Nondestructive Editing In Photoshop -- Part III

Article and Photography by Ron Bigelow

www.ronbigelow.com

Photoshop CS4 Used in this Tutorial

Nondestructive Editing -- Smart Filters

Now, it is time to get into one of the biggest advantages of Smart Objects and that is Smart Filters. Smart Filters are the same filters that are commonly used in Photoshop: filters such as Unsharp Mask, Gaussian Blur, High Pass, Median, and Emboss. So, if these are the same filters that are typically used, what’s the big deal? Well, when applied as Smart Filters, the filters are nondestructive. In addition, changes can be made to the filters at any time. For example, with Unsharp Mask, the radius setting can be changed at any point during the editing. This is not possible without Smart Filters.

Figure 1: Beach Scene

Figure 1 shows the beach scene again. At this point, it is necessary to sharpen the image. This can be done by using a Smart Filter.

Figure 2: Layers Panel

Applying Smart Filters is pretty easy. However, Smart Filters can only be applied to Smart Objects. At this time, the image isn’t a Smart Object. This is obvious because the only layer in this image is missing the Smart Object symbol in the lower right corner of the layer icon (see Figure 2). To convert this image to a Smart Object, choose Layer/Smart Objects/Convert to Smart Object.

Figure 3: Smart Object Layer

The Smart Object symbol is now visible on the layer (see Figure 3), so the image is now a Smart Object.

Figure 4: Unsharp Mask Dialogue Box
Next, the Unsharp Mask filter is applied by choosing Filter/Sharpen/Unsharp Mask. The Unsharp Mask dialogue box appears. For this image, the settings shown in Figure 4 are used.
Figure 5: Unsharp Mask Smart Filter

After clicking OK, the Unsharp Mask Smart Filter appears below the image layer (see Figure 5).

Figure 6: Unsharp Mask Smart Filter Reopened

Suppose that, after several more editing steps, it was decided that the sharpening needed to be adjusted. Now, in the days before Smart Objects, there wouldn't be anything that could be done about the sharpening except to delete the sharpening layer and start over again. Fortunately, the sharpening on this image was done with a Smart Filter, so all that needs to be done is to double click the Unsharp Mask Smart Filter. The Unsharp Mask dialogue box will reopen (see Figure 6), and the Unsharp Mask settings can be adjusted.

The great thing about using a Smart Filter is that the filter can be readjusted as many times as necessary, and the edits will always be nondestructive.

Now, there is a lot more to the use of Smart Filters than was shown here (e.g., the use of Blend modes, Opacity, and masks). For a more detailed tutorial on the use of Smart Filters, please check out the Smart Objects video series in the video section on this web site (Smart Filters start on video V).

Nondestructive Editing -- Camera Raw

Earlier in this article series, it was mentioned that editing with Camera Raw is nondestructive. However, editing in Camera Raw is a special case of nondestructive editing. Camera Raw edits are not only nondestructive; Camera Raw edits are the only edits that do not degrade image detail (Actually, this applies to any raw converter. However, since this article deals with Photoshop, we will concentrate on Camera Raw).

As covered earlier, almost all edits cause image degradation -- even nondestructive edits. 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. In fact, almost all of them do. The key here is the word "almost". The single exception is Camera Raw.

Camera Raw is the single exception because Camera Raw is the only nondestructive editing tool, within Photoshop, that works with analogue data. All of the rest of the tools work with digital data.

Why is this? Well, the image starts off in the camera as analogue data. Only later is the data converted to digital. For the case of JPG images, the data is converted to digital in the camera. On the other hand, for raw images, Camera Raw takes the data from the camera while it is still in analogue form. Camera Raw then converts the image to digital during the raw conversion. Therefore, any image that is in Photoshop (unless it is still in Camera Raw) has already been converted to digital. Thus, raw images in Camera Raw are still analogue images while images in Photoshop are digital images.

Naturally, the next question is, "Why does this matter?" When a person looks at an image in Photoshop or Camera Raw, the person sees an image. When edits are performed, the person sees tonal and color changes occur. However, Photoshop and Camera Raw don't really work directly with images. Photoshop and Camera Raw actually work with numbers. When edits are performed, Photoshop and Camera Raw perform mathematical operations on those numbers. The main point here is that mathematical operations performed on analogue data do not degrade the numbers, but operations performed on digital data do degrade the numbers.

