IT之家 (ITHome) News from October 15: According to Bloomberg, Céline Dion's voice rang out through the Plénitude Arena on the outskirts of Paris; her rendition of the Titanic theme song reached even the spectators in the farthest corners of the giant venue with almost the same clarity as if they were standing beside her on stage.
On the night of this recent show, among the 35,000 spectators in the stands, most eyes were focused more on the pop diva's look, and on her condition as she performed while recovering from a rare neurological disease, than on the dense wall of loudspeakers surrounding the stage. But those black cabinets belong to a sound system developed by the French company L-Acoustics, which is using artificial intelligence to reshape the auditory experience of live events.
Earlier this year, L-Acoustics, a company with a strong reputation in the niche field of high-end audio, launched a brand-new software platform called Source Intelligence. The platform uses machine learning technology,able to separate a singer's vocals from various interfering background sounds in real time. This technology can filter out other instruments, backing vocals, audience noise, and the ambient reverberation mixed within a venue, outputting clear vocals while suppressing the feedback that can easily ruin a performance. According to the company, Source Intelligence can reduce non-vocal signals by up to 40 decibels, a technological advance that has left audio engineers deeply impressed.
Denis Savage, who has worked with Céline Dion for 38 years and served as the front-of-house engineer for this show, said: "It's almost like magic."
He is responsible for using Source Intelligence to debug the Canadian singer's live performances, saying: "When debugging Céline Dion's concert mix, what the audience wants to hear is Céline Dion. You must always ensure that her vocals remain in the most prominent position in the mix. This set of tools from L-Acoustics happens to make that easier."
Live performances are more popular and more high-tech than ever, and multiple audio companies are developing technologies to improve sound quality. Sound systems are a core component of concerts, yet they are often drowned out by noisy crowds and visual effects. Vocal isolation and noise reduction technology has existed for a long time, first pioneered by nightclub DJs. Other audio companies are already using machine learning and neural networks in their software, such as Alpha Labs and Cedar Audio Ltd. German competitor D&B Audiotechnik released a system last month that allows sound engineers to virtually preview the concert sound heard from every seat in a venue.
Products such as Alpha Labs' anti-howling system De-Feedback can be adapted to devices from multiple manufacturers, whereas L-Acoustics has taken a page from Apple's playbook by building a closed ecosystem. Source Intelligence can only be used with L-Acoustics' own sound systems and must be paired with the company's second-generation L-ISA processor, locking customers into its entire hardware and software ecosystem.
L-Acoustics was founded in 1984 by Christian Heil, a physicist passionate about acoustics, with inspiration drawn from a Pink Floyd concert. As the live music industry boomed, the company's order demand rose sharply. L-Acoustics recently completed a full overhaul of the Hollywood Bowl's sound system, and the blockbuster tours of Harry Styles, Bruno Mars, The Weeknd, and Rosalía have all relied on the Source Intelligence platform to support their live sound.
L-Acoustics has long focused on hardware; back in the 90s of the last century, it was this company that drove the popularization of "line array" stacked speaker technology. Now expanding into AI software allows the company to offer a complete, integrated audio solution. Co-CEO Laurent Vaissié said from the company's Nashville office: "Great speakers alone are far from enough—you also need quality audio material and sound processing capabilities."
Vasseur said that L-Acoustics expects revenue to reach 400 million euros next year (IT Home note: approximately 2.998 billion yuan at the current exchange rate), achieving double-digit growth compared with 2026. The privately held company's core business comes from large-scale live performances, while it is also expanding into theaters, sports venues, nightclubs, and some high-end retail settings.
Sound design at the Stade de France stadium presented a series of unique challenges. The venue's stands stretch 13 kilometers in total, can hold up to 45,000 people, was originally built for rugby matches, and was even converted into a swimming pool during the Paris Olympics.
Celine Dion will perform 26 residency shows at the Plénitude Arena, with L-Acoustics' latest-generation L1 loudspeakers flown in the left, center, and right areas of the stage. Compared to the previous generation, the L1 is smaller and lighter, offering better control over sound directionality and coverage. This means Celine Dion's vocals and bass are emitted only by the center array of loudspeakers, making the sound heard throughout the floor area more uniform.
System engineer François Desjardins, who was responsible for the show, said at the central mixing console in the venue after the September 26 performance: "A single sound source ensures stable and consistent sound, avoiding various problems such as crosstalk and echoes."
The sound pressure level on stage area is up to 26 decibels lower than in the audience seating. Savage said: "On stage you can barely hear loud volume." This makes it easier for Céline Dion to hear the actual accompaniment and the cues in her in-ear monitors, rather than the echo of her own voice reflecting back. "It sounds so much cleaner."
Vassie explained that the total cost of the full audio, video, and acoustic system for the Céline Dion concert at Plaines was over 1 million euros (approximately 7.496 million RMB at current exchange rates). The audio budget typically accounts for 10% of production costs for a show, lower than lighting (20%) and video (40%). This professional equipment for Dion's residency shows was procured by the Montreal-based production company Solotech, then delivered to the show's promoter, AEG Presents. In the coming months, Solotech will use this L-Acoustics sound system for other performances; when Dion returns to Paris next year for a new run of ten shows, the equipment will also return to this venue.
The L1 speaker is assembled in Marcoussis, south of Paris, where L-Acoustics' factory and headquarters are located, reflecting the company's French roots. Most components of the core acoustic technology, including some speaker parts, are produced in Germany and other parts of Europe, then shipped to France for final assembly. Vassier said the company currently has 1100 employees, with roughly 20 percent of staff engaged in research and development, and its business has maintained steady growth over the past two decades.
Although there is considerable concern about artificial intelligence within the music industry, Vaslier believes this L-Acoustics software will not take the jobs of sound professionals. "It will not take the jobs of mixing engineers; it only replaces the other traditional tools that engineers would otherwise use."