Proximity Voice Chat in Photon Fusion 2

This page accompanies the Proximity Voice Chat in Photon Fusion 2 (Fusion Voice Tutorial) video and explains how to add basic proximity voice chat to a Photon Fusion project using the Fusion Voice integration.

Use this page as a reference before extending voice chat with custom mute controls, push-to-talk, voice indicators, or other gameplay-specific voice features.

If you want to learn more about Fusion Voice, take a look at its dedicated docs page here.

Overview

The video builds on the Fusion Essentials sample, where connection, player spawning, and movement are already set up.

The goal is simple: each player speaks through their microphone, and their voice comes from their own networked player object. As players move farther apart, their voices become quieter.

The Fusion Voice integration handles connecting each voice stream to the correct player, so most of the work is installing the integration and configuring the right components — no custom networking code is required.

The video covers:

• installing Photon Voice through the Fusion Hub and setting the Voice App ID;
• adding FusionVoiceClient to keep the Fusion and Voice connections synchronized;
• adding VoiceNetworkObject, Recorder, Speaker, and AudioSource to the player prefab;
• configuring the Recorder to capture and transmit the local microphone;
• using a 3D AudioSource to create the proximity effect;
• testing voice between two clients;
• common mistakes that stop voice chat from working.

Video Timeline

Time	Section
00:00	Introduction
00:16	Initial Project Setup
01:28	Setting Up the Fusion Voice Client
03:24	Setting Up the Recorder Component
05:33	Setting Up the Speaker Component
05:55	Setting Up the Fusion Network Object Component
06:06	Testing Out the Results

Related Links

• Photon Fusion Voice: https://doc.photonengine.com/voice/v2/getting-started/voice-for-fusion
• Photon Dashboard: https://dashboard.photonengine.com/
• Sample used in this video: https://doc.photonengine.com/fusion/v2/technical-samples/fusion-essentials-sample/overview

Installing Photon Voice

Starting with Fusion 2.1, the Voice integration can be installed directly through the Fusion Hub:

1. In Unity, open Tools > Fusion > Fusion Hub.
2. Open the Addons tab.
3. Find Voice and press Install.
4. Once installation finishes, open the Voice tab and complete the setup steps shown there.

Photon Voice requires its own App ID, separate from the Fusion App ID:

1. Go to the Photon Dashboard and create a new Photon Voice application.
2. Copy its App ID.
3. In Unity, open the Photon App Settings used by Fusion and paste the ID into App Id Voice.

App ID	Used For
Fusion App ID	The multiplayer session.
Voice App ID	Voice communication.

Setting Up FusionVoiceClient

Find the object or prefab that contains the NetworkRunner and add FusionVoiceClient to it.

FusionVoiceClient keeps the Fusion and Voice connections synchronized: when Fusion joins a session, the Voice client joins the Voice room; when Fusion disconnects or shuts down, the Voice client also leaves and disconnects.

Configure the following on FusionVoiceClient:

• Enable Use Fusion App Settings so the Voice client uses the Voice App ID and connection settings from Fusion's Photon App Settings.
• Disable Use Primary Recorder. A Primary Recorder creates an outgoing voice stream that is not tied to a specific network object. Since each stream should belong to the player who is speaking, a VoiceNetworkObject prefab is used instead.
• No Speaker Prefab needs to be assigned, since the player prefab will already contain a Speaker.

Setting Up the Player Prefab

Open the networked player prefab from the Essentials sample.

1. On the same object that contains the NetworkObject and a NetworkBehaviour, add VoiceNetworkObject.
2. Create a child object under the player, name it Voice, and position it around the player's head.
3. On the Voice child, add Recorder, Speaker, and AudioSource.

Component	Role
Recorder	Captures the local player's microphone and creates the outgoing audio stream.
Speaker	Receives the incoming audio stream belonging to a remote player.
AudioSource	Plays the audio received by the Speaker.

VoiceNetworkObject connects the Recorder and Speaker to the correct Fusion NetworkObject. Photon uses the NetworkObject ID to match each Recorder with the correct Speaker — Player 1's microphone plays through the Speaker on Player 1's remote character, Player 2's microphone plays through the Speaker on Player 2's remote character. Because the Speaker lives inside the player prefab, the voice moves together with the player.

Configuring the Recorder

Select the Recorder component and configure it:

• Set the input source to Microphone.
• Enable Recording Enabled so the Recorder can capture and process microphone audio.
• Enable Transmit Enabled so that audio is sent into the Voice room.

Both settings must be enabled for the player's voice to be transmitted. The remaining Recorder settings can stay at their defaults for this basic setup.

Creating the Proximity Effect

Select the AudioSource attached to the Speaker and set Spatial Blend to 3D.

