Chapter 6. Color Correction

Author: Jeff I. Greenberg

Color is my day-long obsession, joy, and torment.

—Claude Monet

Even if you’ve managed to practice your best videography techniques with expert lighting and export camera work, there is always some improvement in the finished image via color correction.

Your eyes are these amazing, white-balancing organs that constantly interpret and reinterpret your environment. But the same qualities that make your eyes great for the everyday world limit you in the postproduction suite because your eyes are continually adjusting and adapting to what they’re viewing.

For color correction, you need a more consistent way to examine your images, and for this you need to be proficient in using color scopes. Each shot needs to be adjusted and controlled to make sure it doesn’t exceed broadcast limits; even today when the final delivery may only be to digital devices, such as an iPad or YouTube, broadcast limitations are still valuable restrictions for finished work.

Getting two shots to match, especially when they’re shot under different lighting conditions and exposures can be difficult. It should never be a guessing game. The overall goal is to move beyond just getting the picture to “look good.” Color correction can also be an expert storytelling technique, intentionally warming up (adding red/orange) for more romantic shots or cooling down (adding some blue) a shot to make it appear more serious.

Defining “looks” of a scene can further communicate location: For example, images of Miami have a warmer look than images of Alaska.

Color correction is a crucial, necessary technique that can help any production look more professional.

You may be reading this chapter and saying, “I already color correct.” Yes, but are you using all of your scopes? Do you know how to match shots? If you’re not using the Parade scope and if you struggle to match shots, this chapter dives into those topics in depth as well.

Professional colorists, with optimum hardware setups and working as fast as they can, will correct 10–20 minutes of finished work a day. If in 10 minutes they can complete about 200 shots, it takes them, on average, about a minute a clip. Expect that you’ll be slower than this. Make sure you budget and measure how fast you can correct shots to guarantee your delivery schedule. Although you’ll become faster after practice, full-time colorists also have the advantage of using some dedicated hardware.

You need your goals to be more concrete and to serve storytelling.

Here are the four goals of color correction in order:

1. Establish the appropriate tonal range.

2. Remove any color casts.

3. Match shots for shot-to-shot consistency.

4. Create a look (if appropriate).

All of these goals, except the last one, should be invisible to viewers; well done color correction doesn’t call attention to itself.

The first two goals, establishing a tonal range and removing any color casts, are about neutralizing your image. These two goals are about basic cleanup and establishing a starting point for your images.

When possible, it’s ideal to have some communication with a DP or Director about the on-set intentions of the shoot for color correcting. If the intent during shooting included adding a cool filter (blue) to the lens but you weren’t aware of that, you’d be fighting the intentional blue cast on the footage.

Matching shots is more about hiding any shooting defects, such as the color temperature of an indoor shot compared to an exterior shot. It’s not about matching something like a Pantone color. You can get close, but you can’t truly match print items. The reason is that there is a difference between reflected light (print) and emitted light (LCDs and CRTs). Again, if the matching is done well, nobody notices it.

Developing a “look” means adding an artistic impression to the footage. Movies like the Matrix, Minority Report, and Se7en all have stylistic, visual adjustments to communicate concepts like technology, emotion, and danger (Figure 6.1). This type of work is typically left until the end of color correction, because a given scene or series of shots should all be similar before a look is added.

The human eye is sensitive to only a small portion of the energy emitted from the sun—which is called the visible spectrum. It ranges from (ultra) violet on the left through (infra) red on the right (Figure 6.2).

When your clients ask for a warm or cool look (Figure 6.3), you need to understand what they’re thinking.

Adding the maximum R, G, and B to a pixel (Figure 6.4)—let’s call this 100%—yields a white pixel.

The opposite, removing all RGB (0%) yields a black pixel (Figure 6.5).

Gray can be the result of all three sliders at 50%, but it can also be any point where all RGB are identical. All RGB at 5% is a very dark gray, and all RGB at 87% is a light gray. The core concept is that when all three values are the same (Figure 6.6), the color is gray.

Because computers work with these three channels, you should take a moment to look at each channel individually (in the next section “CMY(K) or secondary colors,” you’ll look at the lack of color in each channel with possibly surprising results).

With green and blue at 0, increasing just the red to 255 results in a swatch that is (obviously) red. Similarly, you can do this with just the green control and then just the blue control (Figure 6.7). Although the results may seem obvious, this is an important step in understanding the way computers handle these three color channels.

If you maximize the Blue and Green channels (at 255) and reduce Red to 0, the lack of red results in cyan, making it the opposite of red (Figure 6.8).

If you maximize the Red and Blue channels and reduce Green to 0, the lack of green results in magenta.

If you maximize the Red and Green channels and reduce Blue to 0, the lack of blue results in yellow.

CMY colors are described as the secondary colors to RGB. It’s likely you recognize them from your color printer inks. Computers use emitted light, but paper uses reflected light, which requires a different method to achieve black and white. With a computer, all the colors (RGB) need to be on to get white. But in print, zero information is needed (no ink) to get white. This is the opposite of RGB. Even still, using all three colors in print, CMY only produces a dark brown, so it’s necessary to actually use black ink to yield black.

As an aside, when you look at print, which is reflective of light (versus computers, which emit light), it uses the CMYK (Cyan, Magenta, Yellow, and Black) color space to represent color.

These colors match the layout of the Vectorscope: The scope that allows you to measure the overall color vectors in an image; they also match the layout of the colors in the Three-Way Color Corrector. Once you understand this idea that the Vectorscope and the Three-Way Color Corrector correspond, you’ll intuitively know how to compensate for chroma issues.

Video started out in black and white only, and it was only necessary to record and store the luminance of an image—called Y by video engineers.

Later, in the 60s, color was added to the existing signal. It was possible to describe just the color information without the luminance (brightness). This additional information was only a small amount of extra information because your eyes don’t see color detail as well as luminance details.

In video, the color channels are described as Cr’ (color red prime) and Cb’ (color blue prime), and the green information is mathematically divided into all three channels.

In an ideal world you’d be able to handle all of your video the way it’s stored, which would be in YCr’Cb’ math rather than converting it to RGB math and back again. There are a number of reasons you can’t, many of which have to do with your expected results. You expect the mosaic effect to transform the image in a certain way; YCr’Cb’ wouldn’t produce results that you’d be happy with.

This conversion of from YCr’Cb’ color space to RGB and back can create a small amount of damage to the color information, because multiplication and division are occurring, and at some point, some of the information may be discarded (rounded up or down).

For many, YCr’Cb’ is a mouthful, so it’s acceptable to use the term YUV (technically, it’s the analog description) instead of YCr’Cb’ (the digital description). Adobe Premiere Pro shows this information in color palettes (Figure 6.10).

So, whenever possible when you’re adding effects, you want to do little or no damage to your video’s color space. Adobe Premiere Pro has a specific filter (switch) for effects that are only calculated using YUV space. Figure 6.11 shows the YUV Effects icon.

When you use these effects, you’ll process the picture in its native color space and generally not affect the chroma when you work with luma or vice versa.

What happens to the effects that aren’t YUV? They’re processed using RGB. Because RGB mixes luminance (black and white) and chrominance (color) together, adjusting the controls may affect both the luma and the chroma.

Now that you have a more thorough understanding of color, the next step is to examine your system setup.

What tools do you need?

You need tools to process in YUV space to adjust video with as little damage to its information as possible.

You need video scopes, particularly the Waveform, Vectorscope, and RGB Parade, to accurately measure your pictures.

