Audio Deepfake Detection Benchmark for Brazilian Political Speech
134,400 samples | 40 speakers | 5 regionsParlaSpoof-BR is the first public audio deepfake benchmark targeting Brazilian Portuguese political speech. The dataset is derived from recordings of the Brazilian Chamber of Deputies and includes realistic synthetic utterances generated using state-of-the-art text-to-speech and voice conversion models.
| Label | Subset | Files |
|---|---|---|
| Bona Fide | Original Chamber recordings | 2,000 |
| Enhanced (Resemble Enhance, Demucs, MetricGAN+) | 6,000 | |
| MP3/OGG → WAV variants | 3,200 | |
| Spoof | TTS: OmniVoice, XTTS-v2, Chatterbox, VoxCPM2, Qwen3-TTS | 10,000 |
| Voice Conversion: Seed-VC, kNN-VC, OpenVoice-v2, XVC, EzVC | 10,000 | |
| OmniVoice speech infilling (25%, 50%, 75%, LLM Attack) | 8,000 | |
| Babble noise: three SNR levels (10/15/20 dB) | 60,000 | |
| MP3/OGG → WAV variants | 35,200 |
TTS models synthesize speech from text using a reference voice. Below we show: (1) the source text audio from Beto Pereira, (2) the reference voice from Abilio Brunini, and (3) synthesized outputs from 5 TTS models cloning Abilio's voice to speak Beto's text.
The text content to be synthesized
Voice characteristics to clone
MaskGIT-based architecture with masked infilling
Multilingual (17 languages) zero-shot cloning
Zero-shot voice cloning
Portuguese-optimized synthesis
1.7B parameter multilingual model
VC models transform speaker identity while preserving linguistic content. Below we show: (1) the source audio from Beto Pereira, (2) the reference voice from Abilio Brunini, and (3) converted outputs from 5 VC models transforming Beto's speech to sound like Abilio.
Audio to be converted
Voice characteristics to clone
Self-supervised voice conversion
k-Nearest Neighbors voice conversion
Zero-shot cross-lingual voice cloning
Expressive voice conversion
Easy voice conversion
An LLM (Claude Sonnet 4) identifies semantically critical words to manipulate, selecting targets across categories: NUMBER, ANTONYM, NAME, PHRASE, and NEGATION. OmniVoice's native masked infilling then regenerates only the target segment, conditioned on surrounding acoustic context.
Original: "...diretamente tambem podemos antever no horizonte uma crise economica no Estado com reflexo nacional."
Manipulated: "...diretamente tambem podemos antever no horizonte uma prosperidade economica no Estado com reflexo nacional."
Impact: Transforms negative prediction into positive scenario, completely changing the message
Original: "E se estamos diante de uma crise financeira e de produtividade no campo..."
Manipulated: "E se nao estamos diante de uma crise financeira e de produtividade no campo..."
Impact: Completely denies the existence of the crisis, inverting the political diagnosis
Original: "...venho solicitar ao governo federal que sejam implementadas medidas emergenciais em relacao aos financiamentos agricolas."
Manipulated: "...venho solicitar ao governo federal que sejam implementadas cortes imediatos em relacao aos financiamentos agricolas."
Impact: Transforms request for help into request for resource cuts
Original: "...venho solicitar ao governo federal que sejam implementadas medidas emergenciais..."
Manipulated: "...venho solicitar ao governo federal que sejam suspensas medidas emergenciais..."
Impact: Completely inverts the request from action to inaction
Original: "...eu tenho certeza que nos poderemos ter avancos significativos nas pautas de energia."
Manipulated: "...eu tenho certeza que nos poderemos ter retrocessos significativos nas pautas de energia."
Impact: Completely inverts the positive expectation to negative about agenda progress
Original: "Mas quero aqui tambem dizer aos colegas, como Mato Grossense aqui..."
Manipulated: "Mas quero aqui tambem dizer aos colegas, como Paulista aqui..."
Impact: Changes the declared geographic origin, altering regional context
Original: "...ele que foi autor da emenda das diretas ja no nosso pais, que proporcionou as eleicoes diretas para presidente..."
Manipulated: "...ele que foi autor da emenda das diretas ja no nosso pais, que impediu as eleicoes diretas para presidente..."
Impact: Changes historical role from facilitator to obstructor of direct elections
Original: "...entao eu tenho muito orgulho de como matogrossense homenagear esse cidadao..."
Manipulated: "...entao eu tenho muito vergonha de como matogrossense homenagear esse cidadao..."
Impact: Transforms praise into severe criticism of the honoree
Resynthesizes contiguous segments at varying percentages (25%, 50%, 75%) without altering semantic content. This isolates the purely acoustic detectability of partial synthesis. Each row shows one utterance with original and manipulated versions.
Synthetic audio (OmniVoice TTS) with babble noise at different SNR levels. Lower SNR means more noise, which can help deepfakes evade detection. At 10dB SNR, DF-Arena shows 22.2% evasion rate.
Original synthetic audio
Minimal background chatter
Noticeable crowd noise
Significant masking - 22.2% evasion
Main conclusion: State-of-the-art detectors are fundamentally unprepared for Brazilian Portuguese political speech. Attack methodology (synthesis model choice, manipulation percentage) dominates over demographics (gender, region) in determining detection success.