Olivier Armstrong Voice Impression: The Ice Queen of Briggs
An Olivier Armstrong voice impression captures one of anime’s most intimidating presences — the general who rules Briggs Mountain with absolute authority and zero tolerance for weakness. From Fullmetal Alchemist: Brotherhood to fan Discord servers and cosplay events, that commanding, frost-edged female military voice is immediately recognizable and surprisingly achievable with the right audio setup.
This guide covers the acoustic anatomy of Olivier’s voice across both the Japanese and English dubs, DSP settings you can dial in today, training techniques for live impression work, an AI voice cloning workflow for real-time use, and the ethical framework for keeping your creative project responsible.
TL;DR
- Olivier Armstrong’s voice is commanding and low-mid frequency, built on chest resonance, minimal emotion, and deliberate articulation — not on dramatic pitch extremes.
- Japanese VA Yoko Soumi delivers a fuller, more resonant chest placement; English VA Stephanie Young adds a slight forward presence with elongated vowels on commands.
- DSP settings: -2 to -3 semitones pitch shift, -1 to -1.5 semitones formant shift, 800 Hz–1.2 kHz mid boost, high-shelf cut above 6 kHz.
- AI voice cloning adds timbre accuracy beyond what DSP alone can achieve — sub-300ms latency is achievable on modern hardware.
- Ethical use means personal and creative projects only; commercial use requires licensing.
- VoxBooster runs on Windows 10/11 via low-latency audio capture — no kernel driver, safe with anti-cheat games.
Who Is Olivier Mira Armstrong?
Olivier Mira Armstrong is a major character in the Fullmetal Alchemist manga by Hiromu Arakawa and its 2009 anime adaptation FMAB by Bones studio. She commands Fort Briggs, the northern military fortress protecting Amestris from the nation of Drachma in brutal arctic conditions.
Her personality archetype is often labeled “Ice Queen” — not in a dismissive sense, but because her command philosophy genuinely mirrors the environment she controls. She respects strength, punishes weakness, and operates entirely outside the political maneuvering that defines Central Command. Every line she delivers feels like an executive order that has already been decided before she speaks it.
That character philosophy translates directly to acoustic choices: flat affect, low register, no vocal hedging, no warmth. Pulling off the impression is less about hitting a specific pitch and more about eliminating all the softening habits that creep into everyday speech.
The Acoustic Profile of Olivier’s Voice
Fundamental Frequency and Register
Olivier’s voice sits in the low-mid female range — roughly 160 to 210 Hz in calm authoritative speech, occasionally dipping below 160 Hz for emphasis. This is noticeably lower than most female anime characters, who typically center around 220 to 280 Hz. The lower fundamental, combined with prominent chest resonance, gives the voice its physical weight.
Neither performance (Japanese nor English) uses significant vibrato in dramatic moments. Olivier’s emotional peaks are conveyed through increased intensity and volume, not through pitch expressiveness — another inversion of conventional anime female vocal style.
Resonance Placement
The key differentiator between Olivier’s voice and a generic “serious woman” voice is chest resonance. When you speak primarily from chest register rather than the higher, brighter head register, the voice gains low-mid frequency body — the frequencies that make a voice feel physically present in a room.
Yoko Soumi’s performance in the Japanese dub places resonance squarely in the chest, giving the voice a full, slightly dark timbre. Stephanie Young’s English dub performance is technically comparable but uses slightly more forward placement (resonance toward the front of the chest and lower throat rather than deep chest), which adds a sharper edge to command delivery.
Articulation Style
Both performances share aggressive consonant articulation — plosives (P, T, K, B, D, G) are delivered with hard, clean stops. No swallowed consonants. Final consonants are fully pronounced. Vowels in stressed syllables are held long enough to communicate deliberateness; unstressed syllables are clipped short. The overall effect is of someone who calculates each word’s landing point before releasing it.
Uptalk is completely absent. Every declarative statement ends with a slight fundamental drop (the natural falling intonation of a command), even in interrogatives. This single habit — eliminating uptalk — does more for the impression than any pitch-shift setting.
DSP Settings: The Foundation
DSP-based processing gets you into the right territory quickly without requiring model training or GPU hardware. Use these settings as a starting point and adjust to your natural voice’s register.