You need to have a layout that permits you to optimize your workflow.

You need to be able to quickly turn off or on your corrections to help your eyes reevaluate the picture.

Adobe Premiere Pro has all of these tools and capabilities. Let’s start by looking at the measurement tools, the video scopes.

Even if you think you’re familiar with scopes (Figure 6.12), it’s probably valuable to review this section. Most people don’t have enough experience with the RGB Parade scope, which builds on ideas from the YC Waveform and Vectorscope. The RGB Parade scope provides vital information about color casts and more.

It’s generally accepted that the best use of this scope is to solely examine luminance. Turn off the Chroma setting.

The brighter or denser the clustering of data means that there are more pixels at a given luma value. Look at the sequence 1 Waveform marker 6.14 (Figure 6.14). On the left side of the screen most of the pixels represent a darker area (the wall), which shows a large clustering of energy on the YC Waveform.

A common way to examine the YC Waveform is by breaking it down mentally into three regions, the leftmost, rightmost, and center regions. This makes it easier to help identify “landmarks” in your image, particularly elements that are bright or dark and those that are easily identifiable.

In Figure 6.15 (sequence 1 Waveform marker 6.15), by breaking down the image into sections, it’s easier to look at the center section and identify the wall (brightest part, lower on the scope) and the man’s jacket (darkest part, higher on the scope).

When you examine the gradient ramp in Figure 6.16, (sequence 2 Vectorscope marker 6.16), there is no information on the Vectorscope. This should be expected: There is no chroma (color) in the black to white gradient.

Let’s look at the color bars in Figure 6.17 (marker 6.17).

You now see a dot at each of the RGB and CMY targets. These are at the outer target, which is normally at 100%, the maximum color permitted, but the top of the scope is set at 75%. A general rule is to draw an imaginary shape connecting the dots and keep any color vectors inside of that shape to keep the video broadcast legal.

Now that you’ve seen a gradient and bars, let’s look at what happens when you point your camera at something more interesting—people. Look at the image in Figure 6.18 (marker 6.18). I’ve intentionally added saturation to help make the colors more visible on the Vectorscope.

The sugar packet shows a strong yellow color near the yellow target, and the pixels near the red target represent the man’s shirt. But the flesh tones are more interesting. The line between yellow and red is unofficially called the flesh line by many colorists because flesh tones generally should fall on that line. Regardless of heritage or how much melanin is in people’s skin, nearly all faces (unless sunburnt) should fall along this line.

There are several major signposts that you should look for on an RGB Parade:

Neutral items. If you can find a neutral item (sheet of paper or any gray or white object) on a waveform, you can find it in the RGB Parade, and it should be at the same level in each channel. If not, you have a color cast.

In Figure 6.19, a neutral item (circled in cyan on the image and Parade) would be the road. Blue is a little higher compared to the other two scopes, which means there’s a blue color cast.

Bright/Darks. As images become very bright or very dark, the chroma disappears. If you point a camera at a light, it’s always white regardless of the color of the light. If you have an element in very dark shadows, you can’t identify the color at all. As portions of images approach having black or white values, the RGB elements at the brightest/darkest levels should have similar values.

In Figure 6.19 from sequence 3 RGB Parade at marker 6.19 you see a magenta area, which is a shadow of the bushes; it should have a little extra green (because it’s in the shadow of trees). But notice that on the RGB Parade this region is the same on all three channels, showing that there’s no color cast in the shadows.

Also, note the sky (the yellow circles). The sky is a little higher in the Blue channel than the Red or Green channel, but not terribly so, which is part of the reason the sky looks grayish. If this was higher, toward 100 in all three channels, it would be “neutral” and therefore all the colors would be blown out—not something you’d want in the sky.

Flesh tones. Looking across all three channels, flesh tones should descend from R to G to B (Figure 6.20). The way to think about this goes back to the color wheel. If three channels (RGB) are the same, the result is gray. To create a flesh tone (which is an orange) with just these three colors, you need to add red and subtract blue (adding yellow). Correspondingly, the Red channel would be the highest, the Blue channel would be the lowest, and the Green channel would be in between on the RGB Parade.

The RGB Parade scope is probably the most important scope you’re not using well.

Because you’ll primarily use the YC Waveform, Vectorscope, and RGB Parade, in the next section I’ll talk about mapping them to the keyboard.

Two adjustments should be made: one to your workspace and the other to the keyboard. As always, refine these adjustments beyond the default setups, because it’s important to test and explore what works best for your system.

Depending on your system (you might have two screens or a wide screen), you may want to adjust the workspace for an easier workflow. I’d rather have the scopes in the reference in the center than at the bottom and the effects off to the left. This setup makes it a little faster and easier to adjust using the Three-Way Color Corrector and keep your eyes on the image.

Here’s how I built my workspace:

1. Choose Window> Workspace > Color Correction to start with the default Color Correction Workspace. If you’ve customized it, reset it to the original layout.

2. Click and hold the panel texture next to the reference monitor you want to merge with the top-center frame. Make sure you click the texture area to drag (Figure 6.23).

3. Drag the reference to the center of the top frame. Be sure to hit the center of the trapezoidal overlay (Figure 6.24). By dragging the reference to the center of the trapezoid, it’ll merge with the Source Monitor (if you drag to one of the edges, it will subdivide the panel).

4. Click the Effect Controls on the left pane, and close the keyframe area (Figure 6.25). Although you might want to keyframe some color correction effects, it’s move valuable to close the panel to have more workspace real estate.

5. Save this workspace by choosing Windows > Workspace > New Workspace. By doing so, you’ll be able to recall it quickly in the future.

This workspace is built into the accompanying project for this chapter. It’s called Adjusted Color Correction and is visible in Figure 6.26.

My number one rule for shortcuts is that that any mapping on my keyboard must make sense. I map the scopes to numbers 1, 2, and 3 because I don’t need or use multicam when color correcting. If you do use multicam often, save a separate keyboard shortcut just for color correction.

To remap the keyboard, follow these steps.

1. On a Mac choose Premiere Pro > Keyboard Shortcuts. On Windows choose Edit > Keyboard shortcuts.

You need to map the Waveform, Vectorscope, RGB Parade, and Composite Video of the Program Monitor. The easiest way to do this is to type part of each word into the search box at the top. When the command appears, double-click the text box to the right.

2. Map Waveform to 1, Vectorscope to 2, the RGB Parade to 3, and Composite Video to 4 to quickly toggle between the three.

Make sure you chose the Program Monitor panel, not the Source Monitor panel.

3. Optional: Map “Effect” Enabled to any key you like. I chose the 1 key (which overlaps the Waveform monitor), which is OK because it’s a different window.

Now that your setup is customized and accurate, let’s explore goals and process.

Recall the original goals of color correction: It’s the first two that are important here: Establish the appropriate tonal range and remove any color casts. Using your chosen shot, work backwards and forwards in a scene, correcting other shots to match it. This sets up the third rule: Match shots for shot-to-shot consistency.

Optionally, apply a look to the scene, which is the fourth rule (and the least important when you’re rushing): Create a look (if appropriate).

Three-Way Color Corrector. This effect has become the workhorse of color correcting footage. It offers three adjustments for luma and three for chroma, each divided into the ranges of Shadows, Midtones, and Highlights. It’s the major method of color correction because it easily maps to control surfaces.

RGB Curves. Some colorists prefer curves because they allow for a higher degree of adjustment than the three ranges (Shadows, Midtones, and Highlights). Using RGB Curves is neither worse nor better but rather a different method based on different mathematics.

