AudioShake Breaks Real-Time Barrier for AI Dialogue Isolation at 11ms
AudioShake today announced the availability of its ultra-low-latency dialogue-isolation model, Dialogue RT, delivering end-to-end voice isolation at 11ms. It is the first AI dialogue isolation technology to meet the latency requirements for live broadcast production—unlocking real-time workflows that were previously impractical.
The Problem Audio Engineers Have Absorbed for Years
Live broadcast puts microphones in front of crowds, chaotic fields, and noisy press conferences. Crowd noise bleeds into commentary feeds. Stadium PA bleeds into sideline mics. Field reporters broadcast from streets that have no respect for production schedules.
The consequences show up in two places.
First, ASR and captioning accuracy drops significantly when noisy audio is fed into transcription systems—especially in live workflows operating under tight latency constraints.
Second, broadcast engineers spend significant effort balancing dialogue against other elements using compression, EQ, and noise reduction tools that attempt to suppress unwanted sound but cannot fully separate overlapping sources.
Until AudioShake’s Dialogue RT, no tool could fix these problems at the source in real time without introducing additional latency.
Why 11ms Changes the Equation
Broadcast audio engineers typically require processing latency to stay within roughly 10–15ms to avoid perceptible lip-sync issues on live feeds, with some workflows demanding even tighter constraints. Until now, AI-based dialogue isolation tools have operated outside that envelope.
Waves Clarity Vx Pro, an AI dialogue isolation plugin for post-production, introduces approximately 42ms of latency at 48kHz—well above the live broadcast threshold. CEDAR’s hardware DNS products achieve near-zero latency, but via traditional noise suppression—a different and narrower approach that attenuates noise from a mixed signal rather than isolating the dialogue stem itself.
The distinction matters.
Isolation and denoising are not the same operation. Denoising modifies a mixed signal by suppressing unwanted sound. Isolation extracts dialogue as its own signal, removing everything else. The result is a cleaner, more usable stem that gives broadcast engineers direct, real-time control over dialogue and the rest of the mix—including crowd, PA, and ambient sound.
Dialogue RT’s real-time dialogue isolation turns live audio from a mixed signal into a controllable data stream—usable by both humans and machines.
AudioShake achieves real-time, ultra-low-latency audio separation in part by leveraging NVIDIA
GPUs and the TensorRT SDK to optimize AI models for high-speed inference. By utilizing
NVIDIA Dynamo-Triton, AudioShake can efficiently scale AI workloads across cloud and edge
infrastructure like NVIDIA DGX Spark and Blackwell GPUs to deliver instant stem separation
and noise removal.
What This Enables
Dialogue RT’s practical implications extend across the broadcast chain:
A single source of truth for audio.
Instead of managing parallel feeds for production and transcription, teams can isolate dialogue directly from the main feed and route it downstream—eliminating duplicate workflows.
Fewer microphones, less complexity. In many live productions, engineers deploy dozens of microphones to compensate for bleed between crowd, PA, and commentary. Dialogue RT reduces that dependency—enabling cleaner results with fewer inputs and simpler signal chains.
Better captioning and ASR accuracy.
Using Dialogue RT to feed isolated dialogue rather than a noisy mix directly improves transcription performance by removing background interference at the source. This also enables easier downstream dubbing, localization, and international distribution–all from a single live source.
A more hands-off mix, with control where it matters.
Broadcast engineers can set dialogue isolation on the primary feed and avoid constantly managing crowd noise, stadium PA, or unpredictable field conditions. Unlike denoising tools, which require ongoing threshold tuning, Dialogue RT adapts in real time to changing environments—while still giving engineers the ability to dial in or override as needed.
AI-Media, a global leader in live captioning and translation, is among the first broadcast technology companies to integrate Dialogue RT into its production workflows. By running AudioShake's dialogue isolation upstream of its transcription and translation pipeline, AI-Media is able to feed cleaner audio into its LEXI AI captioning engine. With cleaner dialogue inputs, AI-Media can directly improve its accuracy on live feeds where crowd noise and environmental bleed would otherwise degrade output.
“Dialogue RT’s low-latency performance means dialogue isolation is no longer a post-production tool—it’s now part of the live broadcast chain. Once you can isolate dialogue in real time, you’re not just improving audio quality, you’re giving broadcast teams a fundamentally new level of control over their sound.”
— Jessica Powell, CEO, AudioShake
Availability
Dialogue RT is available now via the AudioShake SDK, designed for direct integration into existing broadcast infrastructure and media workflows. The model runs on NVIDIA DGX Spark as well as NVIDIA’s Blackwell architecture family of GPUs.
Get started at dashboard.audioshake.ai or read documentation at developer.audioshake.ai.