results. The stairstep interpolation resulted in better image detail that required considerably less sharpening than when the Bicubic Smoother interpolation was used.
The only way to know which interpolation method will produce the best results for an image is to interpolate the image with different interpolation methods and see which one works best.
On a side note, for linear interpolations of around 200% or less, I use either Bicubic Smoother or stairstep interpolation since they can be done in Photoshop without the necessity of purchasing any additional software. In addition, they appear to me to produce results as good, at this interpolation level, as the methods that require the purchase of separate software.
One step that affects image quality that is often missed is optimizing the printing resolution. The printing resolution is the number of pixels per inch (PPI) at which an image will be printed. In Photoshop, this is set in the Image Size dialogue box (see Figure 9).
There are actually a number of factors that help determine which PPI will produce the best print for a particular image (for more information on printing resolution, see Optimizing Printing Resolution). Furthermore, a PPI that works best for one image may not be the best choice for another image. In other words, the printing resolution needs to be individually selected for each image. This can be seen be looking back at Figures 7 and 8. The image in Figure 7 printed best at 360 PPI while the image in Figure 8 printed best at 300 PPI even though both prints were made on the same printer.
What this means is that, for a particular image, separate prints need to be made, each at a different PPI, in order to determine which PPI will produce the best print. This means that multiple interpolations need to be made to produce the prints at the different PPI settings.
Once the prints are made, they can be examined to determine which PPI setting produces the best print.
Some points should be kept in mind when determining the best PPI setting for a print:
The last issue that needs to be dealt with is the selection of the printer. Obviously, the better the printer, the better the prints that can be made. Unfortunately, it is not so easy to just switch printers. A good inkjet printer can cost anywhere from hundreds to thousands of dollars.
While good inkjet printers are costly, they affect image quality in two main ways:
For many photographers, the color gamut of the printer is a major concern. Older generation printers have smaller gamuts. As a result, they are not capable of producing as many colors as later generation printers that use newer ink formulations with wider gamuts.
I ran into this issue with my printer. I was unhappy with the printer that I was using because it was incapable of printing some of the more saturated colors. As a result, I was not able to get satisfactory results with some of my images that had fairly saturated colors. This was particularly true for images that I had shot in the spring (with spring's saturated greens) and the fall (with autumn's saturated reds, oranges, and yellows). Several of these images were among my favorites. Yet, the prints looked lackluster due to the rather subdued colors that resulted with my printer.
After debating the issue for about a year, I finally purchased a new printer. The new printer can print most of the saturated colors that the old printer could not. In addition, the new printer has a host of other improved features that result in substantially improved image quality.
Now, I am not suggesting that everyone run out and buy a new printer. However, every photographer should be aware that the printer has a major impact on the quality of the final prints. If a photographer's printer is not up to standards, the quality of the prints it produces will be sub-optimal no matter how good the camera or workflow that was used to create the image.
In short, just as we buy new digital cameras as the technology evolves in order to produce higher quality images, we need to consider upgrading our printers every once in a while in order to get the superior image quality that the improved printer technology delivers.
These methods obviously require time and effort to produce the highest quality prints. Thus, they are not meant for all images. However, when we desire the highest quality, these methods can produce visibly superior prints.