Video Limiter. Using this effect keeps your final work legal for output, which is crucial, even if it’s never meant for broadcast.

Could you use other effects in the panel? Yes, and they do some great things; for example, Leave Color and Tint are stylistic effects. But for the ease and speed of color correction, the three effects in the preceding list will be the key to achieving the goals of color correction.

Another very valuable effect that is new to the Adobe Premiere Pro Creative Cloud release is the Lumetri effect. This effect harnesses the engine of Adobe SpeedGrade in Adobe Premiere Pro CC by allowing you to apply Looks (stylizing the footage) from Adobe SpeedGrade directly into Adobe Premiere Pro.

It can be used in one of two ways:

Primary correction. Primary correction uses the three color wheels and the Input and Output levels (Figure 6.27) to adjust the overall image.

Secondary corrections. When you activate this section of the Three-Way Color Corrector, you limit your adjustments to just part of the image.

Secondary controls can be found under the category called Secondary Color Correction. The controls are a keyer, which uses the same concept as greenscreen effects. It limits the Three-Way Color Corrector to adjusting only what falls on the inside or the outside of the keyer. When activated, the entire effect becomes a secondary effect; it’s generally used after primary correction has been done (with a Three-Way Color Corrector or RGB Curves). In Figure 6.28, a key has been done to adjust the actor’s face with the mask left on.

A check box on the Tonal Range Definition (when opened, Figure 6.29) helps you visualize where each of the three wheels work. Select the Show Tonal Range check box to see the three ranges.

A well-exposed shot will have some exposure in all three ranges.

Adjusting the Black point to the right pulls more darker pixels toward black. Adjusting the White point to the left pulls more lighter pixels toward white. And adjusting the Gray point to the left makes an image darker, whereas moving it to the right makes the image brighter.

The general order of operations should be to adjust the Black point, White point, and then Gamma.

There are three white balance eyedroppers, one at the bottom of each of the three color wheels. Each eyedropper helps to quickly neutralize a color cast in the Shadows, Midtones, or Highlights. The idea is that you use the eyedropper on a part of the image that is neutral (such as a white shirt), and it neutralizes a color cast in that luminance region by adding in the opposite color.

On the rare occasions when you need to magnify the intensity of an adjustment, click and drag on the thin perpendicular line to drag the slider bar (Figure 6.32).

Dragging them upwards brightens that point; dragging them downwards darkens that point. Just click on the line to add a point, and drag the point from the curve area to remove the point.

The advantage of using RGB Curves lies in being able to manipulate each channel directly versus having to “balance” each range. Many use this capability to build a subtle look, such as in Figure 6.34, where I’ve built a high-contrast S curve on the Master Curve and added a little extra blue into the lower portion (shadows) of the Blue Curve.

The RGB Curves controls allow for adjustments of a Secondary Color Correction but have no saturation controls. If you want saturation controls, you’d have to add another effect (such as the Three-Way Color Corrector) either before or, more likely, after the Curves.

It’s important to be able to compare your corrections with the original image. When you’re color correcting, the 90 second rule takes effect: After 90 seconds, your brain begins to compensate for color problems and forgets the original image.

The Split View (Figure 6.35) allows for a quick comparison. Between using Split View and toggling the effect on and off, your eyes will “remember” the original image.

Later in this chapter, I mention a better limiter.

Your system may not necessarily match the exact look or output of what you see in this book (or what was done in the project). The sidebar “Real-world Setup” addresses some of these issues.

Most important is that if the image doesn’t look good to your eye on a calibrated monitor, it doesn’t work!

Use the sequence 5 Common Primary Corrections to follow along with the examples in this section.

1. Fix the luma:

Check the Input Black/White levels against the Waveform scope.

Adjust the Midpoint/Gamma slider until the image “feels” correctly exposed. You may have to go back and adjust the Black/White levels after this step.

2. Fix the chroma:

Adjust the midtones by clicking and dragging from the center of the Midtones wheel in the opposite direction of any perceived color cast in the Vectorscope.

Adjust the Highlights and Shadows color wheels while looking at the RGB Parade.

The shot at marker 6.36 (Figure 6.36) in sequence 5 Common Primary Correction shows a shot that is fairly dark. Let’s assume that this result wasn’t the intention of the DP on set.

Reading the Waveform shows shadows at around 0.3 and highlights at 0.9, and most of the shot is below 0.6, meaning that the bulk of the exposure will be dark.

Additionally, although there is a bright section of the man’s hair (right side), most likely, brightest areas shouldn’t be set all the way to white but rather about 5% below that. Here are the steps you should take to fix luma issues.

1. Shadows first. Because the shadows already fairly dark, the Input Black level probably doesn’t need to be changed.

2. Highlights next. Adjusting the Input White level to the left brightens the highlights. Only a little adjustment is necessary (236 is what I ended up with), because you don’t want the brightest part to be as bright as pointing at a light source.

3. Midtones/Gammas last. Adjusting the Midtone slider to 1.3 produces a brighter picture, but it now feels a little washed out. Adjusting the Black level to 5 helps restore some of the contrast.

Feel free to try the preceding technique on the Timeline with the blown out version of this shot at the marker named Blown out.

At marker 6.37 is a red shot, as shown in Figure 6.37.

Let’s look through the scopes. The damage in the Red channel is significant, and looking at the Vectorscope you’ll see such strong damage that it may be a struggle to fix the shot.

The exposure shown in the Waveform monitor is close enough to be OK. The problem with this image is the chroma. Normally, you’d first adjust the luma, but because the image already has good exposure, let’s skip it and just deal with the chroma issue.

Looking at the Vectorscope, you need to move opposite of the color cast. Because this image has some neutral items (the wall, the newspaper, the man’s hair) you can cheat and click those areas.

Move the Midtones wheel toward cyan. Optionally, you can click in the eyedropper and sample the wall (it’s a neutral item based on other shots). Then adjust the magnification slider. These adjustments help quite a bit, but the blob has moved about half the distance. (Open the Midtones; the number values I ended up with were Magnitude 41, Gain 100, Angle 47.) That’s not enough to repair the image. Let’s switch to the RGB Parade and look at the highlight areas and shadows.

The highlights show too much red and green (the man’s hair on the right side); some blue/cyan needs to be added to compensate. You can use the eyedropper but not in the brightest spot. Better to use the eyedropper in a spot that’s bright but not fully white. (Open the Highlights; the number values I ended up with were Magnitude 46, Gain 71, Angle 29.) But there’s still quite a cast in the image.

The Shadows remain, and his jacket (the darkest element on all three scopes) shows too much red. Using the same procedure you did earlier, try to get all three channels closer (like the highlights) at this point, and try to balance the jacket, the neutral item. (Open the Shadows; the number values I ended up with were Magnitude 36, Gain 20, Angle 47.) The image looks much better, but it still has a strong color.

Reducing saturation is the key. Recall an earlier tip in this chapter: When you’re struggling with chroma, it’s nearly always the shadows that cause problems. Reducing the Shadow Saturation to 50 makes the shot look almost normal. You might also try adjusting the Master Saturation.

The finished effect is on the shot in the sequence that matches Figure 6.38.

Look at the image in Figure 6.39 at marker 6.39; let’s start with luma and then fix the chroma.

