Sonos is tackling the growing issue of inaudible dialogue in TV and movies with its newly updated Speech Enhancement feature. Launching today via a free software update for the Arc Ultra soundbar, the system offers four adjustable levels of dialogue clarity, including a setting specifically designed for individuals with hearing loss.

Recognizing the challenges posed by rushed production, poor audio mixing, and other factors, Sonos collaborated with the Royal National Institute for Deaf People (RNID) to develop a solution that truly addresses the needs of those struggling to hear on-screen conversations. Matt Benatan, Principal Research Scientist at Sonos, says: “It’s not just practical, it’s emotional. One of the most important aspects of TV and movie watching is the opportunity to bond over cultural and entertainment events that are watched in real time. If one viewer cannot fully hear the dialogue, their ability to enjoy and participate in the moment can be lost.”

“By implementing machine learning into our speech extraction technology, we figured out how to separate dialogue from other sounds in the center channel and clarify speech in real time,” said Harry Jones, Sound Experience Engineer at Sonos. “This lets us draw out just the dialogue at the most needed times, without overly impacting volume or taking away from the holistic cinematic experience.”

The new Speech Enhancement offers four levels: Low for subtle emphasis, Medium for better clarity with balanced sound, High for prominent dialogue with reduced background elements, and Max, a dedicated setting for hearing loss that prioritizes dialogue by further controlling the dynamic range of non-speech elements.

Developed through extensive collaboration with RNID and input from 37 participants with varying hearing abilities, the feature was tested across diverse content for nearly a year. “We wanted to ensure that Speech Enhancement would work for all, even those who might not even realize they have hearing loss,” said Lauren Ward, Lead RNID Researcher. “One in three adults in the UK experience hearing loss, and it is reported that just under one in four adults in the USA do too. This tool has the potential to impact a large number of people.”

Award-winning film sound mixer Chris Jenkins also contributed, bringing studio-level speech extraction techniques to the home while preserving the integrity of other audio elements. He lauded the feature as a “huge step forward” in addressing dialogue challenges. Sonos emphasized the central role of user feedback in the development process, ensuring the AI-powered feature delivers a nuanced and effective solution for clearer TV audio.

Dubizzle has launched “Sell with AI”, a tool designed to be intuitive and user-friendly, enabling users to effortlessly post ads once they upload an image. Based on the image, the AI-based solution automatically suggests key details such as category, price, and other attributes, significantly reducing manual input and saving users time and effort. This streamlined process not only accelerates ad posting but also ensures consistency and high-quality standards, enhancing the efficiency of the moderation process and improving the overall marketplace experience on Dubizzle.

Commenting on this launch, Haider Khan, CEO of Dubizzle and Dubizzle Group MENA and Board Member of the Dubai Chamber of Digital Economy, stated, “At Dubizzle, our mission is to continually elevate the experience for our users, keeping innovation at our core by leveraging the use of the latest tech and AI solutions, build products and features that address key pain points. Our new AI-based posting experience not only automates the ad creation process, but it also simplifies the interactions on our platform, setting a new standard for efficiency and quality. This tool allows us to empower our users—whether individuals or businesses—with a seamless, highly intuitive experience, ensuring they can reach their audience faster and more effectively. We are proud to take this next step in our journey to making Dubizzle the most trusted and efficient marketplace in the region.”

After nearly a year of delays and extensive security refinements, Microsoft is finally releasing its AI-powered Recall feature for all Copilot+ PCs. This groundbreaking yet contentious tool, which periodically captures and indexes user activity through screenshots, debuts alongside two other major AI enhancements: a more intuitive Windows Search with natural language capabilities and a new “Click to Do” function reminiscent of Google’s Circle to Search.

From Delay to Deployment
Recall’s journey to launch has been anything but smooth. Originally intended to ship with Copilot+ PCs in June 2023, the feature was postponed following backlash from security researchers over potential privacy risks. Microsoft initially aimed for an October 2023 preview but pushed it back again to November for further safeguards. Now, after months of development, Recall arrives as an opt-in feature—users must explicitly enable it during setup.

According to Navjot Virk, Microsoft’s Corporate VP of Windows Experiences, Recall addresses a common pain point: “We wanted to solve a universal frustration: picking up where you left off.” The tool organizes snapshots of user activity, enabling searches based on contextual memories rather than exact filenames.

A Feature That Divides Opinion
Early testers have described Recall as both revolutionary and unnerving. While it transforms Windows search by interpreting images and contextual data, it also raises significant privacy concerns by storing vast amounts of user activity. Security researcher Kevin Beaumont, who first sounded the alarm on Recall’s vulnerabilities, recently tested the final version and acknowledged Microsoft’s security upgrades, including database encryption, default sensitive-data filtering, and mandatory opt-in. However, Beaumont notes that content filtering remains inconsistent, and the feature can be unlocked with a simple four-digit PIN—despite Microsoft’s claims that biometric authentication is required.

AI-Powered Search and “Click to Do”
Beyond Recall, Microsoft is supercharging Windows Search with natural language processing, allowing users to find files using conversational queries like “that brown dog photo” instead of precise filenames. The new Click to Do feature, activated with Win + Left Click, enables on-the-fly analysis of text and images, offering quick actions such as summarization or object removal from photos. Currently, text-based features are exclusive to Qualcomm-powered Copilot+ PCs, with support for AMD and Intel models expected later this year.