Tutorial on using voice AI to create voice modules

In this tutorial, I'll explain how to use speech AI to create a voice mod for Hogwarts Legacy. All 92 original dialogues (including spreadsheets broken down by gender) are available for download in the Files section.

Create a voice module using voice AI

🎤 **Welcome to the world of voice mod creation! ** This tutorial will teach you step by step how to use AI voice technology to create an exclusive voice replacement module for "Hogwarts Legacy". Whether you want to give your game character Hermione's voice or create an entirely new character voice, here are all the resources and tools you need!

---

📚 Resource preparation

**Main file contains content:**

• 🗣️ 92 original female dialogue files
• 🗣️ 92 original male dialogue files
• 📊 Spreadsheet by gender
• 📦 `unrealpak.exe` tool for packaging mods

**Software that needs to be downloaded:**

• 🎵 Wwiser - audio processing tool
• ✂️ Audacity - audio editing software

**OPTIONAL TOOLS:**

• 🔄 **Rename Master** - Rename files in batches, double the efficiency!

---

📖 Tutorial Directory

• **Software Settings** 🛠️
• **Create AI Voice** 🤖
• **Make voice clips** 🎙️
• **Use Audacity to edit and normalize** ✂️
• **Importing and processing in Wwiser** 🎚️
• **Create Mod** 📦

---

🛠️ 1. Software settings

**1.1** Download and unzip the main file of this module page.

**1.2** Download and install all required software (Wwiser and Audacity).

---

🤖 2. Create AI voice

**2.1** I recommend using **Eleven Labs** to generate AI speech. New users have a free quota of 10,000 characters per month. If you plan to make large-scale voice mods, a paid subscription is also very cost-effective!

**2.2** Log in to ElevenLabs and enter “Voice Lab”. You can clone existing sounds, use platform presets, or generate entirely new sounds. In this tutorial we will generate a new sound.

**2.3** Set the sound parameters to your liking (for example: female, young, British accent, accent strength 200%), give it a name and save it. Remember to generate a preview sentence at least once to activate the sound!

---

🎙️ 3. Create voice clips

**3.1** Open the spreadsheet you downloaded (in the Female or Male original folder).

**3.2** Select the dialogue you want to replace. For example: `696024022` corresponds to the dialogue "Likely this Astronomy Table will be best used when it's dark."

**3.3** Copy the dialogue text to the AI ​​you created in ElevenLabs, and generate it repeatedly until you are satisfied. Sometimes adding spaces or punctuation can improve the effect!

**3.4** Download the voice clip and rename it to the original file name you want to replace (e.g. `696024022_Ginny`). It is recommended to add the `_YourName` suffix to distinguish it.

---

✂️ 4. Use Audacity to edit and standardize

**4.1** Open Audacity and drag in the original file and the new voice clip.

**4.2** Trim the tracks to the exact same length, select all tracks, then click **Effects -> Loudness & Compression -> Normalize**, set to **-9dB** and apply to keep the volume consistent.

**4.3** To mute the original audio track, click **File -> Export -> Export as WAV**.

---

🎚️ 5. Import and process in Wwiser

**5.1** Open the Wwise launcher and install the latest version of Wwiser.

**5.2** Start Wwiser and create a new project.

**5.3** Enter **Project -> Project Settings -> Source Settings** and change the format of "Default Dialog Settings" to **Vorbis Quality High**.

**5.4** Click "ShareSets" on the left, find "Conversation Settings -> Default Work Unit -> Default Conversation Settings", and change "Format" from PCM to **Vorbis**.

**5.5** Switch to "Actor-Mixer-Hierarchy" under "Audio" on the left, right-click "Default Work Unit" -> **New Subitem -> Actor Mixer**, name it after your voice name.

**5.6** Click **Project -> Import Audio File**, select the file exported from Audacity, and make sure the target object is your Actor Mixer.

**5.7** Click on your Actor Mixer, enable **"Enable Loudness Normalization"** …