1. Check the Input Black/White levels against the Waveform scope.

In this step I adjusted the Input Black level by sliding it toward the right a little. I prefer sliding the numbers rather than using the actual arrow. I wanted a slightly more crushed black in the darker areas, so I made more of the image fall toward black. I then adjusted the Input White level.

2. Adjust the Midpoint/Gamma slider until the image “feels” correctly exposed.

I felt the image was a little too dark (keep in mind the DP or Directors might have wanted that!), so I dragged the middle slider/Gamma to the right, brightening up the image. I set the Input Black level to 15.2 and the Input White level to 235.2. The Gamma slider ended up at 1.4.

By turning on Split View as a comparison (Figure 6.40), you can see that the shot is exposed slightly brighter with greater contrast.

Now let’s look at chroma.

3. Start with the midtones and focus on the Vectorscope.

The Vectorscope shows some yellow and red energy and quite a bit of energy between the two spots; yellow + red = brown. And when your eyes look at the image (Figure 6.41), it has a warm (brown) feel.

To compensate, I added just a little bit of blue in the Midtones color wheel. Then I clicked and dragged just a little toward blue (cooling down the shot.)

4. Now, let’s focus on the Shadows and Highlights chroma wheel while looking at the RGB Parade (Figure 6.42).

Generally, any chroma changes in the highlight and shadows are slight, very subtle changes. Looking at the highlights in the image, the brightest spot is the blown out spot on the actor’s face.

As the traces approach white (100), there should generally be a loss of color, yet here there is more red than green and blue. I added a little blue in the color wheel to balance that out while being careful not to neutralize the wall.

The shadow areas show no blue (there is almost no blue at the bottom of the RGB Parade), so I added just a little blue in the Shadows color wheel to balance the blue on the RGB Parade to be similar to the red and green shadow areas.

Both of the adjustments I made are very subtle, but by toggling on and off the entire effect (and viewing full screen), you can focus your eyes on the bright and dark areas to see if you can perceive the chroma change.

RGB Curves isn’t a better or worse filter than the Three-Way Color Corrector; it’s just different. Because it uses different math (see the sidebar “How RGB Processing Differs”), you may have an easier or harder time fixing color problems.

1. Adjust the brightest spots of all three channels to make them similar (as appropriate) while watching the RGB Parade. Repeat this step for the darkest areas.

2. Look at the Vectorscope; add or remove points in the middle of the curves to properly center the blob.

3. If needed, adjust the Master curve to change the luma. Most likely, you’ll use a point in the midtones to lighten or darken the image.

1. Using the same shot, at marker 6.44, I applied the RGB Curves. Then I checked the top of the RGB Parade and saw that the brightest spot was darker in the Blue channel (about 85%).

I added a node on the blue RGB curve that’s approximately 85% of the way up the slope and pulled it upwards to help the top of the Blue channel. Next, I added a second node at the top of the Red channel to pull it downwards (Figure 6.43) to help the Red channel match.

To adjust the shadows, I then added a node on the bottom of the Blue curve to pull up the Blue channel just a little. The idea is to get some similarity in all three channels (the absence of light/shadow means that all three channels should be similar).

Be aware that the image may not look great. In this case, the image has a cyan/blue cast to it. What was the problem? The midtones.

2. Switch to the Vectorscope.

Because there is too much cyan/blue, I needed to add red/yellow. And because the problem was in the midtones, I added a point to the Red curve in the middle, pulling it up and adding red. Similarly, I needed to introduce yellow, so I set a midtone dot in the Blue channel and lowered it, adding yellow (Figure 6.44 on the next page).

3. Now all that was left was to check the exposure on the Waveform and possibly brighten or darken the image here.

I added a dot in the middle of the Master curve and dragged it upwards to lighten the image. In fact, the Master curve behaves similarly to the way a Gamma correction does in the Three-Way Color Corrector.

You can see the final image in the sequence 5 Common Primary Corrections at marker 6.44.

1. Starting with the RGB Parade highlights, the blue highlights were too low and the red highlights were too high. Adjusting the tops of the Blue (1) and Red curves (2) repaired those problems.

2. Next, I repaired the shadows. In the Shadows of the RGB Parade there was too much red, so it needed to be reduced in the Red curve (3).

3. Now to the midtones. The newspaper is a neutral item that exists in the midtones; when I found it on the RGB Parade, it showed too much red. Adding a new point at the halfway mark of the Red curve (4) and reducing it began to repair the problem. At one point, the shot turned green, so I knew I went too far. The Blue channel needed a little assistance, because in the newspaper there wasn’t enough blue on the RGB Parade where the newspaper information was, so I lifted the midpoint of the Blue curve (5).

4. I switched to the Vectorscope and saw that there was still quite a red cast. Using just my eyes (and glancing back at the RGB Parade) I saw that it was in the brighter portion of the midtones. So, I added another red node (6) with a second node to limit its effect (7), reducing the reds in the top of the midtones.

5. After checking the Luminance Waveform (and the image), I added some points to produce some slight brightening (8) of the midtones along with restoring some of the black (9) and white points (10) to produce a slightly higher contrast to the image.

The finished version of the shot is on the Timeline. Feel free to turn off the RGB Curves, add your own, and see which method helps you correct the cast faster—the Three-Way Color Corrector or the RGB Curves.

A primary color correction affects the entire shot. But a secondary color correction affects only part of the shot, which means the results come from adding an additional second effect (or third, fourth, etc.).

Both the Three-Way Color Corrector and the RGB Curves have a section labeled Secondary Color Correction. In that section are chroma keyer-based corrections. You build a key in the effect to limit what part of the shot the effect works on.

Colorists will tell you that there is also a shape-based correction that works on part of the image based on a shape, such as for vignettes or simulating a grad filter. Although Adobe Premiere Pro doesn’t have this feature built in, I’ll describe a method of creating a shape-based correction.

The key is based on three parameters: Hue, Saturation, and Luma (HSL).

The struggle you’ll have when using this feature will always be the same—getting a good key. The best technique is to select a color with the eyedropper and then watch to see what parts of the H, S, and L react.

Make sure the Hue, Saturation, and Luma areas are open to help identify what the keyer selects. Each section has an area that represents the selection, such as the Saturation range shown in Figure 6.47. The two inner squares are the selected region (Start/End Threshold). The two outer triangles (Start/End softens) help the selection gradually include similar nearby regions.

Use the plus and minus eyedroppers to add or remove from the key. For flesh tones, it’s ideal to have a solid selection in the areas of the face (white) and everything else unselected (black).

Here are some additional suggestions to refine your selection:

Luma may require adjustment to either or both the Start/End Threshold. The trick is to look at the image and decide if you’re missing dark regions or bright regions.

Saturation usually needs the Start Threshold adjusted (the areas with less saturation).

Hue is probably the least likely adjustment you’ll need to make.

If moving an adjustment made no change, return the adjustment to where it started (or click Undo).

Adjusting the Thin controls shrinks or expands the mask to help make the adjusted selection a little larger or smaller if needed.

Because face tones are a mixture of red plus yellow (they lack blue), it may be a little tricky to get a good selection if other elements in the image use similar colors. Rarely are walls painted white. Often, they have a little color in them, and a beige tone is all too common, which can be similar to face tones.

Combining the techniques in the section “Shape-based Corrections,” later in this chapter can help you control and limit the secondary to just a region, which is perfect for dealing with a similar unwanted color (e.g., a beige wall would be unwanted when using a secondary on a flesh tone).

In sequence 6 Secondaries (Figure 6.48), on the first shot you can see the results of the effect. It’s in Split View to help you see it in print.