Because the AudioSource is inside the player prefab, it follows the player's position. Unity adjusts the volume based on the distance between the remote player's AudioSource and the local player's AudioListener.

Set Min Distance and Max Distance based on the size of the scene:

• inside Min Distance, the voice plays at full volume;
• as the player moves farther away, the voice becomes quieter;
• once the player passes Max Distance, the voice can no longer be heard.

The 3D AudioSource handles the distance calculation, so no custom script is needed for this basic proximity effect.

After updating the prefab, run Fusion > Rebuild Object Table so Fusion recognizes the updated network prefab.

Testing the Sample

1. Start one client and join a Fusion session.
2. Start a second client and join the same session.
3. Speak through the microphone on one client — the other client should hear the voice through the Speaker attached to that player's remote networked object.
4. Move the players close together — the voice should be clear.
5. Move one player farther away — the voice should gradually become quieter.
6. Move the player closer again — the voice should become louder.

When testing two clients on the same computer, use headphones. Otherwise, the microphone may capture the remote voice coming from the speakers and create echo or feedback.

The Recorder also has a Debug Echo Mode. When enabled, the client receives its own transmitted voice stream, which is useful for testing the microphone and voice connection with a single client. Disable it again before testing normally with multiple players.

Key Concepts

Concept	Description
FusionVoiceClient	Keeps the Fusion connection and the Voice room connection synchronized.
VoiceNetworkObject	Connects a Recorder and Speaker to a specific Fusion NetworkObject.
Recorder	Captures the local player's microphone and creates the outgoing voice stream.
Speaker	Receives and plays the incoming voice stream of a remote player.
Use Primary Recorder	Creates an outgoing stream not tied to a network object; disabled for per-player voice.
App Id Voice	The Photon Voice application ID, separate from the Fusion App ID.
Spatial Blend (3D)	Makes an AudioSource's volume depend on distance to the AudioListener.
Rebuild Object Table	Fusion command that must be run after changing a networked prefab.
Debug Echo Mode	Recorder option that plays back the local client's own transmitted voice for testing.

Best Practices

Use this checklist when adding proximity voice chat with Photon Fusion:

• Install Photon Voice through the Fusion Hub Addons tab rather than importing it manually.
• Use separate App IDs for Fusion (gameplay) and Voice (audio).
• Enable Use Fusion App Settings on FusionVoiceClient so Voice reuses Fusion's connection settings.
• Disable Use Primary Recorder when each voice stream should belong to a specific player.
• Add VoiceNetworkObject to the same object as the player's NetworkObject and NetworkBehaviour.
• Keep the Recorder, Speaker, and AudioSource on a dedicated child object positioned near the player's head.
• Enable both Recording Enabled and Transmit Enabled on the Recorder.
• Set AudioSource Spatial Blend to 3D and tune Min Distance / Max Distance to the scene's scale.
• Run Fusion > Rebuild Object Table after modifying the networked player prefab.
• Use headphones when testing multiple clients on the same machine.
• Use Debug Echo Mode for single-client testing only, and disable it before multi-client testing.

Common Pitfalls

Pitfall	Why It Is a Problem
Voice App ID not set in Photon App Settings	The Voice client cannot connect to a Voice room.
FusionVoiceClient missing from the NetworkRunner object	Fusion and Voice connections are not kept in sync.
Use Primary Recorder left enabled	Creates a stream not tied to a network object instead of a per-player stream.
VoiceNetworkObject missing from the player prefab	Voice streams cannot be matched to the correct networked player.
Missing Recorder or Speaker on the prefab	The player cannot transmit or play back voice.
Recording Enabled or Transmit Enabled left off	The microphone is not captured or not sent, so no voice is transmitted.
Speaker without an AudioSource	The Speaker has no way to play the incoming audio.
AudioSource Spatial Blend not set to 3D	Voice does not attenuate with distance, so the proximity effect does not work.
Forgetting to rebuild the Object Table after prefab changes	Fusion may not recognize the updated network prefab.
Testing two clients on one machine without headphones	The microphone can pick up the other client's speaker output and cause feedback.

Summary

Photon Fusion handles the multiplayer gameplay, while Photon Voice handles the microphone audio. FusionVoiceClient keeps the two connections synchronized, joining and leaving the Voice room alongside the Fusion session.

On the player prefab, VoiceNetworkObject connects a Recorder and Speaker to the correct networked player using the NetworkObject ID, so each player's microphone is played back through the Speaker on their own remote character. Setting the Speaker's AudioSource to 3D makes the voice follow the player and grow quieter with distance, without any custom distance-calculation code.

With the Voice App ID configured, FusionVoiceClient set up on the runner, and VoiceNetworkObject, Recorder, Speaker, and AudioSource set up on the player prefab, proximity voice chat works out of the box through the Fusion Voice integration.

Last updated on 16. 09. 2026

Back to top