What it does
Measures the audio settings of a social media asset against the recommended audio setting of the platform (e.g. TikTok or Instagram Feed).
Why it matters
When designing videos for social media, it's crucial to tailor your content for the specific sound settings on each platform. Depending on the platform, videos may require design considerations for both silent viewing and sound-dependent effectiveness. Here’s why this matters:
-
Viewer Experience: Platforms like Facebook and Instagram typically autoplay videos on mute. Ensure your video’s message is conveyed through visuals, text overlays, or captions so that it is effective even when muted. This improves viewer engagement, as users are more likely to pause and watch when the video is easy to follow without sound.
-
Accessibility: Not everyone can or wants to enable sound—hearing-impaired users or those in quiet environments will appreciate videos that don’t rely solely on audio. Using captions or visual cues ensures that your content is accessible to a broader audience, making your message more inclusive.
-
Engagement Maximization: Videos with captions or text overlays generally perform better in terms of completion rates. Make sure your content is engaging even when watched on mute, encouraging viewers to stick around until the end.
-
Platform-Specific Preferences:
-
Facebook, Instagram, and TikTok: Autoplay videos with sound off—ensure that the message isn’t lost by focusing on strong visuals and subtitles.
-
YouTube: Users expect sound—use it to your advantage by incorporating audio elements like narration, music, or sound effects for a richer experience.
-
Consider both sound-on and sound-off scenarios when creating social media videos. This will maximize reach, ensure accessibility, and create a better user experience, regardless of where your video appears.
How it works
We start by extracting the audio track from the uploaded video file (if available) and feed it into a deep learning-based audio classifier that identifies different types of sound, including "silence." This ensures accurate detection of whether the video is suitable for "sound-off" viewing, even if the video has a valid audio track.
Simultaneously, we analyze the visual components of the video, extracting and assessing on-screen text to ensure the content can effectively convey its message without relying on sound. By combining audio classification and text analysis, we provide a comprehensive evaluation of how well your video performs in both sound-on and sound-off environments.
How to achieve great results
For sound-off platforms like Facebook, include on-screen text to ensure viewers can comprehend the message without sound. However, for sound-on channels like Youtube, prioritize the inclusion of sound as it is more engaging and aligns with user expectations in that context
AI models used
Text extraction is performed by a text recognition model. Audio content is classified by a deep-learning based audio classification model.