A primary correction was built with the first Three-Way Color Corrector and a second Three-Way Color Corrector was added to adjust just the flesh tones. Let’s work through the steps.

1. Make the initial selection. I made the initial selection by clicking on the left portion of the man’s cheek (Figure 6.48) and then selected Show Mask. The mask was missing the darker and brighter parts of his face, so I adjusted the Thresholds rather than use the eyedroppers to add more of a selection.

2. Adjust Luma Thresholds. Because I wanted to include the darker and brighter parts of his face, I adjusted the Luma Start and End Thresholds to increase the matte selection. Moving the Luma Start Threshold to the left included more of the left side (darker parts) of the face. Moving the Luma End Threshold included a little more of the right side (brighter parts) of the face.

3. Adjust Saturation Thresholds. The right side of his face was still somewhat unselected, because it’s the brightest part of the face. And, as you recall from the “RGB Parade” section, as brightness increases, usually less chroma (saturation) exists. Decreasing the Saturation End Threshold helped include the bright areas but also included the bench. In this image it was acceptable to allow the bench selection in order to adjust the face because the adjustment is subtle.

4. Always Soften. As always, soften the selection (mask) just a little (in this case, set Soften to 3.0).

Deselect Show Mask and make color adjustments to just the flesh tones.

In this example, I added some saturation (Master Saturation at 140) and slightly lifted the Midtone Luma/Gamma to 1.1.

At marker 6.49 (Figure 6.49) you can see the finished version of a sky correction in Split View, comparing the sky in before and after images. The original sky (on the right) is very bright and has nearly no chroma at all, leaving it feeling very stark.

A primary correction was built with the first Three-Way Color Corrector and a second Three-Way Color Corrector was added to adjust just the sky. Let’s walk through this example.

1. Make the initial selection. Skies are usually the brightest element in an image, and in this case, have little information other than being very bright. There was almost no chroma. Making a selection with the eyedropper would have been a mistake; it would have chosen a hue, and with the sky so blown out, I wasn’t sure it would have selected the color blue because there was little to no blue in the sky. So, in this case, instead of using the eyedropper, I wanted the selection to include only the “bright” areas (luma)—but also make sure I kept the hues and all of the saturations—leaving just a question of selecting “bright” (pixels of high luma). Therefore, rather than using the eyedropper, I adjusted just the Luma controls section directly. I skipped making the initial selection and selected Show Mask to see the mask (which was completely white; everything was selected).

2. Adjust Luma Thresholds. The entire luma range was selected, from the dark pixels (the Luma Start Threshold) to the bright pixels (the Luma End Threshold). I wanted to select all of the bright pixels, not the darker areas of the image. I then dragged the Luma Start Threshold to the right (to ignore all the darker pixels) until only the sky was selected. I added some start softness (by moving the triangle) to make the selection more graduated.

3. Adjust Saturation Thresholds. There was no saturation of color in the sky, so I needed to leave the saturation adjustments alone, because adjusting them would only lessen what was selected. If there was some blue in the sky and white clouds, using the Saturation Thresholds could help because blue clouds have some saturation of color, whereas white clouds do not.

4. Soften. As always, I softened the image just a little, using a setting of 2.0.

With a decent selection made, I needed to adjust the second Three-Way Color Corrector to make the sky look like a sky. Because the sky is all highlights, I adjusted the Highlights color wheel and gave the sky a strong blue adjustment, which helped but wasn’t convincing.

Looking at the RGB Parade explains why. The red and green “bright” areas are just too bright. Lowering the Output White level reduced the brightest part of the image, but because I had added blue on the Highlights wheel, the result is a good, natural-looking sky.

The finished version is on your Timeline at marker 6.49 for you to examine.

Adobe includes a prebuilt effect called Leave Color that you can use to apply a single color effect. But the advantage of doing it with a Three-Way Color Corrector is that you take advantage of the full Mercury Engine acceleration as well as more subtle controls.

The technique is to use a Three-Way Color Corrector’s secondary feature, select the color you want to keep, desaturate that area, and at the very end invert the mask, causing the effect to flip-flop and desaturate everything but the selected color.

At marker 6.50 a primary correction was built with the first Three-Way Color Corrector, and a second Three-Way Color Corrector was added to adjust the inversion of the selected color. Let’s walk through the steps.

1. Make the initial selection. This time, I relied on using the eyedroppers because the yellow of the image was very distinct. It was quick and easy to use the plus eyedropper to add darker and lighter yellows, because there was no other yellow in the image.

2. Adjust Luma Thresholds. By selecting Show Mask, luma was given a broad adjustment, selecting darker yellows and brighter yellows.

3. Adjust Saturation Thresholds. What was still missing was some of the less saturated and more saturated yellows. I adjusted the Saturation selections much wider, which really picked up on the details.

4. Soften. As always, I softened the selection just a little, using a setting of 2.0.

With a decent selection created, I then needed to open the saturation controls of the Three-Way Color Corrector and lower all four controls: the Master, Shadows, Midtones, and Highlights, desaturating everything.

I then used the Invert switch at the bottom of the effect to reverse the key by setting it to Invert Mask (Figure 6.51), which desaturated the rest of the image, leaving just the color yellow of the sugar packet.

For example, currently, it’s trendy to darken the edges of an image to give it a vignette. The reason it’s used so often is that it attracts information to “bright” parts of an image and mimics the way your vision falls off to the edges of an image when you focus on something.

Because there isn’t a built-in plug-in, I needed to build a stack of a specific recipe. When I use a shape-based corrections technique, I have to copy and paste a video from V1 onto V2, so the same video is on two layers. Then, I need to add a third “top” layer that includes some form of black-and-white shape that I would create to cut the video on V2 onto itself on V1.

This is very much a recipe; if you get the stack right, it works.

At marker 6.52, take a look at the example of a vignette in Figure 6.52:

V3 Shape + Blur. The top layer needs to be a shape. It can be built anywhere, such as with the Title tool or in Adobe Photoshop. On the Timeline, I built it using the Title tool. It also needs a blur to soften the edges, so I used the Gaussian blur.

V2 copy of the video + Track Matte. The middle layer was a brighter version of the color correction done on V1. It needs an additional effect called the Track Matte Key. This effect uses the top layer to cut the middle layer over the background. In the Track Matte effect, you must choose V3 in the Matte drop-down menu. This cuts the V2 track, showing only V2 in the white areas from the V3 shape—in this case the circle.

V1 copy of the video. The bottom layer is a darker version of the color correction.

Building this recipe is solely about getting the recipe right.

Here are some suggestions when you’re building shape-based corrections:

Don’t set the Gaussian blur on the V3 Shape until the very end of positioning and building the stack. You need to be able to differentiate where the shape is cutting. You know you have the right value when the transition from the foreground to background is seamless.

Your eyes adapt quickly: Toggling the effect on and off seems dramatic, but after your eyes adapt, it looks natural.

You may have to move (keyframe) the shape if there’s movement in the frame. One of the advantages of the very soft edge (heavy blur) is that the movement is very subtle.

Toggling back and forth between shots quickly (using the up and down arrow keys) is a great way to compare shots. If you need to do this for a frame other than the head frame of a shot, add markers (press M) and toggle using the Go to Next Marker (Shift+M) and Previous Marker (Shift+Command+M [Shift+Ctrl+M]) keys. Another way to work is to place the sequence you’re working on in the Source Monitor. The sequence will act as reference with its own separate playhead.

