Crazy Talk Tutorial

Online Help
Free Trial Download
Introducing the Project Page
You may save your creation as a project file to continue modifying whenever you want. Built-in and
custom projects are stored in the Project Gallery for you to easily apply.

: Click to create an empty new project file.

1. New/Open/Save

: Click to browse to and open a desired project file.

: Click to save the current project as a file.

: Click to apply one of the project files from the

Template or Custom tabs.
2. Project Gallery
: Click to add the current project into the Custom tab.

: Click to overwrite the selected project in the gallery

with the current project.

: Playback script.

: Pause playback of script.

3. Playback Control

: Stop playback of script.

: Loop playback of script.

: Adjust volume.
Creating a blank project:

You may click the New Project

button to start a new project with only one model on a
white background with no script present throughout the project.
Creating a new project:
Create a model by loading an image or using one of the images on the right side of your
screen. To use one of the images from the right side of your screen, double click on the
image. To create a model or use a saved image, do the following:

Getting Started - Creating CrazyTalk Models

You can create your own customized CrazyTalk Models using any digital snapshot of yourself, a picture
of your friends, colleagues, family members, i.e. any photo from your digital camera.

Loading Image
There are two ways to load your image:
Import Image: Click to import a jpg, bmp, tga or png file with one or more faces to generate
models.

Camera Capture: Click to take a picture with your web camera. Another panel will open for
capturing an image.

Cropping, Rotating and Enhancing Imported Image

Image Processing is the first step of the wizard. You may also click the Image Processing
button to open the image processing tools and enhance the image after exiting the wizard.

Crop Image: Select the required image area.

/
Rotate CW / Rotate CCW: Rotate the image by 90 degrees in the clockwise or
counter-clockwise direction.

Mirror: Mirror the image on the horizontal axis.

Smart Level: Adjust the color levels of the image automatically.

Color Level: Adjust the brightness, contrast, hue and saturation levels of the image
manually. Use the sliders in the menu box to adjust the values or enter the numerical values for
each parameter in the boxes next to the sliders.

Color Balance: Adjust the color balance manually. Drag the sliders to adjust the Cyan Red, Magenta - Green, and Yellow -Blue levels. Click Highlights to apply these color settings
to emphasize the brighter areas. Click Midtones to apply these color settings to the entire image
with average intensity. Click Shadows to apply these color settings to emphasize the darker
areas in the image.

Face Fitting
The Face Fitting step is next. Click and move the indicators numbered 1, 2, 3, and 4 on the image to
provide the four points for the fitting process as displayed in the sample image. You may also click the
Face Fitting
models.

button to start the Face Fitting step again after exiting the wizard to add or delete

Adjust the four points as close to the correct position as possible to get the best fit for the eyes and
mouth features.
If you merely want to have a single model in the project, please skip the following steps.
To generate multiple models (you may add up to four models):
i.

Click the Add Model

button.

ii.

Drag on the image to select another character on the image. Please note that the top left corner of
the cursor is the base for framing out a model.

iii.

Locate the four feature points again to best fit the new model's eyes and mouth features.

iv.

You may click on the target quadrilateral and click the Delete Model
model.

button to remove the

Click the OK button when you finish adding models and go to the next process.
(3D) Face Orientation
The Face Orientation step displays. Here you may assign different face styles to every model by
clicking the model on the image and then selecting a suitable style.

You may also click the Face Orientation

button to invoke the Face Orientation step again after
exiting the wizard to re-assign different face styles to each model.

Utilize the
tool to rotate the oval wire to best fit the orientation of the face on the image. Do
not worry if the wire does not perfectly fit the face on the image since this panel is only for
deciding the rotation strength of the profile.
Strength: Drag the slider to adjust the movement strength of the profile as the model rotates.
Preview: Toggle this button and move your mouse around to see the effects of the Strength in
the viewport. Toggle it again to leave the Preview mode.

Click the OK button to exit the wizard and go back to the model page with the newly-generated models.
You may click the
change the settings.

, and

buttons on the Model page to re-launch the panels above to

Separating Models from Background

Click the Background Mask Editing

button to modify the area of the image that reveals the
background image, which does not move along with the talking models.
If you load TGA or PNG images with alpha channel, the alpha channel will be automatically converted into
the Background Mask.

Full: Extend the background mask over the complete image area.

Clear: Delete the background mask.

Inverse: Invert the background mask area. The area that was previously defined as the
background mask and the animated area are swapped.

Paint: Paint on the image to enlarge the background area or add more parts to it.

Erase: Paint to erase parts of the background mask.

Brush Size: Drag the slider to adjust the size of the brush tool.
Mask Color: Click and specify the color for the background mask via the color palette.
Opacity: Drag the slider to specify the opacity of the background mask.
Contract/Expand: Drag the slider to contract or expand the background mask.
Blur: Drag the slider to blur the edges of the background mask.

