How to Configure Altered Real-Time Pro as a Virtual Microphone in Discord and OBS for Low-Latency Streaming and VTubing

Live broadcasters, VTubers, and virtual event hosts frequently face a frustrating dilemma when transforming their voices during live streams: traditional cloud-based voice changers introduce noticeable latency (often exceeding 200–400 milliseconds), causing awkward pauses in live conversations and throwing avatar lip-sync out of alignment. Furthermore, cloud streams introduce transmission jitter and privacy concerns.

Altered Real-Time Pro solves this bottleneck by executing AI voice transformation and speaker isolation entirely on local desktop CPUs. Operating as a native system-level virtual audio device (Microphone (Altered Virtual Audio Device)), it morphs vocal timbre, applies expressive Voice Skins, and isolates speech from background noise with algorithmic latency as low as 135 ms.

Here is the complete walkthrough for configuring Altered Real-Time Pro as a virtual microphone, routing it through Discord and OBS Studio, and fine-tuning your stream for pristine, low-latency audio.

1. System Requirements and Hardware Preparation

Because Altered Real-Time Pro performs deep learning inference locally on your CPU rather than offloading to external cloud servers, your computer must meet specific computational baselines.

Minimum vs. Recommended Hardware

Essential Acoustic Rules Before Starting

1. Always use closed-back headphones: Open-back headphones or desktop monitor speakers leak audio back into your physical microphone. In a real-time AI morphing loop, this acoustic bleed creates metallic comb-filtering, robotic stutter, or uncontrollable feedback shrieks.
2. Keep the physical microphone close to your mouth: Position your boom or dynamic microphone 2 to 4 inches from your mouth with an effective pop filter. Clean input signal direct to the capsule delivers the best vocal tracking for the AI model.
3. Use wired audio connections: Bluetooth headsets introduce 100–250 ms of wireless hardware latency on top of any DSP processing. Use a USB interface or wired 3.5mm connection.

2. Setting Up Altered Real-Time Pro Desktop

Before routing audio into Discord or OBS, you must initialize the virtual audio driver and start the real-time processing engine inside Altered Studio.

[Physical Microphone (USB/XLR)]
               │
               ▼
   [Altered Real-Time Pro Engine]
   - Select Voice Skin (1 of 20)
   - Enable Isolate (Advanced / Biometric)
   - Local CPU Inference (~135ms Latency)
               │
               ▼
[Altered Virtual Audio Device (Virtual Mic)]
        ┌──────┴──────┐
        ▼             ▼
  [Discord Voice]   [OBS Studio Stream]

Step 1: Launch and Switch to the Real-Time Interface

1. Open Altered Studio on your desktop.
2. Click the Real-Time tab in the main navigation bar to enter the real-time processing console.
3. Under the input settings, select your Physical Microphone (e.g., Focusrite Scarlett USB, Shure MV7, or Headset Microphone).

Step 2: Choose Your Vocal Model (Voice Skin)

1. Altered Real-Time Pro includes 20 distinct Voice Skins, ranging from diverse genders, vocal textures, character timbres, to specific vocal personas.
2. Select the voice skin that matches your streaming character or VTuber persona.

Step 3: Configure Isolate (Background Noise & Speaker Rejection)

If you stream from a noisy room, near mechanical keyboard switches, or around roommates, enable the Isolate engine:

• Advanced Mode: Studio-grade acoustic suppression that strips ambient room reflections, AC hum, computer fan noise, and background clatter.
• Biometric Mode: Locks onto your unique vocal signature (Voiceprint). Even if another person speaks loudly in the same room or shouts right next to your desk at up to -6 dB Signal-to-Interference-Ratio (SIR), Biometric mode rejects the competing speaker and processes only your voice.

Step 4: Start the Real-Time Engine

Click Start the Engine. Speak into your microphone and verify on the real-time input/output VU meter that your voice is active and modulating smoothly. Altered installs and activates Microphone (Altered Virtual Audio Device) at the OS driver level.

3. Configuring Discord for Voice Chat and Collaboration

When playing multiplayer games or streaming group collaborations on Discord, routing through Altered Real-Time Pro ensures your teammates hear your transformed character voice in real time.

Step-by-Step Discord Setup

1. Launch Discord and click the User Settings (gear icon) located near your avatar in the lower-left corner.
2. Under App Settings, click Voice & Video.
3. Under Input Device, click the drop-down menu and select Altered Virtual Microphone (or Microphone (Altered Virtual Audio Device)).
4. Leave Output Device set to your regular physical headphones.

Crucial Discord Settings to Avoid Audio Clipping

Discord comes with built-in voice processing tools designed for uncompressed human speech. When combined with an AI-transformed virtual microphone, these filters can conflict and cause muffled speech or sudden dropouts:

• Disable Krisp Noise Suppression: Under Noise Suppression, select None or standard noise suppression. Altered's built-in Isolate engine handles denoising with higher precision locally; stacking Krisp on top causes artificial gating and choppy syllable cutoffs.
• Echo Cancellation: If you wear closed-back headphones, toggle off Discord's aggressive Echo Cancellation.
• Automatic Gain Control: Turn off Automatic Gain Control if your voice volume fluctuates unnaturally during animated streaming reactions.
• Input Sensitivity: Toggle off Automatically determine input sensitivity and manually adjust the slider so that whispered speech triggers the gate while idle breathing remains silent.

