This is the bake feature of LipTalk. You need to specify the start and end frames for the animation. Then, press Bake and LipTalk will generate keyframes for the userdatas of the Null-Object. You can only bake if the correct user-datas are present. If you have not yet added user datas, you can add the required user datas using the Add User Data button (2). The correct ones will be added with the correct name. DAZ characters already have some userdatas from the visimes export.
After the baking process is done, you can disable the Realtime checkbox (5). This will make the plugin use the keyframes to control the Pose-Morph sliders.
You can Undo the bake or simply press Delete. Pressing Delete will delete all keyframes on this track.
It is also possible to bake the first transcript for instance and then the second one  and so forth.

You can use the LipTalk tag to make your character's lips move according to the User Datas. The User Datas are created when you press the Add User-Data Button in the LipTalk interface. They have the right names and values for the lip sync. If you have baked the lip-sync the User Datas hold the keys and you can choose between Realtime (5) Mode or keyframe or motion-clip mode. Realtime mode animates the lips based on the transcript or the text if in text-mode (6) in realtime, while keyframe or motion-clip mode uses the baked keyframes animations or motion-clips.

If you need to create poses/visemes for your custom character, you can use the Add Poses Button. This button will automatically generate poses with the correct names and assign them to the Pose Morph tag. This just works if you have linked the Pose Morph tag in the Pose-Morph Tag link field.(7). It creates all necessary poses with the correct naming etc  and you can directly begin to sculpt.
If you have all the poses seperately as geometry then you can also name each mesh according to the visemes for LipTalk and drag them all together into your Pose-Morph Tag. Cinema 4D will give them the correct name and also links them. Then you don't need to press Add Poses. Or you press Add Poses and drag each geometry manually into the link field in the advanced group of the pose.  

Some DAZ Characters older than Genesis8 may export user-datas and morphs with incompatible names like "Vis AA". The Plugin requires "AA" only. You can use the Rename DAZ button to fix this. It will rename the user-datas of the DAZ Character Controller Null Object and the main Pose-Morph tag if linked. (7)
For further explanations about all needed poses and morphs go to the section  "Custom Characters".

Realtime mode and keyframe/motion-clip mode are two options that you can choose by toggling the Realtime checkbox. In Realtime mode, the plugin uses the transcript or the text input (if you are in text-mode (6)) to create the animation. In keyframe/motion-clip mode, the plugin uses the keyframes that are stored in the user-datas of the Null-Object. You can also make your own motion-clips from the keyframes and save them for later use. The plugin can also work with motions that you have already created.

The Text Mode checkbox allows you to switch between audio/transcript mode and text mode. In text mode, you can see some extra GUI Elements that let you enter some text.

To control the lip movements of your character, you need to link a Pose Morph Tag that contains the blendshapes for each viseme. To do this, simply drag and drop the Pose Morph Tag from the Object Manager into the Pose-Morph-Tag link.

To use audio-data as input, you need to provide a transcript file that was generated by the LipTalkAudioConverter. If you specified a path to it, you can load it  by clicking the "Load Transcript" button (9). 

If you click the Load Transcript button it loads the transcript from the path in the Transcript (8) link field. Everytime you want to load or reload a transcript you have to click this button. If you load a scene with LipTalk it automatically loads the transcript. Loading different transcripts by animting this parameter is currently not avaiable. If you want to use different transcripts you have to bake each (1) or use a second LipTalk tag with a different Playback Start and End Range (24).

The Advanced Mode checkbox is only available if you converted the audio file in Advanced Mode using LipTalkAudioConverter. It is only recommended to use Advanced mode if the audio-speech files contain long vowels. These transcripts additionally contain all the information about the amplitudes of the audio file and have a much larger file size. If you have converted in Advanced mode, you can use both modes and toggle between both algorithms in LipTalk. It is an internal algorithm in the LipTalk Plugin called the "Frame-Range-Method" which was developed by 3Deoskill. In Normal Mode, the lips are immediately closed after a phoneme is detected using the Release parameter. In Advanced Mode, the lips remain in this position as long as the internal algorithm still associates the amplitude with the phoneme. If the signal falls below a certain level, the Release parameter kicks in and closes the lips. The Threshold (16) parameter can be used to influence this threshold value, similar to a noise gate. The parameter works dynamically and automatically adjusts to the detected original level. Internally, the audio signal is of course smoothed to avoid irregularities

The Mode menu allows the user to select either the Advanced or the Comic mode. The Advanced mode requires the user to create or export (DAZ) all the necessary visemes for their characters, such as "AA, EE, IY, OW, UW, SH, M, T, F, W". This mode produces the most realistic results. The Comic mode only needs a few viseme poses for cartoon-like characters, such as "AA, EE, OW, M, T". The Custom Characters section explains how to set up these modes.