Here are some general guidelines for practical shot matching:

Match tonal range. Try to get a similar tonal range between the shots first: the black point, the white point, and most important, the midtone/Gamma. Look for landmarks like where the highlights of flesh tones appear in the Waveform scope.

Identify vectors and saturation. Look at the Vectorscope, and compare the saturations, especially.

Use RGB Parade. Like color grading, the RGB Parade is the most important scope. Try to find matchable elements in both images.

Limit your vision I. Cheat by temporarily surrounding a color you want to match using a Crop effect or a Garbage Matte effect on both images. These methods of cutting an image may help to isolate the image to a single color, and then use the RGB Parade to compare them.

Limit your vision II. When you’re trying to get two shots from very different moments, it may be worth temporarily moving one shot to V2 and adding a left or right side crop to see both shots simultaneously to compare them in the scopes.

Eliminate shadow pollution. Often, when you’re struggling to match two shots, the reason is that there’s some saturation of color in the shadows. Consider desaturating the shadows.

Use Secondary Color Correction. Yes, you can use the Secondary section of a Three-Way Color Corrector or RGB Curves to isolate and fix part of an image that doesn’t match, but then you’ll just be adjusting a small portion of the total shot. It’s part of the arsenal of ways to correct, but don’t go to it first.

Avoid noise. Watch for noise, particularly in poorly shot (consumer camera) material. In these cases you may have to rely on some type of noise reduction effect.

On the Timeline in 7 shot matching at marker 6.53 are two shots (Figure 6.53), both of a model walking outside. The first shot looks pretty normal (but shots can always use some primary color correction), but the second shot is very dark (exposure difference) and has some blue coloration, most likely because the light was fading at the end of day!

A basic correction was done to both shots at marker 6.54 (Figure 6.54). Almost no exposure change was done with the shot CU model walking outside, and a little warmth was added to the midtones. The MS model walking shot had lots of exposure change, brightening the shot. Constant comparison was done with the Waveforms to try to match the luminance of the faces. Saturation was added, which helped visually in getting the overall chroma to match each other, but pushing too much warmth led to the outfit becoming somewhat orange.

A secondary color correction was needed to warm up just the flesh tones (Figure 6.55). By adding a second Three-Way Color Corrector and using the keyer for the just face, extra warmth was added in the midtones and the Master Saturation was increased.

If these shots recur on the Timeline, it’d be smart to reuse them, which leads to learning how to save and reuse effects.

Copy and paste. Copy and paste one or several effects via the Paste Attributes command (Option+Command+V [Alt+Ctrl+V]).

Smart Preset bins. Create a Preset bin called Color looks. All your color corrections should live in that bin. Using the Three-Way Color Corrector to prebuild a look that is already slightly cool or warm can give you a better starting point than just using the default Three-Way Color Corrector.

Better camera starting points. Build a correction for the neutral point for each camera. It’s faster to start with a neutral point than with the default effect. This is particularly valuable on multicam shoots; once built, you’ll have a starting effect where each camera matches each other better.

Export presets. Export your presets and keep them organized outside of Adobe Premiere Pro. On the off chance you need to fully reset the software, this export will act as a backup of your customizations. Right-click on the preset/preset bin in the Effects panel to export or import the presets.

Multiple versions. If you’ll be working with a client, it’s a smart idea to build several alternate looks on a shot. Just disable the ones you don’t like. Make sure you right-click on the name of the effect and give it a description.

Should the series resemble a warm environment like Miami? How can you visually depict London as different from Chicago? Creating a “look” to represent a location provides a visual method of communicating location to the audience viewing the video. Similarly, warm looks are often called romantic. Cooler looks are frequently used for more serious scenes. Technology looks typically have some green in the shadows. Placing one of these looks over an entire scene (rather than just a shot) adds a subtle, consistent visual treatment to a scene.

This S curve looks like Figure 6.56 at marker 6.56 from the sequence 8 Looks. Based on the exposure of the shot, it may be necessary to go in and tweak the curve to make it stronger or weaker.

Often, the black and white information doesn’t produce an image with strong contrast. Adobe Photoshop users know this, so instead they use the Black and White adjustment, which permits the output to source from just the Red, Green, or Blue channels rather than the Luma channel.

The Channel Mixer can be set to Monochrome to create a black and white image utilizing only the Red, Green, or Blue channels (or a mixture of two.)

On the Timeline at marker 6.57 are four versions of a shot: one built from the Black and White effect and one built per channel with the Channel Mixer (Figure 6.57). Note that the red controls are the only controls functioning when the Channel Mixer is set to Monochrome. Quickly move from shot to shot to see the difference.

To build the shots at marker 6.58 (Figure 6.58), I used the Tint effect and took some careful adjustments of the white point to a Hue of 25 degrees (orange, exactly halfway between red and yellow) and a Saturation of 40% (more than that was too strong).

Then the S curve from earlier was added to give it a more contrasty “pop.”

The next shot on the Timeline after the Sepia shot (Figure 6.58) was built with a Three-Way Color Corrector with an increased contrast in the Input levels, reducing the saturation across the Master, Shadows, Midtones, and Highlights Saturations, and finally adding some blue to the midtones and shadows.

Whereas Bleach Bypass leaves the image washed out, a Summer look in an image of a bright summer day would have some extra green in the midtones (needing a node in the middle of the Green curve of the RGB effect) and some extra orange in the brightest areas (adding some Red in the highlights and reducing blue in the highlights). The advantage of using curves is in the ease of adjusting just part of the highlights and midtones.

A Three-Way Color Corrector was added to boost the Midtones Saturation and reduce the Shadows Saturation to avoid getting too much color in the blacks.

Simulating a Gradient filter required the technique used in the section “Shape-based Corrections.” The V3 track had a black to white gradient and was adjusted and rotated to cover the area of the sky. The shot is the second to last on the Timeline. The “orange” version of the shot is on V2 along with a Track Matte Key effect. The normal version of the clip is on V1. The image now looks like the sun is warming up the sky (Figure 6.59).

If you’re not familiar with it, Adobe SpeedGrade is a full-color correction environment. But frankly, the interface is daunting for novices. Instead, Adobe has leveraged the Lumetri Engine from Adobe SpeedGrade into a single effect, permitting the application of Adobe SpeedGrade looks directly inside Adobe Premiere Pro.

Instead of using the Lumetri effect in the Color Correction Bin when you want to add Lumetri Looks, skip to the Lumetri Looks section (Figure 6.60) at the bottom of the Effects panel, because the effects in these categories already have a preset attached to them.

Instead, it’s better to use one of two methods to accomplish the same goal with a single effect: adjustment layers or nesting.

An adjustment layer was created in the project and applied to a higher video track, as shown in Figure 6.61 on track V2. By applying the effect to the adjustment layer, Adobe Premiere Pro “adjusts” the appearance of all the clips beneath that layer.

The adjustment layer is easy to turn on or off globally by turning on or off the Track output on the Timeline. You can do this by right-clicking on the adjustment layer and choosing Enable, or by pressing Shift+Command+E (Shift+Ctrl+E).

Using the Enable command, it’s also possible to build several scenic looks on several adjustment layers and enable only the one that you want to apply.

To use nesting, select the shots and choose Clip > Nest. A new sequence representing the clip(s) appears in the Project panel, and the clip(s) are replaced by the new sequence.

Adjustment layers are preferred when applying looks, because they can be easily added, removed, or modified in length, whereas nesting can’t.