4. Configuring OBS Studio for Stream Output & VTuber Avatars

In OBS Studio, you want your stream audience to hear your transformed character voice clearly while avoiding audio duplication or desynchronization with your VTuber model.

Method 1: Global Audio Devices (Simplest Setup)

1. Open OBS Studio and navigate to Settings (lower-right Controls dock).
2. Select the Audio tab on the left sidebar.
3. Locate Global Audio Devices > Mic/Auxiliary Audio.
4. Click the dropdown menu and select Altered Virtual Audio Device.
5. Ensure Mic/Auxiliary Audio 2, 3, and 4 remain Disabled so your raw physical microphone is never broadcast simultaneously.
6. Click Apply and OK.

Method 2: Dedicated Audio Input Capture Source (Recommended for Scene Flexibility)

If you switch between your natural voice (e.g., "Just Chatting" intros) and your character voice during gameplay scenes, setting up a per-scene audio source gives you modular control:

1. In your OBS Sources panel, click the + (Add) icon.
2. Select Audio Input Capture.
3. Name the source Altered Voice (Real-Time).
4. In the device selection dialog, choose Altered Virtual Audio Device.
5. Click OK.

5. Synchronizing VTuber Lip-Sync and Avatar Tracking

For VTubers using facial tracking software (such as VTube Studio, Warudo, 3Tene, or Live2D tracking plugins), video tracking latency introduces a subtle timing offset:

• Camera face-tracking and model rendering typically incur 40 ms to 90 ms of visual latency.
• Altered Real-Time Pro has an algorithmic processing latency of around 135 ms on modern CPUs.

If unadjusted, your avatar's mouth may open slightly before or after your voice is heard on stream.

Aligning Audio and Video in OBS

1. In OBS Studio, locate the Audio Mixer panel.
2. Click the three vertical dots (or gear icon) next to your Altered Voice source and select Advanced Audio Properties.
3. Locate the Sync Offset (ms) column:
   • If your avatar's mouth movements lag behind your voice: Add a positive delay (e.g., +40 ms to +80 ms) to your virtual microphone.
   • If the transformed audio arrives slightly behind a webcam capture: Add a positive delay to your webcam or capture card source under its own properties.
4. Record a 15-second test clip in OBS clapping and speaking sharp plosives ("P", "B", "T") to confirm mouth sync alignment before going live.

6. Optimization, Troubleshooting, and Best Practices

Eliminating Robotic Stutter and Audio Dropouts

If your voice crackles, sounds metallic, or skips frames while playing a demanding game:

1. Check CPU Resource Allocation: Real-Time AI inference demands sustained CPU performance. If your game or VTuber tracking maxes out all CPU cores to 100%, Altered's inference thread may starve. Cap game frame rates in your GPU control panel (e.g., lock at 60 FPS or 120 FPS) to leave CPU overhead for real-time DSP.
2. Standardize Sample Rates: Ensure Windows Sound Settings, OBS Studio, and Altered Studio all run at the identical sample rate—strictly 48,000 Hz (48 kHz). Mismatched sample rates (e.g., 44.1 kHz to 48 kHz) force software resampling buffers, creating clicks and cumulative drift.
3. Audio Buffer Settings: In your hardware audio interface control panel (ASIO/Windows WASAPI), use a buffer size between 128 samples and 256 samples. A buffer below 64 samples increases CPU strain and crackling risk, while a buffer above 512 samples introduces unacceptable conversational lag.

Preventing Voice Duplication ("Double Voice")

If stream viewers report hearing a weird echo or "two voices at once":

• Verify OBS Audio Mixer: You must see only one active microphone meter. If you see both your physical mic (e.g., Mic/Aux) and Altered Voice, mute or delete the physical microphone source from the active scene.
• If streaming Discord voice chat with friends: In OBS, make sure your Discord desktop audio capture does not capture your own microphone loopback.

7. Plan and Feature Comparison

Altered Real-Time Pro offers accessible entry tiers alongside professional multi-seat options:

For streamers broadcasting standard 2-hour sessions, the Voice Skins tier provides adequate daily headroom with full access to the character roster. Full-time VTubers doing marathon 4+ hour streams will benefit most from an Unlimited tier, ensuring uninterrupted character continuity throughout the broadcast.

References

• Real-Time Voice Changer | Altered
• How Altered Real-Time Pro Works | Altered
• OBS with Altered Real-Time Pro Voice Changer
• Discord with Altered Real-Time Pro Voice Changer
• Real-Time Voice Isolation & Speaker Isolation | Altered
• Real-Time Pro Pricing | Altered
• Real-Time Pro Frequently Asked Questions | Altered