With the Sync Start value you can determine when your animation should start. Or if your audio is not in sync you can change the value here or in the audio (21)

The Speed value is just available in Text Mode (6). It is a factor. The higher the value the slower the animation. The default value is 1.

The Release parameter is very important and controls how quickly a blendshape returns to zero after being activated by a phoneme. It is similar to the release time in audio effects like hall, compressor etc. For example, when a viseme like "AA" is detected from the transcript, it sets the corresponding blendshape to a certain value. Then, the Release parameter gradually decreases the value to zero over time. The higher the value, the faster the decrease. The Release parameter should be adjusted according to the speed of speech of the characters. If the speech is slow, the Release parameter should be lower, and vice versa. If the Release parameter is too low, it may interfere with another blendshape, such as "EE", because the PoseMorph tag blends both shapes, "AA" and "EE". Therefore, it is important to find the optimal value for the Release parameter. This parameter can be animated. In Advanced Mode(10), this parameter has less effect but is still important.

The Limiter is a feature that controls the audio amplitude range, similar to a brickwall limiter in audio effects and is just available in audio mode. The LipTalkAudioConverter uses the audio amplitude to measure the intensity of a phoneme, which affects the mouth animation. By using a variable percentage for each value, the system creates more realistic and natural animations. For example, a character who speaks louder will have a wider mouth opening than a character who speaks softly. However, some phonemes may have a lower or higher amplitude than others, which can sometimes cause unnatural or inconsistent animations. To avoid this, you can set the Limiter to a value between 0 and 1, which represents the maximum amplitude of the audio. For example, if the lowest phoneme has an amplitude of 0.18 and the highest 0.8, you can set the Limiter to 0.2, which will reduce the amplitude of all phonemes to 0.2 or lower. This will make the animation more smooth and stable. When the Limiter has lower values, it means that it is reducing the signal more. To compensate this, you need to increase the Srength (17) parameter to restore the signal level.

The Threshold parameter is a type of Noise Gate that filters out unwanted sounds. It only works in Advanced Mode (10). It determines the maximum level of sound that is considered as silence. For example, if your character says a long "AA" for one second, LipTalk measures the sound intensity in intervals of 1000 frames. If the sound intensity is lower than the Threshold, it means the "AA" has ended. Then the Release parameter (14) slowly reduces the blendshape.
The Threshold is important because it helps to adjust for different levels of background noise in the audio files. It is very sensitive and may need to be set in the two or even three-digit decimal point range. A 0db signal in Liptalk is at the value 1. The threshold for silence can then be below 0.1. The default value is 0.085

You can adjust the overall Strength or for each visime individually. These parameters are animatable and work in both audio and text mode. The Strength parameter affects all visemes at once. If you want to change the intensity of a specific viseme, you can lower or raise its value or animate it.

Here you can load a link to the audio file corresponding to your transcript. In Cinema 4D, you can insert the sound by clicking the Add Soundtrack button (22).  This is an animatable parameter. Cinema 4D will update the sound automatically if you animate it and there is already a soundtrack.
TIP: If the parameter is not animated and you link to another soundfile and the track is already created it will  only refresh the sound at frame 0. So here you could  remove and re-add the track to avoid going back for refresh.

Here you can mute the soundtrack. This is not animatable.

It shows the name of the soundtrack if the sound is loaded into Cinema 4D.

The Sync Start value lets you control the timing of your audio playback. This value works the same way as the Sync Start (12) for the transcript.

The Add Soundtrack button creates a soundtrack on the Null-Object with the LipTalk tag and shows it in the link field (20)

The Delete button will clear the audio track of the Null-Object

The LipTalk tag has a Playback Start and End option. This lets you control when the tag affects the Pose Morph Tag. You can use this to make different Null-Objects with LipTalk tags for the same Pose Morph Tag. For example, if your character has four dialogues with four audio files, you can make four Null-Objects with LipTalk tags and set the Playback range for each one. This way, they won't interfere with each other. The Playbacks may not overlap, otherwise one LipTalk Tag blocks the other.
You could also create more LipTalk tags on the same Null-Objekt but with 
Don't forget to set also the Sync Start(12) and the Start (21) for  each tag.
Tip: If you use baked motion-clips, you only need one Null-Object. You can mix all four baked motions non-linear.