Pitch Shift
- Target: -2 to -3 semitones
- Rationale: Brings a typical female voice from its natural center into Olivier’s lower register. Male voices with already lower fundamentals may need only -1 semitone or zero shift, compensating through formant and EQ instead.
Formant Shift
- Target: -1 to -1.5 semitones
- Rationale: Pulling formants down slightly enlarges the perceived vocal tract, adding resonant depth without the “robot” quality you get from pitch shift alone. Avoid going below -2 semitones — it creates unnatural darkness that breaks the impression.
EQ
- Low-mid boost: +2 to +3 dB at 800 Hz–1.2 kHz (Q around 1.5). This is the projection frequency — where chest resonance lives and where Olivier’s voice has most of its authority.
- High-shelf cut: -3 to -4 dB above 6 kHz. Reduces sibilance and shrillness that pitch shifting can introduce.
- Sub-bass roll-off: High-pass filter at 80 Hz to remove rumble that adds muddiness without contributing to the impression.
Compression
- Settings: 3:1 ratio, medium attack (15–20 ms), fast release (60–80 ms), threshold set so the compressor engages during louder command delivery.
- Purpose: Projection compression. Olivier’s voice never wavers — a compressor makes quiet syllables punch forward and prevents loud moments from spiking. The result is the “walls of this room belong to me” vocal presence.
Noise Gate
- Threshold: Set just above your room noise floor.
- Why it matters: Olivier’s silences are as deliberate as her words. A gate keeps the channel clean between phrases, which amplifies the impression of controlled authority.
Japanese vs. English Dub: Comparative Settings
| Parameter | Yoko Soumi (JP) | Stephanie Young (EN) |
|---|---|---|
| Pitch shift | -2.5 to -3 semitones | -2 to -2.5 semitones |
| Formant shift | -1.5 semitones | -1 to -1.2 semitones |
| Low-mid boost | +3 dB at 900 Hz | +2 dB at 1.1 kHz |
| High-shelf cut | -4 dB above 5.5 kHz | -3 dB above 6.5 kHz |
| Resonance character | Fuller, darker chest | Forward, slightly sharper |
| Vowel length | Short, clipped | Slightly elongated on commands |
| Vibrato | Essentially none | Essentially none |
Training Drills for Live Impression
Software handles timbre, but delivery is your responsibility. These exercises address the three live performance habits that break an Olivier impression fastest.
1. Uptalk Elimination
Record yourself reading ten declarative sentences. Play them back and mark every instance of rising intonation on the last syllable of a statement. Practice repeating those sentences with a deliberate falling intonation on the final word. After twenty repetitions per sentence, falling-final delivery becomes the default. This is the single highest-leverage drill.
2. Consonant Hardening
Speak a line from the script while placing a finger lightly against your lips. You should feel a clear, distinct tap or burst on each plosive. If the sensation is soft or absent, your consonants are being swallowed. Exaggerate initially — then dial back to natural delivery. Hardened consonants carry across noise and make every word land with physical weight.
3. The Stillness Drill
Read a ten-line command speech and pause completely for two full seconds between each sentence. Do not fill the silence. Train yourself to be comfortable in the authority of the gap. In live roleplay or streaming, this pause is the moment your audience understands that Olivier’s next statement is worth listening to.
4. Emotional Ceiling Practice
Olivier expresses intensity through volume and pace, not through pitch rise or tremor. Practice delivering angry or threatening lines at exactly the same pitch as calm lines — only louder and with shorter syllable durations. Any involuntary pitch rise sounds like a different character entirely.
AI Voice Cloning Workflow for the fma olivier voice mod
AI voice cloning adds the timbre accuracy that DSP alone cannot match — the specific spectral fingerprint of Yoko Soumi or Stephanie Young’s performance.
Step 1: Collect Clean Audio
Gather 10 to 30 minutes of isolated Olivier dialogue. FMAB mixes most scenes with background orchestration, so you need to select moments with minimal underscore. Scene-by-scene isolation or community-curated clean audio packs provide the cleanest source material. Target a variety of emotional registers: calm command, cold contempt, aggressive order, rare soft moment.
Step 2: Prepare the Dataset
Segment the audio into clips of 3 to 15 seconds each. Remove any clips with music bleed, background noise, or other speakers in the frame. Normalize each clip to -16 LUFS to ensure consistent input volume during training. Label clips descriptively if your training tool supports metadata.