Preview: View the results of your editing.

Fine Tuning Face Fitting in Wireframe Mode
CrazyTalk automatically generates a wire-frame surrounding the face according to the 4 points you set
in the Face Fitting panel. You may click Basic Facial Mode
or Detailed Facial Mode
to
view the wire-frame. Drag the control points or the lines to fit the characters face much more precisely
in these two modes.
Please also use the
locate the points more precisely.

Basic Facial Mode

tools above the viewport to zoom in/out or pan the image to

Detailed Facial Mode

Zoom in for Editing

The wire-frame structure is divided into two parts, Character Frame and Head Frame (includes Skull
Mesh and Hair Mesh).
Head Frame
Character Frame

Skull Mesh

The character frame

decides the range on
the image to generate

The skull mesh is

used to frame out the
skull, eyes, nose and

Hair Mesh

The hair mesh, also

known as Pharaoh
mesh, is used to

a model.

mouth of the model.

frame out objects

attached to the head
Cover the whole
The more proper the such as hair, hat or
body of the character
points are aligned, the long ears (of animals).
on the image with the better the facial
character frame so the animation the model
It prevents the
whole model can be
presents.
image from breaking
transformed in the
and increases the
timeline editor.
realism as the model
rotates.
Defining Custom Background

Click the Background Settings

button to specify a solid color or an image as the background.

Color: Click to specify a solid color as the background.

Background:
o Original Background: Use the background of the original image.
o Color Background: Select this button to enable the solid color as background as specified
in the previous step.
o Import: Browse to your desired image and apply it as the background of your project. You
may also specify the Position of the image in the background.

The original look

Define the background area (blue area)

Scene image imported as the background.

Result

Specifying Virtual Eyes

Click the Eye Settings

the Eyes.

button to specify the settings for

i.

Select one of the eye templates from the Eye Style dropdown list.

ii.

Click Modify. You may then adjust the position and the
color of the Eyeball, Eyelash, Eye Shadow as well as
the Specularity of the eyes.

Before

After

You may also click the Import

button to import custom images (JPG, BMP, TGA, PNG,) with or
without alpha channel as the texture of the character's Eyes (256 pixels x 256 pixels).
Specifying Virtual Teeth and Inner Mouth Color

Click the Mouth Settings

button to specify the settings for the Teeth, Mouth and Lips.

Before

After

You may also click the Import

button to import custom images (JPG, BMP, TGA, PNG,) with or
without alpha channel as the texture of the model's Teeth (256 pixel x 166 pixels).
Assigning Standby Motions

Pick the model and click the Model Motion Settings

button to assign different Standby Motions
to individual models. The models' head will rotate when they are not speaking.

1. Modify the script optionally.

The Script page is the critical phase in CrazyTalk's process of creating talking images. Here you decide
what and how your models talk and specify expressions, facial movements and special effects that go
along with the animated script.
Note: You can download additional text-to-speech voices from Microsoft or browse Reallusion website for
more information.

To record your own script, use the Red Record button on the left
Adding Model Audio

Click the New Script

button to start a new script without any keys. The length of the new
script is 15 seconds by default.
To import WAV/MP3 files, click the Import Audio
button to import a speech file in PCM
WAV/MP3 format. Specify the name and location of the WAV/MP3 file using the dialog box and
click Open to open the file.
To record sound, click the Record Audio

button.

Change to Front Mics if using the

headsets and mics plugged into
the front of the computer.

Click the Record button to start recording. Recording from the specified audio source starts.
To convert text to speech, click the Text-to-Speech

button.

i.

If you have more than one TTS engine installed, you may need to select the engine first.

ii.

Type the text in the editor window.

iii.

From the Voice Mode drop-down list, select the type of voice to be used by the model for
speaking.

iv.

Adjust the voice by using the Volume, Pitch, and Speed sliders to achieve the desired
effect. Click the Preview button to play back the result. Click the OK button when done.

a. New Script: Click this button to add one blank script into the model's
track. The default length of the new script is 15 seconds.
b. Import Audio: Click this button to import a pre-recorded speech files in
PCM WAV/MP3 format.
c. Record Audio: Click this button to record voice via microphone installed
on your system.
d. Text-to-Speech: You may convert typed text into speech with this
feature.
Note: You can download additional text-to-speech voices from Microsoft or
browse the Reallusion website for more information.

Modifying a Talking Script

There are three methods to modify a talking script:
A. Double-click or right-click (select Edit in the right menu) on the target script in the Stage.

B. Select a script in the track and click the Timeline tab in the Stage panel.
C. In the Timeline Editor, select the target script in the script drop-down list.

