Voice Changer in REAPER: VST Chain Guide

REAPER is the DAW of choice for podcasters, voice-over artists, and streamers who want surgical control over their audio — and it pairs remarkably well with a real-time voice changer when you get the VST chain order right. Get it wrong and you’ll fight noise, latency spikes, and unpredictable gain staging all session long.

This guide walks the full signal path: from raw microphone input through EQ, gating, and your voice mod plugin, all the way to ReaRoute or low-latency audio capture output for streaming and recording. It also covers which JSFX scripts complement voice modification without adding latency, and how to wire everything together for Discord, OBS, and podcast clients.

TL;DR

Signal order matters: gate → EQ → voice changer → compression, not the other way around.
REAPER’s low-latency audio capture shared-mode input picks up any Windows-compliant virtual mic with no driver install.
ReaRoute lets ASIO-capable apps receive REAPER’s processed output at near-zero latency.
JSFX pitch-correction and ducking scripts add zero cost and integrate cleanly after your voice mod.
VoxBooster’s low-latency audio capture virtual mic is recognised by REAPER as a standard input device — sub-300 ms AI cloning with sub-20 ms DSP, no kernel driver.
A well-tuned REAPER chain solves monitoring delay that plagues direct-monitoring setups.

Why REAPER for Voice Changing?

Most voice changers run as standalone applications — you pick a preset, they process your mic, and they output to a virtual device that your streaming app or chat client selects. That works, but you lose two things: the ability to layer additional DSP on top of the voice mod, and a permanent recording of every session for review or editing.

REAPER gives you both. Because it can accept any low-latency audio capture or ASIO input — including virtual microphone devices — it can sit downstream of a voice changer and apply its own processing chain to the already-modified voice. Alternatively, if your voice changer ships as a VST or VST3 plugin, REAPER hosts it directly in-chain alongside native Cockos plugins.

The result: a voice-changed, EQ’d, compressed, de-essed signal that records to disk, streams to OBS via ReaRoute, and monitors in your headphones — all simultaneously, at DAW-grade quality.

Audio Driver Choice: low-latency audio capture vs ASIO vs ReaRoute

Before building any FX chain, you need to choose the right audio backend in Preferences → Audio → Device.

Backend	Typical Latency	Best For	Notes
low-latency audio capture Shared	10–30 ms	Most voice mod + recording scenarios	Shares device with other apps; easiest setup
low-latency audio capture Exclusive	5–15 ms	Low-latency monitoring, no other apps need mic	Locks device — Discord can’t share the mic
ASIO (hardware driver)	1–10 ms	Pro audio interfaces, ultra-low monitoring	Requires ASIO driver from interface maker
ReaRoute (ASIO)	~1 ms	Sending REAPER output to OBS/second DAW	Cockos virtual ASIO driver, install separately

For the majority of voice-changer-in-REAPER setups — podcast recording, streaming commentary, voice-over work — low-latency audio capture shared mode is the right answer. It lets VoxBooster (or any virtual mic) and REAPER share the same device simultaneously, with latency low enough that monitoring through headphones is comfortable.

Switch to ASIO only if you have a dedicated interface with an ASIO driver and need sub-5 ms round-trip for live performance monitoring.

Setting Up Your Input Device in REAPER

Open Options → Preferences → Audio → Device.
Set Audio system to low-latency audio capture.
Under Input device, click the dropdown. Any low-latency audio capture-compliant virtual microphone — including VoxBooster’s virtual mic — appears here alongside your physical devices.
Set a buffer size of 256 samples at 48000 Hz for a good balance of latency (~5 ms at the buffer level) and stability.
Check Allow input from all devices if you want to use multiple inputs in the same session.
Click Apply and close.

With VoxBooster running, its virtual mic shows up as a standard input device. Select it here and every new track defaults to receiving that device when you arm it for recording.

The VST Chain: Signal Order for Voice Modification

This is where most people go wrong. The correct order depends on what you want the voice changer to react to.

