> For the complete documentation index, see [llms.txt](https://docs.cgpoly.io/lip-sync-documentation/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.cgpoly.io/lip-sync-documentation/quick-start-guides/quick-start-shape-keys.md).

# Quick Start - Shape Keys

{% embed url="<https://youtu.be/eRGC6uxhwKM>" %}

## ⚡ Quick Start: Shape Keys

This guide will help you set up your first lip sync animation using **Blender's Shape Keys**.

Shape Keys mode is ideal for 3D models with no facial Rigs.

### 🎨 What You Need

* A character with at least one visible face
* Shape Keys representing phonemes
* An audio file (e.g., `.wav`, `.mp3`, `.mp4`, etc.)

{% hint style="info" %}
You need to define your own Shape Keys.&#x20;

Sadly, this is a thing Lip Sync cannot automate.
{% endhint %}

***

### 🪄 Step-by-Step Guide

#### Configure Language

* In the `Lip Sync` panel, select your **Language Model**

<figure><img src="/files/WtlsdBbRKwmxpCArgIr8" alt="" width="375"><figcaption><p>Lip Sync works with ~25 languages</p></figcaption></figure>

{% hint style="info" %}
Lip Sync will download your Language Model and store it in cache so you won't have to download it again.

Models are downloaded from <https://alphacephei.com/vosk/models>. Only **small sized models** are downloaded.
{% endhint %}

#### Add Lip Sync to your Model

* Click on **Add Lip Sync**<br>

<figure><img src="/files/ZQ1NcqOQUQBn9BaNiFg0" alt="" width="318"><figcaption></figcaption></figure>

#### Set Up Animation type

<figure><img src="/files/VahB7qLP9hl7pWyLayq8" alt=""><figcaption><p>Character made by Blender Studio &#x3C;3</p></figcaption></figure>

* Ensure that Animation Type is set to **Shape Keys**

#### Viseme Settings

This is the most tedious part.  You have to manually assign **Shape Keys** to **Visemes**.

<figure><img src="/files/7ghFikhUTinGmw85Do55" alt=""><figcaption><p>Use Drop-down lists to assign your Shape Key to corresponding Viseme</p></figcaption></figure>

{% hint style="success" %}
You can use the **same Shape Key** for **different Visemes.** \
It can be useful if you only have a few Shape Keys for your character.
{% endhint %}

{% hint style="info" %}
`sil` viseme is the position for "Silence".&#x20;

`UNK` viseme is the position for unrecognized phonemes.
{% endhint %}

#### Animation Settings

Animation settings enable you to fine tune how Keyframes are inserting.

<figure><img src="/files/bjauIT8z9mTrn5NfHSyE" alt=""><figcaption><p>Fine tune your animation</p></figcaption></figure>

<table><thead><tr><th width="169">Settings</th><th width="365">Description</th><th>Infos</th></tr></thead><tbody><tr><td>Lip Close Duration</td><td>Time required to close your character mouth. It will be used when a silence is detected and insert sil keyframe accordingly</td><td></td></tr><tr><td>In between</td><td><p>Minimum Threshold between two keyframes. </p><p>If a keyframe should be inserted before, it will be ignored.</p><p><br><em>e.g: 417ms means that next keyframe will be at least at 417ms from the previous one.</em> </p></td><td><p>Decreasing this value will ensure you to animate all phonemes but may result in jerky motion. </p><p><br>Increasing this value will give a smoother motion but sync will be less precise.</p></td></tr><tr><td>Silence</td><td>Time to have between 2 words to be considered as a "Silence".</td><td>Too short value will force you character to close their mouth often.<br><br>Too high value will force your character to have longer mouth transition between two words. </td></tr></tbody></table>

{% hint style="warning" %}
Timing is very important and need to respect this constraint:

* `Silence` should always be greater than `Lip Close Duration` and `In between`&#x20;
  {% endhint %}

#### Add Audio

<figure><img src="/files/TkfXm7xbxM4rvgZ2MHaX" alt=""><figcaption></figcaption></figure>

* Add a Video Sequencer Editor
* Drag & Drop (or Add) your media file

{% hint style="info" %}
You should only have one enabled media File. Otherwise speech recognition may product unexpected results.
{% endhint %}

#### Baking

Press `Bake Audio` and wait!

<figure><img src="/files/CZaRkxx53JFaFh2rMz04" alt=""><figcaption><p>Lip Sync inserted 1372 Keyframes in a few seconds!</p></figcaption></figure>

{% hint style="warning" %}
Baking will remove previously inserted keyframes on defined Shape Keys.
{% endhint %}

{% hint style="info" %}
For Shape Keys, Keyframes use a **Linear Interpolation**
{% endhint %}

{% hint style="success" %}
Lip Sync works with **all Frame Rates** (24 fps, 30 fps, 60 fps, etc..)
{% endhint %}

Now your character is talking! 🎬