Step 3: Train or Import a Model
Train a voice conversion model using your prepared dataset. Community repositories sometimes host pre-trained models for popular anime characters — check before investing training time. Import the model file into VoxBooster via the AI Voice tab.
Step 4: Configure Real-Time Conversion
VoxBooster’s AI voice cloning achieves sub-300ms latency on Windows 10/11 hardware with a modern CPU. For best results:
- Set input to your physical microphone.
- Load the Olivier model in the AI Voice section.
- Apply the DSP chain described above as a post-processing layer.
- Set output to VoxBooster’s virtual microphone device.
- In Discord, OBS, or your game’s audio settings, select the VoxBooster virtual microphone as input.
No Python environment, no command-line setup, no kernel driver installation required.
Step 5: Latency Calibration
With AI conversion active, run a monitoring test: speak into the microphone and listen to the output on headphones. Adjust the buffer size in VoxBooster’s audio settings until latency falls below your comfort threshold for live use. Most users find 250 to 280 ms acceptable for Discord roleplay; streaming voice narration can tolerate higher.
Routing to Discord, OBS, and Games
Once VoxBooster’s virtual microphone is configured, routing is identical across applications.
Discord: Settings → Voice & Video → Input Device → VoxBooster Virtual Microphone. Enable noise suppression at the software level rather than Discord’s built-in Krisp, which can interfere with the processed voice character.
OBS: Add a new Audio Input Capture source, select VoxBooster Virtual Microphone. Set monitoring to Monitor and Output if you want to hear the processed voice while streaming.
Games: In the game’s audio or voice chat settings, select VoxBooster Virtual Microphone as the microphone input. Because VoxBooster uses low-latency audio capture and no kernel driver, it does not trigger anti-cheat detection in titles that use EAC, BattlEye, or Riot Vanguard.
Ethics of AI Voice Impression Work
Using AI-assisted voice impression responsibly requires understanding what the technology does and where the boundaries are.
Personal and fan use — Discord roleplay, cosplay events, fan streams, tabletop RPG sessions, non-monetized content — represents minimal ethical and legal risk. The voice belongs to a fictional character, and the original performances are already freely available in a public-facing commercial product.
The line to watch is commercial use: monetized content where the impression is used to represent the voice actor rather than the character, products sold using the voice, or anything that could be construed as the voice actor endorsing a product or service. These scenarios require explicit licensing from the rights holders and the voice actors’ consent.
AI cloning specifically raises an additional concern: the models trained on Yoko Soumi or Stephanie Young’s performances capture not just Olivier’s character voice but the actor’s own vocal instrument. Using that model to voice content unrelated to FMA — where a listener might mistake it for the actor — moves into territory that affects real people’s livelihoods. Stay within the character context.
Deepfakes and deception: Never use an AI voice model to impersonate a real person in contexts that could mislead listeners. This applies to voice actors, public figures, and anyone else. Fictional character roleplay does not fall into this category.
Soft Gear Reference
For reference if you are building or expanding your setup:
- Microphone: A condenser mic with a cardioid pattern minimizes off-axis room noise. Large-diaphragm condensers pick up the chest resonance range (below 500 Hz) more accurately than most USB headset mics.
- Interface: Any USB audio interface with low-latency ASIO drivers keeps processing overhead minimal.
- Headphones: Closed-back headphones prevent mic bleed from monitor audio; essential for real-time AI conversion where the output is playing simultaneously with your input.
Quick Start Checklist
- Install VoxBooster on Windows 10 or 11.
- Apply DSP chain: -2.5 semitone pitch shift, -1.5 semitone formant shift, 900 Hz mid boost, 6 kHz shelf cut.
- Optional: Import or train an Olivier Armstrong AI voice model and enable AI conversion.
- Set VoxBooster virtual microphone as input in Discord, OBS, or your game.
- Run the uptalk elimination drill until falling intonation is your default.
- Record a test clip of ten command lines and verify consonant hardness and absence of pitch rise on emotional peaks.
- Keep use personal, non-commercial, and clearly fictional.
That covers everything from acoustic fundamentals through live performance technique and AI-assisted real-time conversion. The Ice Queen of Briggs is not an easy impression — she demands deliberateness and discipline — but those same qualities make her one of the most satisfying character voices to master.