Option A: Voice Changer as a VST Plugin in the FX Chain

If your voice mod ships as a VST/VST3:

Physical mic input
  → ReaGate (noise gate)
  → ReaEQ (high-pass filter at 80 Hz, presence cut at 3–5 kHz if harsh)
  → Voice Changer VST
  → ReaComp (light compression, 3:1, slow attack ~20 ms)
  → ReaEQ second pass (optional polish — roll off above 12 kHz)
  → Output / Send to master

Why gate first? The gate ensures that only intentional speech reaches the voice changer. AI voice processing algorithms are sensitive to room noise — even subtle HVAC hum can cause the formant-shifting algorithm to produce artefacts. A noise gate with a threshold set just below your speaking level (typically −40 to −35 dBFS) solves this cleanly.

Why EQ before the voice changer? A high-pass filter at 80 Hz removes low-frequency rumble that the voice algorithm might interpret as chest resonance and over-exaggerate. The presence cut (if needed) prevents the algorithm from amplifying harsh consonants.

Why compress after? Voice changers — especially pitch-shifting and formant-shifting modes — introduce gain variance. A syllable that was −12 dBFS might come out at −6 dBFS after a +3 semitone pitch shift. Compression after the mod catches these peaks before they clip the recording or surprise your stream listeners.

Option B: Voice Changer as a Virtual Mic Input

If your voice changer runs as a standalone app and exposes a virtual mic (e.g., VoxBooster):

VoxBooster virtual mic input (already voice-changed)
  → ReaGate (tighten any residual noise)
  → ReaEQ (final tone shaping — boost presence at 2 kHz, HP at 100 Hz)
  → ReaComp (3:1, fast attack for speech control)
  → Output / Send to master

In this setup VoxBooster handles the voice transformation — including AI voice cloning at sub-300 ms latency and DSP effects at sub-20 ms — and REAPER adds the final recording-grade polish. This is the cleanest workflow: VoxBooster is responsible for the voice mod, REAPER is responsible for the mix.

Input Monitoring with Low Latency

REAPER’s input monitoring lets you hear your processed voice in real time through your headphones. The catch: if you’re also monitoring through your physical interface, you’ll hear two signals — a direct unprocessed copy from the interface and the delayed processed copy from REAPER. This creates a comb-filtering effect that makes it almost impossible to perform naturally.

Fix it:

Disable direct monitoring on your audio interface (or mute the input channel in your interface’s mixer software).
In REAPER, arm the voice-changer track and click the record-arm button until the green monitor icon appears (this enables input monitoring for that track).
Set Preferences → Audio → Recording → Monitor input to Enabled when armed.
Keep buffer size at 256 samples (low-latency audio capture) — this keeps round-trip latency under 20 ms for monitoring, which is imperceptible.

Now you hear only the REAPER-processed signal in your headphones, with full EQ and compression applied, at low latency.

ReaRoute: Sending REAPER’s Output to OBS or Discord

ReaRoute is Cockos’s ASIO driver. Once installed, REAPER exposes up to 16 stereo ASIO channels that any ASIO-capable application can read. This is the gold-standard method for routing REAPER’s processed voice output directly into OBS, a podcast client, or a second DAW instance.

Setup steps:

Download and install ReaRoute from reaper.fm/reaplugs.
In REAPER, go to Preferences → Audio → Device and set the output device to ReaRoute ASIO.
Add a send from your voice-changer track to Hardware Output: ReaRoute 1/2.
In OBS, add an Audio Input Capture source. Set the device to ReaRoute ASIO Channel 1/2 (or whatever virtual cable the ASIO driver exposes in Windows audio).
In Discord, set your input device to the same ReaRoute output if it appears as a Windows audio device, or use a low-latency audio capture loopback from REAPER’s output channel.

The result: Discord and OBS both receive REAPER’s processed, EQ’d, compressed, voice-changed output — not the raw mic. If you update VoxBooster’s voice preset mid-stream, both apps receive the updated sound instantly.

For low-latency audio capture-only setups (no ReaRoute), use REAPER’s Send to hardware outputs and enable low-latency audio capture loopback on the output device. Then select that loopback as input in Discord or OBS. Slightly more latency than ReaRoute but zero additional software.