Adding Expressions to a Talking Script

There are 4 methods to add expressions to a talking script:
A. Emotives Library
Click the Emotives tab to view a set of basic templates such as sad, angry, happy, and more.
Double-click on the desired template to apply it.
You can adjust the strength of the expression on the model's face by using the
slider.
B. Motion Clip Gallery
Drag-and-drop the desired template from the Motion Clip gallery.
(These templates contain expressions and special effects)

To create a custom motion clip, drag to select the desired portion in the Motion Clip track and
click the Group

button.

C. Utilizing Puppeteering Panel

Click the Puppet Control

button to invoke the Puppeteering Panel to create custom motion clips.
A motion clip will be overlapped as the top layer of the Face, Transform, Eyes or Shoulders tracks.
Select a Puppet Profile and full face control, click the Preview button and press the Space bar to
view the result. Click the Record button when you are satisfied with the settings.

Pick a desired profile and click the Record button, then press the Space bar to start recording the
motion clip. Move your mouse to move the model, the movements will be recorded
simultaneously. Press the Space bar again to stop recording.
D. Setting Keys
There are two methods to set keys to individual tracks:
Double click on any empty place of each track to invoke the related dialog box to add desired
template.

Track

Double Click

Voice

Select all tracks.

Lips

Invoke Lip Synching

Description

Select desired template.

dialog box.

Motion Clip

No reaction

Transform

Invoke Key Editor

dialog box.

Face/Face
Strength

Invoke Facial
Expression dialog
box.

Adjust the strength.

Drag in this track to make a range for creating custom

clips.

Group the data from Transform, Face, Face Strength

Eye, Shoulder, SFX and SFX-Swap to convert into on
Motion Clip.

Set transform keys for Models manually or automatica

Select desired template.

Adjust the strength of the expression.

Modify the Facial Feature and Muscles.

Add/Delete customized template.

Set keys for Head, Facial Muscles, Eyes, and Should

manually or automatically.

Specify the Facial Muscles in an Expression key.

Select Image Theme, Image Overlay, and Text SFX

effect.

Type texts as the contents of the text special effects.

Two tracks of SFX for creating more special effects in th

background of the image.

Head
Eye

Invoke Key Editor

dialog box.

Shoulder

SFX/SFXSwap

Invoke the Special

Effect dialog box.

Invoke the Key Editor panel, pick Head, Transform, Eyes or Shoulders in the Key Editor
panel, drag on the sample image to adjust the model, and click the Set Key button to add a key
into the timeline. Check the Auto Key box to update the key automatically instead of having to
press the Set Key button manually.

2. Click the Save Project

button. You are prompted to save the project.

Note:
You may also use the Add button to save the project into the Project Gallery.
You may also save your current project while on the Model or Script page.
Opening a project:

Click the Open Project

button on the Project page.

Double-click on the thumbnail in the Template or Custom gallery. Alternatively, select the
desired project thumbnail and click the Apply

button to open the project.

Double-click on the project file (*.ctp) to open it.

Saving a Project
Please load an image and (optionally) modify the script for the model before saving.
Save on the Project page

1. Select the Custom tab in the project gallery.

2. Pick the target folder.

3. Click the Add

button to save the current project.

Save on any page

1. Click the

button anytime on any page. The Save As Project File dialogue will appear.

2. Specify the name for the file, select the desired folder and click Save.

Note:
If you click the Save Project
button with the Ctrl key pressed down,
CrazyTalk saves the model of each frame in JPEG format, which can decrease the
size of the project file.
Opening and saving (with the Ctrl key pressed down) one project file repetitively will
cause the quality of the model to decrease since CrazyTalk opens the project in BMP
format but saves in JPEG format.

You may create subfolders under the Project or Desktop by clicking the

button in the Project Gallery/Custom tab.

CrazyTalk prompts you whether you want to save the unused script clips in the
disabled character tracks on the stage. If you decide to save them, the clips can be
reused when another model is added. This will cause the file size increase however.

Introducing the Output Page

CrazyTalk now offers the convenient feature of exporting your talking images to multimedia formats
that are compatible with a wide range variety of devices. For instance, you can convert your talking
message to an AVI file with DVD and HD quality, or for uploading to YouTube. You can convert it to
sequenced image (BMP/TGA/PNG) files or a streaming RM (Real Media) file to broadcast on a web server.
Exporting Flash files is also supported so you may output your talking model as a flash animation.

: Playback script.

: Pause playback of script.

1. Playback
Control

: Stop playback of script.

: Loop playback of script.

: Adjust volume.

: Zoom In - Enlarge image to increase clarity.

2. Viewing Tools
: Zoom Out - Reduce magnification to increase the area
of visibility.

: Fit to Window - Resize image to fit it in the viewport.

: Original Size - Display image at original size and

magnification.

: Drag Zoom - Drag a marquee across the image and

enlarge the selected area.

: Output Video/Sequenced Images/Wave Files.

3. Mode Tabs

: Output Widget Files.

: Output Files for Web.

: Upload Files to YouTube.

4. Output Settings

Provides options for specifying the output media format, area

for export, target resolution, etc.