Working of a Morpher

Let’s run through a quick example of how Morpher works to begin with and then we will move into facial animation using morphers. We will morph a Cabernet wine glass type to a Burgundy one (see Figure 9-1). Of course there are a lot of other types of glasses, so let’s use this to understand how morphs work.

Fire up 3ds Max and load Morph_Begin.max. (If you get an error about missing the image Wineglasstypes.jpeg, browse to the reference image folder of Chapter 9 and choose the image from the folder to assign it.) I modeled the Cabernet wine glass using the polygon modeling tools. If you right-click on the Cabernet glass that is modeled and go to Object Properties, you should see the vertices and faces count, as shown in Figure 9-2.

Now we need to model the Burgundy wine glass, since we need both shapes for the morph to work. You can model the burgundy following the reference image, as set in the Begin_Morph.max, but you need to remember that both objects should have the same number of vertices. It’s going to be hard to keep track of the number of vertices as we model (it’s possible to model by keeping the vertex count in check for every step of your model, but it is not ideal to do it this way).

The easier and better approach when you want to animate using morphs is to make a copy of the source object (in our case, the Cabernet glass), hold down the Shift key, and move it (using the Move tool) to the position of the Burgundy glass in the reference image. When you let go, you will be prompted with a dialog, as shown in Figure 9-3.

We need a copy of it, so choose Copy. If you need more copies, you can change the number of copies. For now, we need only one. Rename the new copy Burgundy. When you let go, you should have two objects—one named Cabernet and the other named Burgundy. A checkpoint file has been created up until this point called Morph_Begin01.max.

Since we copied/cloned the first object, the second object will share the same vertex count. Now, with the second object aligned to the reference image, select it and go into subobject mode of Editable Mesh ➤ Vertex from the modifier stack. Begin tweaking the vertices to align them to fit as per the burgundy glass in the reference image. See Figure 9-4.

Remember not to use any modeling tools, as they will add/remove vertices based on the tool. See Figure 9-5.

Once you are happy with the settings, you will have two different-looking objects with the same number of vertices and faces, which is a must-have for any object to morph. A checkpoint file has been created called Morph_Begin02.max.

Our intention is to morph Cabernet to Burgundy, so select the Cabernet image and choose the Modify tab in the command panel. Add a Morpher modifier by clicking on the Modifier list and choosing Morpher. See Figure 9-6.

Select the Morpher in the Modifier list and go to the channel list section of the Morpher modifier. The channel list will show the number of morph targets you can add. We will discuss this more in a later section, as we move into facial animations. For now, right-click on the first empty button in the channel list. You should be prompted with the Pick From Scene option, as shown in Figure 9-7.

Click on Pick From Scene and click on the Burgundy Object in the viewport. Once you have chosen the Burgundy object, the channel list should show the first channel with the Burgundy name and a green light, which denotes that there is a morph. See Figure 9-8.

At this point I would like you to use the spinner next to the Burgundy name in channel list and see what happens in the viewport. Your Cabernet glass should morph into a burgundy glass as you increase the value from 0 to 100. You can now turn on Auto Key and change the values as needed to animate the morph.

This is morph animation at a basic level. There are more than just two types of wine glasses, though. By modeling the other kinds using the same technique of clone and tweak and adding them to the subsequent channels in the channel list, you could animate to many targets concurrently—such as a Burgundy and a short glass. By adding both to their respective channels, you could even animate from a Cabernet to a short Burgundy glass (both morphs are applied at the same time!).

Emotions: Facial Expression

When a human character shows any of these expressions, there are a number of muscles involved in contorting the face to show what we see. At this point, I suggest you mimic these expressions in a mirror and note the changes happening in your face, especially when compared to a neutral state when no expression is shown.

Let’s go ahead and create facial animation of emotions with morphs.

I detached the head of the character that we previously used for rigging and deleted the body so that we can focus on the facial features alone. I created a checkpoint file with the head alone called Human_Head.max.

I recommend you find images or take photographs of a person who matches the persona of the character you are trying to animate and use that as a backdrop or reference. I created a checkpoint file with all the tweaked emotions. The reference file is called Emotions.max.

You can see these emotions in Figures 9-10 through 9-12.

In Figure 9-10, apart from the obvious mouth open and closed state, there is a lot happening around the mouth area. Notice the shape of the chin in both states. Next up, Figure 9-11 shows anger and fear.

Much of the tweaks in Figure 9-12 are around the eyes and mouth areas.