JSFX Scripts That Pair Well with Voice Mods

REAPER ships with a large library of JSFX (Jesusonic FX) scripts — zero-cost, zero-latency plugins written in REAPER’s scripting language. Several pair exceptionally well with voice modification.

Pitch correction (JSFX): REAPER includes pitch/ReaTune and community scripts that apply light pitch correction. After a formant-shifting voice mod, stray semitone drift can make the voice sound unsteady. A subtle pitch corrector (speed: slow, range: ±30 cents) smooths this without adding a robotic effect.

Stereo widener: Most voice changers output mono or near-mono. A JSFX stereo widener placed after the voice mod adds width by introducing a short (5–15 ms) Haas delay on one channel. Sounds natural for podcasting; use sparingly for gaming comms where spatial cues matter.

Ducking script: Place a JSFX ducker on your music/game audio track and key it from your voice track’s send. When you speak, background audio ducks automatically. This is a staple of podcast production and works seamlessly with voice-changed input because the keying signal comes from REAPER’s post-processing signal, not the raw mic.

De-esser (ReaFIR in subtract mode): Voice changers sometimes emphasise sibilants (“s”, “t” sounds) when shifting formants upward. ReaFIR in dynamic subtract mode, tuned to 6–10 kHz, tames this without affecting the rest of the spectrum.

Podcast and Streaming Routing: A Complete Signal Map

Here is a full session layout for a solo podcaster or streamer using REAPER + VoxBooster:

[VoxBooster Virtual Mic] ──→ REAPER Track 1 (Voice)
  FX: ReaGate → ReaEQ → ReaComp → ReaFIR de-ess
  Sends: Master Mix, ReaRoute Out 1/2 (for OBS), ReaRoute Out 3/4 (for recorder)

[Game Audio / Music] ──→ REAPER Track 2 (Background)
  FX: JSFX Ducker (keyed from Track 1)
  Sends: Master Mix, ReaRoute Out 1/2 (blend with voice for OBS)

[Master Mix] ──→ REAPER Master Track
  FX: Limiter (-1 dBTP true peak)
  Output: Speakers / Headphones

[ReaRoute Out 1/2] ──→ OBS Audio Input Capture (stream mix)
[ReaRoute Out 3/4] ──→ Podcast recording app (voice only, no game audio)

This architecture means your stream mix (voice + game) and your podcast recording (voice only) come from the same session, with no re-routing required between episodes.

Whisper Transcription: Turning Recordings into Show Notes

Once you’ve recorded a session through REAPER, VoxBooster’s built-in Whisper transcription can generate a transcript of the voice-changed recording. Because VoxBooster processes audio at the Windows audio level, the same Whisper pipeline that transcribes live speech also works on REAPER-recorded WAV files fed back through the VoxBooster audio engine.

For podcast production, this means automatic show notes and chapters with no third-party transcription service needed.

Cockos REAPER Docs and External Resources

The configurations above are tested against REAPER 7.x. For version-specific menus and option names, the official reference is reaper.fm/userguide. The ReaRoute installer and ReaPlugs bundle (which includes ReaEQ, ReaComp, ReaGate, and ReaFIR as standalone VST plugins usable in other DAWs) is at reaper.fm/reaplugs.

For low-latency audio capture technical details — buffer sharing, exclusive mode negotiation, and how Windows audio sessions interact with virtual devices — the [Microsoft low-latency audio capture documentation](https://learn.microsoft.com/en-us/windows/win32/coreaudio/low-latency audio capture) is the definitive reference. For VST3 plugin format specifics relevant to hosting in REAPER, see the Steinberg VST3 developer documentation.

VoxBooster + REAPER: The Practical Combination

VoxBooster is built for Windows 10/11 and exposes a low-latency audio capture virtual microphone that REAPER treats as a standard input device. There is no kernel driver, no ASIO4ALL workaround, and no virtual audio cable to configure. You select the VoxBooster virtual mic in REAPER’s device preferences or on a per-track basis, and it behaves exactly like a physical mic input.

