DSP-based audio cuts out annoying echoes

AMI Semiconductor has an echo cancellation algorithm optimised for the ultra-low-power BelaSigna 200 and recently launched BelaSigna 250 DSP-based audio processing systems.

AMI Semiconductor has an echo cancellation algorithm optimised for the ultra-low-power BelaSigna 200 and recently launched BelaSigna 250 DSP-based audio processing systems.

The algorithm provides high-quality, echo-free voice communication in office telecomms headsets and Bluetooth mono and stereo headsets.

The AMIS echo cancellation algorithm eliminates acoustic echo, a common problem for headset users whereby the signal being transmitted out of the headset speaker is recaptured by the headset microphone.

This results in the person at the other end of the communication link hearing a distracting echo of his or her voice.

The algorithm processes the speaker (received) and microphone (transmitted) signals together and removes any part of the speaker signal that may have been captured by the microphone.

With the ability to cancel the echo, the algorithm allows the speaker and microphone to be designed in close proximity to one another, enabling smaller headset designs.

The algorithm design is also robust in noisy environments and preserves the signal in double-talk situations.

In standard echo cancellers, performance is determined by the algorithm's ability to identify when the headset user talks.

This task becomes difficult in double-talk and in the presence of loud ambient noise, where the echo canceller may become confused and intermittently cancel the user's speech by mistake.

Although the headset manufacturer may have the ability to optimise the echo canceller for specific environments, a tradeoff usually must be made between echo cancellation performance and signal degradation in these challenging environments.

By taking advantage of AMI Semiconductor's patented WOLA filterback coprocessor within the BelaSigna 200 and 250 systems, the echo canceller does not rely on the ability to identify when the headset user talks.

Even in extremely noisy environments, performance is consistent and cutoffs are avoided entirely, resulting in very consistent behaviour across all situations.

Tradeoffs are not necessary, which reduces the amount of headset-specific tuning when optimising the echo canceller for a new headset.

The AMIS algorithm is also flexible enough to be integrated with control functionality in today's advanced headsets, such as an FM radio, LCDs and nonvolatile memories for record/playback features.

"The smaller size and lighter design desired for communication headsets make echo management an increasingly common challenge", said David Coode, Audio Group Manager at AMI Semiconductor.

"This technology not only allows for more intelligible conversations, but also sets the stage for smaller, less intrusive headsets".

"By adding this algorithm to our existing portfolio, we continue to set new standards for communication clarity".

Building on this technology, AMIS has developed the BelaSigna Echo Cancellation Bundle that integrates this new echo cancellation algorithm with powerful noise reduction to deliver exceptional audio performance with ultralow power consumption.

The microphone (transmitted) signal is processed with the spectral noise eliminator (SNE) noise reduction algorithm, achieving between 10 and 14dB of noise reduction in typical listening situations.

Back to top