Note that fear and disgust have a lot in common, except for the eyes. This should do well for the animation.

If you are unable to find a reference image, this chapter’s images are provided along with the source files in the reference images folder.

If you want to succeed as an animator, observe the motions around you. There is a rhythm to all motions and facial expressions.

Another key point to note is that like in rigging, where we recommend a T pose for the character, in facial animation always have the character’s mouth opened in its default pose. This is because closing the mouth using subobject modes of the geometry is easier than opening it.

A checkpoint file has been created called Emotions.max for you to reference if need be. You can use it to follow along in getting the morph animation for the character.

Your morphs are set! If you have additional morphs, you can load them into subsequent slots. We have set six emotions for the character, with a total of seven, including the neutral state. Your channel list should look like Figure 9-13.

At this point, you can select all the other heads except the neutral one and hide them so that they are not visible in the renderer.

You can turn on AutoKey or use SetKey and tweak the values in the channel list. Keyframes are automatically created when Auto Key is on. To see the keyframes for modifier elements, be sure to check the modifiers in the Keyfilter mode. See Figure 9-14.

A checkpoint file is available for your reference called Emotions_Morph.max. Now you can tweak the values of multiple morphs to get diverse results.

Phonemes

If we want our character to talk, we need to begin setting up morphs for that as well, but how many morphs do we need to set? Maybe 26 for the letters of the alphabet? Alphabet sounds are different when they are pronounced in conjunction with other letters. Even though the English language has only 26 letters, there are actually 44 phonemes.

What exactly are phonemes? To break it down into simpler terms, it’s the sound we create when we say a particular word. We don’t speak by letters but by sounds. For example, the word catch is comprised of three sounds—c (it’s more like ca), a (more like aeh), and tch—as one sound. It would be hard to explain this even further. Words as phonemes need to explained through verbal training or sound files. There are a total of 44 phonemes. Does this mean we need to create morphs for all the 44? It would be best because we would have much more control over the character with 44 morphs, but if you want to simplify this process, you can do away with some of these phonemes. In our case, we will be using nine phonemes.

I provide a file for you to see the morphs created for the phonemes; it’s called Phonemes.max. The phonemes can be created by duplicating the head using the Move, Rotate, and Scale tools with the aid of soft selection for fine tweaks. Again remember not to increase the polygon count, as morphs rely on the polygon count to be the same.

Creating these morph positions is the biggest and cumbersome task. Once the morphs are done, the animation is easy. There are a lot of third-party tools to aid you in creating morphs easily. If you are exporting a mesh from the Autodesk Character Generator, you can export the character with morphs and a preset bone rig that has bone weight.

Audio Lip Sync: MorphTarget

Let’s now use an audio file to create a lip sync. I provided an audio file called Morph_Targets.wav for your reference. You can follow along with this file or with any voiceover files you have to lip sync the character. Be sure you have the audio in the .wav 16-bit pcm format, or your audio might not be supported by 3ds Max. Feel free to use any converter online to convert it to .wav format.

Fire up 3ds Max if you haven’t already and open FacialAnimation_Begin.max. Note that the file has copies of a human head tweaked for the following morphs:

The same process needs to be followed for the other words as per the audio. Look at phonetics and when that sound is pronounced, use the relevant morph to animate the face. There are two ways we can proceed. One is by setting keyframes for key phonemes for the entire sentence for voiceovers and then coming in and filling in between. Or, you can go in a linear process and set all keys at a particular frame before going forward to the next one. I recommend you keep the keyframes and then come back and edit/add as necessary.

1. 1.

   Keep playing a particular part over and over until you are satisfied. Reposition the keyframes as needed based on the audio cues.

    
2. 2.

   Once the keyframes are set, open the mini Curve Editor and on the left pane, choose Object ➤ Phoneme_Rest ➤ Modified Object ➤ Morpher. Select the morph. You should see the graph. Tweak the graphs to your liking using the graph tools and for fine refinement of the animation. See Figure 9-29.

    

I went ahead and made a few tweaks in terms of adding more morph keyframes for phones. I also used blink and other expression morphs. A checkpoint file has been created called FacialAnimation_Begin03.max.

Hybrid : Facial Animation Using Morphs and Bones

Now let’s take the head and make it look around as it talks. We could have created four morphs with the head, making it look up, down, left, and right. We are going to use bones to get this done.

A checkpoint file has been created called FacialAnimation_Complete.max. The Neck_Control can be a child of a complete human rig so that the head moves along with the body.