If you turn on the Text Mode button the interface elements 25-33 appear. Your character is now no longer controlled by a transcript. You are now able to use text to animate the lips of your character. You can choose the language that suits you best from this menu. Currently, 3Deoskill supports five languages: English, Russian, German, French and Italian. Each language has a large vocabulary of words that you can use to design your models. English has around 160000 words, Russian has about 100000 words, German similar to English. French and Italian each have around 300,000 words.. 3Deoskill is constantly working to improve its dictionaries and add more languages in the future.

You can enter your text in this field. Please avoide special characters like  commas, dots etc. Liptalk will automatically remove them.

The "Convert" converts the text into phonemes. but he asks you beforehand so that you don't accidentally overwrite the phoneme parameters you created

The "External" button opens an external text editor, so that you can edit the text more comfortably. It opens the standard editor of your system and shows the converted phonemes. Therefore it saves first the content of the phoneme text field a file in your plugin folder and loads it into the editor. Now you can add your parameters and save the file again.

The "Paste" button adds the phonemes from the special file which was created with the "External" button and  pastes it  into the phonem panel. Only the latest file status is inserted. If you forget to save the file, the old version will be inserted. Please be careful - this button carries out the action immediately and does not ask you beforehand

The "Word Spacing" parameter controls the pause between the words. Actually we tend to connect words fluently , nearly without any pause. But since this is hard to synchronize later with an voice-over,  you can set the pause according to your taste by increasing this value. It is in milliseconds and increases the default value which is 100 ms. 

This panel displays the converted phonemes  and inserts a blank space before the first word. This is necessary for the internal algorithm to synchronize the lip movements.
But this field can do more than just display the phonemes. This field is also used to enter special parameters to controll each of your  phonemes. You can read about that down below in the topic : Usage of the text converter.

The  Parameter Error label shows an error message if you made some mistakes in the syntax for the parameters, you added to the phonemes.  How to add prameters and the parameter syntax is described below.

In this group you can find the Manual Recording settings.  Here you can turn on the manual recorder, which enables you to block the automatic workflow to set manuel keyframes e.g. if you want animate specific noise which was not dedected correctly from the AudioConverter for instance laughters, click noises etc.

Manual ON lets you record manually and stops the automatic workflow. You can animate this parameter. It only works in the playback range you set in Playback Start/End (24)

You can directly manipulate the Pose Morph Tag by adjusting the strength parameters of each viseme. If you would animate the visemes in the Pose Morph tag, LipTalk will overwrite them in automatic mode. Moreover, you will lose the ability to use motions later, which require the keyframes of the User Data of the Null Object.

TIP: To transfer your lip movements to a game engine, you can bake the Pose Morph tag in the dope sheet.

The Record Sliders button creates a keyframe for all visemes at the current frame.

The Reset Sliders button resets all visemes to zero

In this group you will find the settings after baking. This section allows you to control the strengths, even if you have already baked the lips into keyframes or motions. If you want to reuse motions from a library for another character but only need minor strength adjustments, these values give you the opportunity to do so.

This checkbox turns the After Baking on

The strength values. Default value is 1. It is a factor. The baked keyframes will be multiplied by these values. They are all animatable.

With the Record Sliders button you can set keyframes for all strength values at the current frame.

With the Reset Sliders  button you simply set all strength values to it's default value.

Add the Jaw Joint Solver tag to the jaw joint of your characer

Create a PoseMorph Tag on the characters geometry

Add the sliders for the visemes (they can be empty if you just want to move the jaw joint)

Load a transcript into the LipTalk plugin or type in text and link the Pose Morph tag to it.

Drag the Pose Morph tag also into the Jaw Joint Solver tag

The joint will be now animated.

(optional) If you want also the lips to move you have to modify the pose morph poses by adding data to them by sculpting

Here you can drag the Pose Morph tag which is controlled via the LipTalk tag.

The Rotation-Axis sets the direction of the rotation of the joint.

The Strength parameter controls the overall strength of the rotation.

This vector shows the "Initialized" angle.

This "Initialize Angle" button freezes the actual position from that the Jaw Solver Tag starts to work.
The Jaw Solver Tag does nothing until it is not initialized. It is so to say the default position.

The "Free" Button  releases the joint/object to the state before initialization

Here you can adjust the strength  of the rotation for each phoneme. The strength values in the LipTalk tag do also affect this but they also affect the strengths of the poses in the Pose Morph Tag.

The Invert checkbox inverts the rotation of the joint/object