KUNITACHI, Japan - Jan. 27, 2026 - PRLog -- Brainstorm Co., Ltd. (Headquarters: Kunitachi, Tokyo; CEO: Takayuki Nakamura) today announced the release of VoiceSterize, a macOS application that fully automates noise reduction, audio mastering, and transcription for video and audio content creators, including Podcasters and YouTubers.
Background: Liberating Creators from "Audio Struggles" In recent years, individual content creation for platforms like YouTube and Podcasts has exploded. However, many creators face significant challenges with environmental noise, inconsistent volume levels, and tedious transcription work. Leveraging years of expertise in professional game sound production, Brainstorm Co., Ltd. developed VoiceSterize as a tool that allows anyone to create "broadcast-quality sound" with a single click.
Key Features:
- "SpectralNoiseGate": Protects the voice while surgically removing noise.
- Automated Mastering: Adheres to global broadcast loudness standards.
- Total Privacy: 100% local processing for AI transcription.
- One-Time Purchase: Professional audio tools for just $9.99—no monthly subscription fees.
Technical Deep Dive: Under the Hood For audio engineers and tech enthusiasts, VoiceSterize offers transparency and precision in its signal processing chain:
- Spectral Processing via FFT: Unlike traditional noise gates that simply cut audio below a certain volume threshold, VoiceSterize utilizes Fast Fourier Transform (FFT) to analyze audio in the frequency domain. It distinguishes between the "voice band" and "noise floor" across the spectrum, allowing for the removal of noise without unnatural "pumping" artifacts.
- Broadcast-Standard Loudness Normalization: The application integrates a mastering chain compliant with ITU-R BS.1770 standards. It automatically normalizes audio to -16 LUFS (Integrated), the de facto standard for Podcasts. Furthermore, the built-in True Peak Limiter prevents inter-sample peaks and digital clipping, ensuring broadcast-safe audio levels for any platform.
- Algorithmic Precision, Not AI Generation: Unlike recent "Generative AI" tools that synthesize artificial voices, VoiceSterize relies on SpectralNoiseGate, a pure Digital Signal Processing (DSP) algorithm, for noise reduction. It does not use AI to alter the audio signal. AI is utilized exclusively for the text transcription feature. This approach preserves the speaker's original emotion, intonation, and authenticity without generating fake audio data.
- CoreML Optimized Whisper Engine: The transcription engine is built upon OpenAI's Whisper model but has been heavily optimized for Apple Silicon (M1 through M5) using CoreML. By quantizing the model for the Apple Neural Engine, VoiceSterize achieves high-speed inference with low memory footprint.
Quote "As sound professionals, we know that bad audio ruins good video," said Takayuki Nakamura, CEO of Brainstorm Co., Ltd. "We built VoiceSterize to give every creator a virtual sound engineer on their desktop. You don't need to learn complex EQ settings or pay monthly fees to get professional results."
Availability VoiceSterize is available now on the Mac App Store. $9.99
Product URL:
https://apps.apple.com/us/app/voicesterize/id6754917929?mt=12
Youtube:
https://youtu.be/ngA8-X9Ubco?si=oOIT9BioRbCOcvPM
Web: https://www.nakataka.net/voicesterize-en/
Photos: (Click photo to enlarge)
Source: BRAINSTORM co.,ltd.
Read Full Story - One-Click Pro Audio for Streamers: "VoiceSterize" Automates Noise Reduction & Mastering on Mac | More news from this source
Press release distribution by PRLog
