Ticker

6/recent/ticker-posts

Sonos finally makes it easier to understand dialogues with AI

Sonos finally makes it easier to understand dialogues with AI

The problem particularly affects the over-50s, more than half of whom suffer from hearing loss, according to UK data. But even with perfect hearing, you sometimes have to strain to hear lines drowned out by sound effects and music.

Four settings for all ears

Gone are the days of the basic on/off system: Sonos now offers four levels of enhancement (Low, Medium, High, and Max). The Max level is specifically targeted at people with hearing loss, while the others are suitable for those who simply want to hear dialogue better in noisy scenes.

“By incorporating machine learning into our speech extraction technology, we’ve discovered how to separate dialogue from other center channel sounds and clarify speech in real time,” explains Harry Jones, an engineer at Sonos. Unlike older techniques that simply boosted vocal frequencies (with often disappointing results), this approach better preserves the overall sound balance.

The artificial intelligence works in real time and only activates when necessary. No dialogue? The system stays on standby. Quiet scene? No need to intervene. But as soon as a conversation risks being masked by explosions or loud music, the AI kicks in.

This technology didn't come out of nowhere. Sonos has been working on it for four years, including a full year of testing in collaboration with the Royal National Institute for Deaf People (RNID). The British association held workshops with people with different types of hearing loss to fine-tune the settings.

The company also enlisted the help of Chris Jenkins, a renowned film sound engineer, to maintain the balance between dialogue clarity and artistic respect. Result: Dolby Atmos effects and low frequencies remain preserved, even with voice enhancement enabled.

During the demonstrations, the first two levels significantly improved understanding without altering the sound ambiance. The High mode begins to make the voice slightly artificial, while the Max level clearly prioritizes intelligibility at the expense of everything else.

This update has been available since May 13, but only on the Arc Ultra. Other Sonos soundbars do not have the necessary computing power. The company has not specified whether its future models will benefit from this technology, but we can reasonably hope so.

Post a Comment

0 Comments