When part of the signal in duplex (simultaneous two-way) communication returns back to the source it is called “echo”. Even small echo may be very annoying and disrupting due to a significant roundtrip delay in long distance analog and almost all digital communication systems.
Acoustic echo occurs in a voice communication terminal as a result of acoustic coupling between the speaker and microphone. Due to non-linear processing applied in communication channels, such as lossy vocoders and transcoding, acoustic echo must be treated (cancelled) locally inside the device.
- Small, low power and often overdriven speakers produce high level of non-linear distortions
- They are often used in very noisy conditions
- Limited processing delay allowed
- Limited computational resources (MIPS and memory) due to cost and power consumption constrains
- Highest grade voice quality and seamless duplex in capable voice terminals
- Flexible and easily configurable to work optimally in all possible applications and environments
- Capable to withstand problems associated with pure acoustic design and components’ quality
- Small processing delay
- Modest requirements of computational resources on most platforms (DSP, MCU, host CPU)
- Supporting wide dynamic signal range on 16 bit processors
- Supporting narrow band (NB, 4 KHz), wide band (WB, 8 KHz), super-wide band (SWB, 12 KHz) and ultra-wide band (UWB, 16 KHz) speech
- Supporting stereo communication channels (stereo echo cancellation)
- Strictest ITU and ETSI specifications compliant (including automotive specifications)
Use control buttons on the interactive screenshots to play/stop/navigate
Raw signal, as recorded by the microphone (containing echo and user's speech):
Processed signal after AEC - the echo has been eliminated, the user's speech is preserved:
Mobile phones, Bluetooth headsets, automotive hands-free kits, video/audio conferencing equipment, intercom systems and many others.
Alango Acoustic Echo Cancellation is seamlessly integrated with other Alango technologies into one Voice Communication Package (VCP). Besides echo cancellation, VCP includes stationary and transient noise suppression, dynamic range compressor, automatic volume and equalization control, speech enhancement, (optional) adaptive dual and multi-microphone arrays and several other functions.
- Fast convergence (< 300ms) with no initial echo
- Convergence in double talk and high noise
- Robustness to speaker signal distortions (including special non-linear processor distortions)
- Echo canceller filter length up to 700ms
- Residual echo level: < -60dB
- Processing delay: 30ms
- Supported sampling rates: 8kHz, 16kHz, 24kHz, 32kHz
- Implementations for variety of platforms including fixed- and floating-point processors