The reason that this occurs is because of the nature of analogue vs. digital numbers. Analogue numbers are numbers that are continuous. From a simple viewpoint, this means that analogue numbers have decimal points with an unlimited number of decimal places. So 3.542 and 8.124 are analogue numbers. Thus, analogue numbers can be very precise. If more precision is needed with analogue numbers, more decimal places will simply be used. Conversely, digital numbers are numbers that are discrete. In other words, only certain values can be represented digitally. Often, digital numbers are integers (whole numbers). When this is the case, as it is with digital images, numbers like 4 and 9 are digital because they are integers, but 5.46 is not a digital number because it is not an integer.

A little demonstration will illustrate how mathematical operations do not degrade analogue numbers but do degrade digital numbers. Figure 7 shows a series of numbers in the first row. These numbers are analogue numbers. The second row shows the numbers after a mathematical operation was performed on the numbers. In this case, the numbers were divided by 4.0. The third row shows the numbers after the numbers in the second row were multiplied by 4.0. In essence, the second operation reverses the first operation. Looking at the third row shows that the third row ends up the exact same as the first row. In other words, two operations were performed on these analogue numbers, and the final numbers ending up being the same as the original numbers. In other words, no degradation of the numbers occurred.

 

Figure 7: Analogue Numbers
Original
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
Divided by 4.0
0.25
0.50
0.75
1.00
1.25
1.50
1.75
2.00
2.25
2.50
Multiplied by 4.0
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00

Figure 8 shows the same process except the two mathematical operations were performed on digital numbers. The first row shows the original numbers (which are integers as with digital photography data). The second row shows the numbers after the numbers were divided by 4. Since the digital numbers being used in this example are integers, the numbers in the second row must be rounded off to the nearest integer. The third row shows the numbers after the numbers in the second row were multiplied by 4. Looking at the third row shows that the third row does not end up the same as the first row. Even though the two operations reversed the mathematical process, the digital numbers in row three are different that the original numbers. Basically, the numbers have been altered.

Furthermore, in addition to the data being altered, there are fewer numbers at the end of the process than at the beginning. The process started out with ten numbers and ended up with only four (i.e., 0, 4, 8, and 12). In essence, the numerical data has been degraded.

Figure 8: Digital Numbers
Original
1
2
3
4
5
6
7
8
9
10
Divided by 4
0
1
1
1
1
2
2
2
2
3
Multiplied by 4
0
4
4
4
4
8
8
8
8
12

Clearly, this shows that editing analogue data does not degrade the data, but editing digital data does.

It is now time to see how Camera Raw produces images with no image degradation while other nondestructive techniques result in degradation.

Figure 9 shows the beach image raw file that has been opened in Camera Raw by choosing File/Open and accessing the file. In Camera Raw, the Tone Curve has been used to increase the contrast of the image.

Figure 9: Camera Raw Edit
Figure 10: Histogram after Camera Raw Edit
Figure 10 shows the histogram of the image once the raw conversion has been completed and the image has been opened in Photoshop. The important point here is that the histogram is very smooth. This indicates that there has been no image degradation
Figure 11: Layers Panel

Now, it is time to perform a similar edit in Photoshop rather than in Camera Raw. Figure 11 shows the Layers panel of the exact same beach image that was edited in Camera Raw (except the Tone Curve was not applied in Camera Raw). This time, the contrast was increased by adding a nondestructive Curves adjustment layer (see Figure 12).

Figure 12: Curves
 

Figure 13: Histogram after Curves Edit

Figure 13 shows the histogram for the image. The gaps and spikes clearly show that the image has been degraded after the edit was performed -- even though a nondestructive edit was used. Now, this image was deliberately edited as an eight bit image to emphasize the image degradation. If the image had been edited as a sixteen bit image, the image degradation would have been less. However, even with a sixteen bit image, the degradation still occurs; it is just smaller in magnitude.

The point of this should now be apparent. When high quality images are required, as much of the editing as possible should be carried out in Camera Raw.

Summary

For the purposes of preserving the original data; flexibility; and image quality, nondestructive editing should be used. When the highest image quality is desired, Camera Raw should be used for as much of the editing as possible.

Articles

Nondestructive -- Part II