Nesting’s major advantage is in collapsing multiple clips and/or tracks into a single element (such as collapsing the shape-based secondary corrections mentioned earlier in this chapter).

Legalization has no bearing on web delivery, because the web is very much the Wild West with no set rules.

What can further complicate legalizing video is that some broadcasters will run your work through a hardware legalizer (which will compress the way it looks, guaranteeing it’ll be safe but also ruining your hard work). Some broadcasters publish detailed specifications, such as the Public Broadcasting System in the United States, which has specific guidelines available for download called the Red Book (www.pbs.org/producing/red-book).

Even though legalizing video mostly applies to broadcasting, it can apply to other tools in the production chain (such as online hosting or video compression headaches) when you exceed the broadcasting standards for where you live.

In essence, it’s better to try to keep your video broadcast safe rather than not.

The Video Limiter effect works by using a “Smart Limit” of luma plus chroma across all the ranges—Shadows, Midtones, and Highlights.

If you have different signal needs, you can adjust the effect’s job of limiting the signal as needed for output.

When you apply the Lumetri effect (from the Color Correction category of effects, not the Lumetri Looks section), you’ll be faced with a File Chooser dialog. Navigate and select either Legalize NTSC (US) or Legalize PAL (Europe) to take advantage of these presets that are included with the project downloads for this chapter.

The original workflows of Adobe Premiere Pro and Adobe SpeedGrade were a little rocky, but Adobe has really listened. The developers seem to have approached the problem with fresh eyes and really got the workflow right, meeting users needs.

Now, Adobe SpeedGrade can read and write into an Adobe Premiere Pro project. Any clip that is adjusted in Adobe SpeedGrade gets a single Lumetri effect in Adobe Premiere Pro.

Unfortunately, this short section in this book cannot truly do Adobe SpeedGrade justice. It provides a minimalist crash course, a “connect the dots” to what’s come prior in this chapter, and it’s far from comprehensive.

Adobe SpeedGrade has a very different interface (Figure 6.63) than any of the other Adobe Creative Cloud applications. This section offers a great “try it” quick start to the application.

This back-and-forth software switching process is, on the whole, a little slower than working solely in Adobe Premiere Pro. But if you decide to learn Adobe SpeedGrade on a deeper level, you will eventually be rewarded by being able do complex color corrections faster. After all, Adobe SpeedGrade is dedicated to performing this task. Yet it won’t be as fast as adjusting a single shot directly in Adobe Premiere Pro.

Prepare your sequence. It’s probably a good practice to duplicate your sequence prior to doing anything else. If needed, you can “undo” any changes you make. Additionally, you might want to consider stripping any Three-Way Color Corrector effects from your sequence. If you’ve used any sort of stylistic effect, such as a Tint effect, decide if you want to keep it or rebuild it in Adobe SpeedGrade.

Send to Adobe SpeedGrade. Select any sequence in Adobe Premiere Pro, choose File > Directly Link to Adobe SpeedGrade. Adobe Premiere Pro will close because both applications can’t directly work on the sequence simultaneously.

Perform color corrections in Adobe SpeedGrade. This is where you’d adjust your sequence. I’ll cover the equivalents of the Three-Way Color Corrector as it relates to Adobe SpeedGrade and shot matching.

Send back to Adobe Premiere Pro. When you’re finished, you just press a button in Adobe SpeedGrade that will close down the project and reopen it in Adobe Premiere Pro. Every clip that was adjusted will now have a Lumetri effect, because that’s the engine behind Adobe SpeedGrade.

The following sections contain essential pieces of information for Adobe SpeedGrade to actually be useful to you.

The top of the screen displays your footage. A Timeline is underneath it along with what looks like a Three-Way Color Corrector at the bottom.

Because the only real effect in Adobe SpeedGrade is color grading, when you adjust the color wheels, you’ll be making corrections to your footage.

To the left of the color wheels is a Layers stack showing the various adjustments that may be layered on a clip.

The panel on the right has a couple of tabs across the top that are meant for two uses, saving specific frames for reuse (Snapshots) and building masks that can be used for a secondary correction.

The most important button is Direct Link to Adobe Premiere Pro because you’ll use it to send the Timeline back to Adobe Premiere Pro. You won’t need to click Undo, Redo, or Save, because they’re all the common keyboard commands you already know: Command+Z (Ctrl+Z), Command+Shift+Z (Ctrl+Shift+Z), and Command+S (Ctrl+S).

In SpeedGrade the playhead controller (Figure 6.66) is below the clips. You can grab it at the large rectangle with the time-code on it. Optionally, you can click where you want it to appear above the Timeline. Notice the number “1” on it. That will become important later when we deal with shot matching. What you want to avoid is clicking on a clip directly.

To play the clip forward and backward, you can use the JKL keys, but the up arrow key will also play forward, and the down arrow key will play backward.

The left and right arrow keys play forward and back a single frame.

To move forward or backward a single clip, press Command (Ctrl) plus the left or right arrow keys.

The Home and End keys go to the In point and Out point.

Table 6.2 shows a list of some common keyboard shortcuts that are immediately valuable as you jump into Adobe SpeedGrade.

To zoom using the mouse, press Option (Alt) and the scroll wheel (a two-finger gesture on a trackpad).

Commonly, editors like to use zoom to Fit. But I prefer 50%, so the screen is being perfectly scaled rather than some odd scaling like 57%.

If you have a numeric keypad, the plus and minus keys zoom in and out.

To zoom using the mouse, press Command (Ctrl) and the scroll wheel (a two-finger gesture on a trackpad).

A great toggle is to press Shift+H, which shows the viewer full screen and hides the interface. Either pressing the shortcut again or pressing Esc will return you to the Adobe SpeedGrade interface.

Press the A key to reveal the Analysis panel shown in Figure 6.71. It’s also the small triangle at the top next to the left edge of the viewer. At the bottom you can choose between one and four scopes. Three is the amount you need, but there’s nothing wrong with also seeing the Histogram scope, although it’s generally the least-used scope.

Right-click on any scope section to choose which scope you prefer to place in that quadrant.

Some colorists prefer this method, even though it’s a little confusing at first. But the advantage is that it’s possible to see a color cast by looking at a neutral item and then trying to adjust that section of the Waveform so the Red, Green, and Blue traces are equal, which makes it turn white (Figure 6.72).

If you want to turn off RGB, just right-click on the Waveform and choose Switch to Luma. Then it behaves like a Luma only Waveform monitor.

The Waveform scope doesn’t read in IRE; instead, it has percentages on the left edge and a numerical setting based on the color space pop-up menu in the bottom-right corner. I suggest you change it to Float (which keeps the highest quality). Then 0 becomes Black and 1 becomes White.

This is a good reason to use the NTSC or PAL safe color Lumetri filter preset in Adobe Premiere Pro that’s available with this chapter’s media in a folder called Extras. The two .look files were saved from Adobe SpeedGrade.

A Layers panel on the left shows the Primary layer (Figure 6.74). Yes, you can have more than one layer as well as specialized secondary layers, but for this minimal guide to SpeedGrade we’ll just focus on this one layer.

Note the Reset button. It’s the fastest way to reset an entire look. A trashcan is also available to delete a layer that is inactive when there is only a single layer.

Traditionally, three-way corrections (in any tool) have a luma range where each wheel—usually called Shadows, Midtones, and Highlights—works. Adobe SpeedGrade instead calls them Offset (Shadows), Gamma (Midtones), and Gain (Highlights) as in Figure 6.75.