The AI voice cloning pipeline runs at sub-300 ms end-to-end latency; the DSP effects (pitch shift, formant mod, reverb, chorus) run at sub-20 ms. For the REAPER + streaming workflow, this means the voice transformation is effectively imperceptible — the lag you notice is REAPER’s own buffer, not the voice mod.

If you’re running the soundboard alongside voice modification, VoxBooster mixes both into the same virtual output, so REAPER sees a single clean input with voice and sound effects already combined. No secondary track, no complex routing.

For AI voice cloning in a podcast context — voicing a character, maintaining a pseudonymous persona, or producing multi-voice content solo — the REAPER chain above adds the final broadcast-quality polish that turns a good AI voice into a professional recording.

Conclusion

REAPER and a real-time voice changer are a natural combination. REAPER’s flexible routing, native Cockos plugins, JSFX scripting, and ReaRoute ASIO output give you a broadcast-quality signal chain around your voice mod. The key is getting the order right — gate and EQ before the voice changer, compress and de-ess after — and choosing the right audio backend for your use case.

Download VoxBooster and try the low-latency audio capture virtual mic with REAPER. With REAPER’s 60-day free trial and VoxBooster’s pricing starting at $6.99, you can have a full professional voice-changing and recording chain running in under an hour — no kernel drivers, no virtual cables, no compromises on audio quality.

For related reading, see our guides on setting up a voice changer in Discord and the best VST voice effects for streaming.

FAQ

What is the correct VST chain order for a voice changer in REAPER? The recommended signal path is: physical mic input → ReaEQ (HPF + presence cuts) → noise gate → your voice changer virtual mic or VST plugin → ReaComp → output. Gating before the voice changer prevents the algorithm from reacting to room noise, and compression after it controls unpredictable gain from pitch or formant shifting.

How do I get REAPER to recognise a virtual microphone as an input? In REAPER Preferences → Audio → Device, choose low-latency audio capture (shared mode) and refresh the device list. Any low-latency audio capture-compliant virtual audio device — including Windows-native loopback sources — appears as a standard input. No custom ASIO driver is needed for most real-time voice mod scenarios.

What is ReaRoute and when should I use it for voice modification? ReaRoute is Cockos’s ASIO driver that lets other ASIO applications send and receive audio from REAPER. Use it when your recording software, streaming encoder, or a second DAW instance needs REAPER’s processed output — for example, routing a voice-changed track directly into OBS or a podcast recorder at ultra-low latency.

Can JSFX scripts work alongside a voice changer plugin in REAPER? Yes. JSFX scripts run in REAPER’s FX chain like any other plugin. Useful companions include pitch-correction JSFX (tune corrected voice to key), stereo-width expanders placed after mono voice output, and custom ducking scripts that lower background music when you speak — all zero additional licence cost.

What latency should I expect with low-latency audio capture shared mode in REAPER? low-latency audio capture shared mode typically achieves 10–30 ms round-trip on modern hardware at 48 kHz/256-sample buffer. If you need sub-10 ms, switch to low-latency audio capture exclusive or ASIO. For real-time voice monitoring during recording — not live streaming — 30 ms is imperceptible to most speakers and avoids the buffer negotiation overhead of exclusive mode.

Does VoxBooster work as a REAPER input source without an ASIO driver? Yes. VoxBooster exposes a low-latency audio capture virtual microphone that REAPER recognises as a standard Windows audio input in shared mode. You select it in REAPER’s Audio Device preferences or on a per-track basis. No kernel driver, no ASIO4ALL, no virtual cable — just pick the device and record.

How do I route REAPER’s processed voice into Discord or OBS for streaming? The simplest method: run VoxBooster’s virtual output as your mic in Discord or OBS, and use REAPER as a monitoring and recording chain that taps the same source. For direct REAPER-to-app routing, use ReaRoute (ASIO) to send REAPER’s master output to an ASIO-aware virtual cable, then select that cable in Discord/OBS. low-latency audio capture loopback from REAPER’s output device is a simpler fallback.