In a broad spectrum of applications, user experience with streamed audio depends critically on the ability to deliver high fidelity under variable network conditions, as packet loss often results in disruptive audio artifacts and major degradation of perceptual quality. The main obstacle to effective recovery from packet loss in streamed or networked music is due to the polyphonic nature of the signal which makes it notoriously difficult to predict. Current approaches employ ad hoc and crude methods to conceal missing portions of polyphonic music. Instead, InfoCoding Labs’ proprietary technology is optimally tailored to the structure of polyphonic signals and offers the capability to accurately predict or interpolate missing music portions, and thus effectively combat packet loss, ultimately achieving considerably improved reconstruction fidelity at the receiver.
Central to InfoCoding Labs’ packet loss concealment capabilities is the introduction of its revolutionary "cascaded long term prediction" (CLTP) filters, which adapt to the multiple periodic components that constitute the polyphonic music signal at any given time, thus enabling a “divide and concur” technique that separates and predicts each individual periodic component without interference from the others, and recombines the predictions to obtain the best possible estimate of a missing portion of the music signal.
This technology is broadly applicable wherever resilience to audio data loss is required. Typical application scenarios include audio restoration and recovery from damaged media, audio communication to mobile devices and/or over unreliable networks (in various industry sectors including personal, commercial, and defense communications, health monitoring and telemedicine, and more), live and interactive applications (live broadcast with interactive participation of reporters on the scene, teleconferencing, interactive music events/concerts, and a broad spectrum of interactive gaming applications).