Luma adjustments are performed by clicking on the triangle (Figure 6.76) that sits around any of the edges of each color wheel. The adjustment order is Offset (black point), Gain (white point), and then Gamma.

To adjust luma, click the triangle icon on a wheel and drag the mouse left or right in a straight line. Don’t drag the mouse in a circle. The farther left or right you drag, the larger the change.

To reset the luma adjustment of any wheel, click the triangle in the lower-left corner.

Let’s try a luma adjustment.

1. In Adobe Premiere Pro, duplicate the sequence called 11 Direct Link to Adobe SpeedGrade. Rename it to whatever you like, and then choose File > Direct Link.

2. Move to the third shot (the first shot of an actor). Above the Timeline click the All button and then the 6s button to make sure you’re zoomed in.

3. Press A for the scopes. Adjust the scopes in such a way that they feel comfortable. For these sections I tend to work with a one scope layout for each section, but normally I work with a three scope layout. Remember that you can right-click on any scope and reassign it or reconfigure it if needed.

4. In the viewer, press the Toggle In/Out button; an in/out mark should appear around the clip on the Timeline. From this point on, the best navigation method to use is to press Command (Ctrl) left/right arrows to move from shot to shot.

Although this is a fairly full-range shot, my intention is to have you try using the controls.

5. Adjust the Offset triangle and crush some of the darker areas.

6. Adjust the Gain triangle. Go a little too far until the bright part of the actor’s neck hits 100 and then back off until it looks good to you.

7. Adjust the Gamma triangle. Try making the shot bright and then try making it a little darker and moodier. You may need to readjust the Offset/Gain wheels based on your Gamma changes.

Just like color grading in Adobe Premiere Pro, color grading in SpeedGrade is a combination of technique and “feel.” And if the result looks terrible on your (calibrated) system, reset and try it again.

To make a quick adjustment, click anywhere in the circle, and drag to adjust the chroma. For a virtual trackball, right-click in the wheel and move the mouse in the direction of the adjustment; then right-click again to accept the adjustment.

First make adjustments in the Gamma color wheel, paying attention to the Vectorscope. Then, examine the RGB Parade and adjust the Gain color wheel and the Offset color wheel.

Let’s try an adjustment.

Press Command (Ctrl) right arrow and move down the Timeline until you’re at the third to last shot, which is called Road.

Go through the Luma adjustments first by adjusting the Offset, Gain, and then the Gamma triangles.

Based on the Vectorscope and the RGB Parade, the midtones have some extra blue in them. Using the right-click method, click in the middle of the Gamma color wheel, moving away from blue (toward yellow). The road becomes neutralized.

This results in (potentially) nine ranges for minute adjustments. And because you’re already familiar with how each wheel works, you can dial in adjustments precisely as needed.

If you want a challenging shot to work with to try the extra ranges, the shot prior to the Road shot, called 00020.mp4, has a severe red cast. Try removing it.

Here’s a quick overview of the global controls:

Input Saturation. This adjusts the saturation before anything else happens on a shot. You’ll tend to use it if a shot is too monotone at the start.

Final Saturation. Whereas the Input Saturation works before any adjustments have been made, this adjusts the saturation after you’ve performed a color correction. You’d use this when you’re happy with your grade, but you want to pull back or push all the colors in a shot.

Pivot + Contrast. Yes, I’ve grouped these together. Contrast adjusts the difference between the lightest and darkest portions of the image. When you increase contrast, brighter pixels move toward white and darker pixels move toward black. The question becomes where should the midpoint be? At 50%? The pivot is the midpoint (default of .5). You usually contrast to make bright areas brighter and dark areas darker, and if you’re unhappy with which pixels become brighter or darker, you adjust the pivot. Generally, it’s better to make a gamma adjustment than a Contrast + Pivot because gamma leaves black/white sections alone.

Temperature. This allows you to quickly cool down (add blue) or warm up (add orange tones) a shot.

Magenta. You can probably skip this control, but this slider is the opposite of the Temperature control. Whereas the Temperature slider goes from Blue to Orange, this goes from Magenta to Green. If you’re seeing too much magenta or green in your images, there’s usually a white balance problem.

You’ll use the same recipe that you used in the “Three-Way Color Corrector” section earlier in this chapter.

1. Look at the Waveform and adjust the Offset (Black point). Then adjust the Gain (White point). The Gamma is subjective.

2. Looking at the Vectorscope look for the blob to include the center. Adjust the Gamma color wheel to accommodate for that. For the midrange, this is where most of a color cast tends to be visible.

3. Check the RGB Parade. Adjust the Offset wheel to balance/even the amount of color in RGB at the bottom/darkest and the top/brightest areas of the scopes. (Remember that as you approach black or white, you lose chroma.)

4. Look for neutral elements in a shot and make sure they’re balanced in the RGB Parade.

Yes, you could use the hover plus pressing C method. But a great example of how well Adobe SpeedGrade is thought out is by demonstrating the Shot/Reverse technique. A full keyboard (with a numeric keypad) is required for this. You must use the plus and minus and number keys on the keypad.

Set Adobe SpeedGrade to copy from prior shots by pressing Option+ - (Alt+ -). (The opposite, Option++ (Alt++) copies from the next shot.)

Press Option (Alt) and a number key to copy the look from the prior/next shot based on the number you press. So, in a Shot/Reverse shot situation where you’ve graded the first appearance of an actor, you’d move to the next appearance (two shots later) by pressing Option+2 (Alt+2)and be done.

If you tried this and ended up with a second playhead, you didn’t use the numeric keypad. See the section “Adding a second playhead” to find out how to set your system to a single playhead again.

Color match always matches the first playhead to the second playhead.

I prefer the dual screen interface where the second viewer appears on the left side of your screen. The viewer acts as a reference and stays in place, allowing you to quickly match several shots to the same reference shot.

The fastest way to create a second playhead is by pressing Option+2 (Alt+2). To go back to one playhead, press Option+1 (Alt+1).

If you want the second playhead to travel with the first, press the stop/play button on the edge of playhead 2 (Figure 6.81).

If you want to try using AutoColorMatch with the project, you’ll find the two shots you matched earlier in the section “Practical Shot Matching” are the last two shots on the Timeline.

Some topics were just too deep to cover and keep this section as a “hit the ground running” look at Adobe SpeedGrade. You’ll need to investigate the following areas on your own:

Looks. Adobe SpeedGrade offers you the ability to save looks and recall looks from its built-in library.

Layers. The layers stack on the left side of the screen permits multiple primary and secondary layers. This also allows for variations when clients aren’t sure they want a cool or a warm look.

Secondaries. Secondaries in Adobe SpeedGrade adjust just portions of the image based on a keyed selection. For example, you might increase the saturation of a sky or grass with a secondary layer.

Masks. A mask adjusts part of a layer based on a circular or rectangular shape. Darkening the outside of an oval to create a vignette would be a classic use of a mask.

Tracking. When masking, it’s common to need to track a region onscreen so a mask follows something moving in the image (such as a person walking through the screen). This way a vignette would follow someone on screen.

Keyframes. Lighting situations change. It’s useful to have your color adjustments vary on a single shot (similar to keyframes in Adobe Premiere Pro or Adobe After Effects).

Snapshots. Snapshots are a great way to save a particular look for comparison.

Adobe SpeedGrade does so much that I can barely scratch the surface in a book of techniques for Adobe Premiere Pro. But the workflow made it too valuable of an application to ignore!