Today Hume is introducing Real World VoiceEQ: a benchmark for measuring the human quality of voice AI across 40+ proprietary and open-source models, 15+ dimensions, and 60+ metrics—developed from more
The Hook
“wall. do you being what beyond most instead by over leading 1 million models the right voice EQ Kairos, tested. build And now,”
Problem Agitation
“The audio track contains extensive spoken dialogue from a primary male narrator, interspersed with short, simulated voice snippets and a few other male voices. The dialogue focuses on the importance of understanding human emotion in voice AI.”
Product Reveal
“The audio track contains extensive spoken dialogue from a primary male narrator, interspersed with short, simulated voice snippets and a few other male voices. The dialogue focuses on the importance of understanding human emotion in voice AI.”
Call to Action
“The audio track contains extensive spoken dialogue from a primary male narrator, interspersed with short, simulated voice snippets and a few other male voices. The dialogue focuses on the importance of understanding human emotion in voice AI.”
Synthesizable in Remotion · key components of the hume-voice-eq-launch-template template.
Short, spoken words appear as animated captions in a rounded pill-shaped background. The text often has a subtle bounce or blur-in effect, sometimes accompanied by a larger, blurred version of the word appearing in the background.
Use spring() for position and scale, interpolate() for opacity and blur. The pill background can be a View with borderRadius. For the background blur, use a Text component with a filter: 'blur(Xpx)' and opacity interpolation.
Soft, evolving gradient backgrounds transition smoothly, often serving as a canvas for animated text or data points. Small, glowing particles or dots float and interact within this gradient space, suggesting data flow or emotional reactions.
Implement with LinearGradient or RadialGradient components, animating colors and stops using interpolate(). For particles, use Canvas or a series of View components with position and opacity animated via spring() or interpolate(), potentially with a filter: 'blur()' for glow.
Text appears as if typed in a terminal, often accompanied by a progress bar or loading indicator. This conveys a sense of real-time evaluation or processing.
Use Sequence and AbsoluteFill for text typing effect (character by character reveal). A View with width animated via interpolate() can create the progress bar. Color changes for progress segments.
The screen divides into two or more distinct sections, displaying contrasting or complementary visuals simultaneously (e.g., live-action on one side, UI on the other, or abstract math vs. practical application).
Use AbsoluteFill with flexDirection: 'row' or 'column' to divide the screen. Each section contains its own Sequence or Composition. Ensure smooth transitions for the split itself, perhaps a wipe or slide effect using interpolate().
Large, bold sans-serif text animates into view, often with a strong, quick motion blur or a subtle spring effect, sometimes against a gradient background. It's used for key statements or numbers.
Animate scale, opacity, and transform: 'translateY' using spring() for a punchy entrance. Apply filter: 'blur()' for motion blur during the animation.
Purpose. Establish the core problem: AI's inability to truly understand human emotion, creating a gap in interaction.
Execution. Rapid cuts between relatable human scenarios (live-action) and abstract digital representations of AI's current limitations (e.g., voice snippets, blurred text, simple UI).
Purpose. Agitate the problem further by showing how current evaluation methods fall short, highlighting the complexity and nuance AI misses.
Execution. Transition to more technical visuals: terminal screens, data points, and UI showing superficial analysis, contrasted with the depth of human emotion.
Purpose. Reveal Hume AI's VoiceEQ as the groundbreaking solution that quantifies and measures the human quality of voice AI.
Execution. Clear, impactful text reveals of 'VoiceEQ' and 'KAIROS' on dynamic gradient backgrounds. Introduction of the evaluation UI, showcasing its comprehensive metrics and models.
Purpose. Detail the extensive research, models, dimensions, and metrics behind VoiceEQ, building credibility and demonstrating its thoroughness.
Execution. Split-screen visuals (e.g., scientific equations alongside UI evaluation), animated data tables, and infographics that visually represent the scale and depth of the benchmark.
Purpose. Illustrate the transformative impact of VoiceEQ, enabling more human-centric and emotionally intelligent AI.
Execution. Uplifting live-action shots combined with aspirational motion graphics, showing the positive outcomes of emotionally intelligent AI.
Purpose. Call to action, inviting developers and companies to leverage VoiceEQ for their voice AI solutions.
Execution. A strong, clean branding sequence with the Hume AI logo and a clear tagline, followed by a call to visit the website or engage.