At CES 2024, Waves was able to show multiple solutions that demonstrate its clear leadership in DSP applications for the most diverse range of use cases, including conferencing systems and laptops. Among other solutions, Waves showcased its latest examples of Perceptual Distortion Reduction, which analyzes audio in real-time and delivers well-balanced, consistently louder output with less distortion even from highly constrained sound systems, such as laptop speakers. And even Emmy Award-winning Clarity Vx Pro noise removal plugin can now be experienced in select consumer devices from manufacturers such as Dell.
Transversal to those consumer applications is the suite of Waves Maxx technologies, developed to intelligently maximize the sound quality of all sorts of home and personal devices no matter how challenging the device's design, the device's position, the content being delivered, or the environment surrounding it. Those include popular tools such as psychoacoustic bass enhancement, content adaptive compression, stereo enhancement, and adaptive tuning linked to system volume.
This suite has been evolving in recent years with the MaxxVoice family of algorithms, purpose-built for conferencing, communications and speech recognition. From noise reduction to full-duplex echo cancellation, the MaxxVoice suite evolved to include sophisticated tools for far-field voice pick-up and adaptive microphone optimizations.
This range of Waves solutions is now being taken to whole new levels as Waves introduced its own Neural Networks engine, originally applied to the company's range of studio and post-production plugins, and gradually transposed to consumer applications, where it powers current adaptive and perceptual algorithms. This is being now applied to improve areas such as sound muting during conference calls, where the algorithm softens background noises during pauses in voice input while maintaining just enough ambient noise for conversation continuity.
Now Waves is launching a new noise removal capability that also takes advantage of its proprietary Neural Networks engine. Waves SoloMode AI-based software for smartphones, PCs and other devices is able to recognize the user's unique vocal signature, and intelligently filter out background noise, including other voices.
This new audio solution for consumer devices, draws on the company's advanced signal processing technology enhanced with AI/ML to solve a problem that previous audio filtering products have struggled to master: how to distinguish the user's voice. Smartphones or PCs running Waves' MaxxVoice SoloMode will give users the sense that they are holding a conversation in a room as quiet as a library, no matter how noisy the environment actually is.
"SoloMode uses advanced AI to radically enhance the precision with which we can distinguish the sound that the user wants to hear or project, and to shape the output from speakers or microphones to perfectly fit the sonic environment," says Eitan David, VP and General Manager of the Consumer Division at Waves.
This is a great example of the power of AI to filter out everything but the user's voice. While standard noise-reduction architectures can identify and remove unwanted sounds using AI models that have been trained to recognize common sources of background noise, such as dogs barking, traffic, or wind, those solutions will struggle in situations in which multiple human voices are picked up on the capture side of the audio processing chain, such as in crowds, call centers or public spaces.
SoloMode software enables real-time user voice isolation for best-in-class audio clarity during calls, regardless of the number of voices and other noise sources in the background. When setting up SoloMode for the first time, the user is prompted to create their unique voice signature by recording a standard set of sentences for around 25 seconds of audio. After testing the sample, the SoloMode runs a user-isolation algorithm tuned by AI to recognize the voice signature and distinguish it from all other voices.
The SoloMode feature can be turned on and off as required. This is useful in situations in which multiple people are on one side of a call, since SoloMode will remove any non-registered voices from the rendered audio. If SoloMode detects the continued presence of another potential non-registered user, it will provide a prompt to the user in case the call requires the algorithm to be turned off.
At its CES 2024 hospitality suite, Waves also provided a preview of the power of AI-enabled audio processing algorithms, using the latest advances in generative AI technology to achieve studio-quality vocal reconstructions from demo and archive-quality audio recordings, with full-spectrum frequency restoration. Very impressive.
www.maxx.com
www.waves.com
