Importance of better Speech Quality

With the way we communicate changing, improving and going digital day by day, improving speech quality is of utmost importance to be at a position to lead the markets in the space. Let's think about the standard environment in which speech communication occurs. There are multiple issues in the environment that hinder the quality of audio and speech intelligibility. We discuss the top issues here about how it affects the voice experience, and how the Speech Processing Solutions from Analog Devices can help mitigate these issues to provide the best hearing experience. Echo Remember hearing yourself at times while you were talking on the call to your friend? These are acoustic echoes caused because your friend had put you on the speaker. The block diagram below explains the acoustic echo scenario. This is a scenario of me and my friend on a call over the IP. Let's get familiarized with few of the jargons here. Audio from my friend is called the Far End Signal, and my audio is called the Near End Signal. Correspondingly, the place where the Far End Signal is played out (my loud speaker) is the Far End Out and the place where the Far End Signal is received (My friend's microphone) is the Far End In. Similarly, my microphone is the Near End In and my friend's loud speaker is the Near End Out. When I have the Far End Signal played out on my speaker, that voice can be fed back into my microphone, which creates an acoustic echo. Due to this, my friend will be hearing himself, with a delay and creates a very uncomfortable hearing experience. The Acoustic Echo Cancellation Solution from Analog Devices solves this issue by cancelling this echo from the Near End Signal. Background Noise No matter how good the environment is, there always exists some kind of a persistent background noise. The examples of such noises maybe fan noise, AC noise, machine noise, wind noise, engine noise, noise introduced due to the data and electrical interfaces in the system, etc. These noises affect the intelligibility of speech audio, whether it's for human perception, or for processing on a Voice UI. The Noise Reduction Solution from Analog Devices estimates these persistent background noises and suppresses them so that the most intelligible pure voice audio can go through. Unwanted sound sources around Though the background persistent noises are taken care of, one might be in an environment where there are other sources of audio/sound too. It might be other people around talking to themselves, music, kids running around, or the sounds of traffic from a window. In such scenarios, it is important that the system picks up speech only from the desired source and suppresses audio from the other sources. The microphone beamforming technology from Analog Devices helps in concentrating on voice input only from the direction of the desired speaker and suppresses the audio from other unwanted directions. Beamforming is an essential part of any conferencing system these days and the solution from Analog Devices makes this possible with the best performance and minimal resources. Delay Audio and Video calls over IP are the new normal. One major problem with this is the significant delay that exists, mainly due to the transmission over IP. Often, one would experience that a person has to speak, and pause for a brief amount of time to hear the reply from the other side. Any delay that gets added due to audio processing, though negligible compared to the delay over IP, can still cause an added negative experience on the already hampered scenario of delay. The Speech Processing Solutions from Analog Devices introduce minimal delay with fast convergence rates so that it can improve audio even in most time critical scenarios.

Acoustic Echo Cancellation

What is Acoustic Echo Cancellation? The Acoustic Echo Cancellation (AEC) block is designed to remove echoes, reverberation, and unwanted added sounds from a signal that passes through an acoustic space. As shown in the diagram below, the sound coming from the remote person speaking, known as the Far End In, is sent in parallel to a DSP path and to an acoustic path. The acoustic path consists of an amplifier/loudspeaker, an acoustic environment, and a microphone returning the signal to the DSP. The AEC block is based on an adaptive FIR filter. The algorithm continuously adapts this filter to model the acoustic path. The output of the filter is then subtracted from the acoustic path signal to produce a “clean” signal output with the linear portion of acoustic echoes largely removed. The AEC block also calculates a residual signal containing nonlinear acoustic artifacts. This signal is sent to a Residual Echo Cancellation block (RES) that further recovers the input signal. The signal is then (optionally) passed through a noise reduction function to produce the output, which is a known as the “Far End Out.” The filter pauses adaptation when it detects sounds in the acoustic path unrelated to the far end in. This allows sounds in to be added to the far end out. For example, in the case of a hands-free system or speakerphone, adaptation pauses when a person speaks directly into the microphone. The person at the far end hears only the local talker and not the echoes and reverberation from the far end in the near end space. This is absolutely necessary for clear, full duplex conversation over communication channel. The adaptive filter in the AEC block will remove any signal that is part of the far end in. This can be extremely useful. There are two blocks that, together, comprise the Acoustic Echo Cancellation algorithm as implemented on the SigmaDSP. The primary block (AEC) removes the linear portion of the echoes and reverberation. operates in the stream processing domain and performs most of the functionality. A second block (RES) runs in the frequency domain and supplements the primary AEC block by performing nonlinear processing on the residual echo signal. The residual echo removal improves the performance substantially, and all performance benchmarks have been made using the two blocks together. However, the RES block is, strictly speaking, optional. An additional block, noise reduction, is also shown in the diagram below. While not formally part of the AEC algorithm, they are used together more often than not. The noise reduction block also operates in the frequency domain and can follow the residual echo suppression in the block/frequency domain space with no additional overhead. The noise reduction block has use in many applications, however, that do not requires AEC. Most AEC implementations, including this one, are optimized for speech. As such, the bandwidth is reduced to minimize processing resources. There are three widely accepted bandwidth standards: Narrow band (NB) uses a sampling rate of 8 kHz (bandwidth of 4 kHz). This is the bandwidth historically used in telephony (PSTN). Wide band (WB) uses a sampling rate of 16 kHz (8 kHz bandwidth). This is the most commonly used implementation for new products, and is the implementation documented below. The base implementation of SigmaDSP is narrow band. However, a wide band implementation can be constructed using two narrow band blocks, one for the lower frequencies and one for the upper frequencies. The documentation and example projects are all wide band and are implemented in this way. Ultra-wide band (UWB) uses a sampling rate of 32 kHz (16 kHz bandwidth). UWB AEC is not often used in speech applications. Why do I need Acoustic Echo Cancellation? AEC is needed when a far end signal is played over a loudspeaker into a reverberant acoustic space and is picked up by a microphone. In the full duplex hand-free telecommunication systems remote user's voice will be delayed and returned to the remote user. AEC will remove those echoes, reverberation, and unwanted added sounds. What are the design considerations? Tunable Parameters Algorithm Parameter Range Optimal AEC Taps 320 to 640 640 HPF Cut Off 20 to 300 Hz 150 Hz RES Temporal Smoothing 0.1 to 1.0 0.2739 Spectral Smoothing 0.1 to 1.0 0.2 Smoothing factor 0.1 to 1.0 0.7 Note that the standard compliance testing was performed with the above mentioned optimal parameters and there is no guarantee of performance if this parameter is changed. Resource Requirement Algorithm MIPS (16kHz) Data Memory (16kHz) AEC 24 6k RES 8 6k More information about AEC available on ADI wiki Note that the above information is for AEC on the ADAU1467 and ADAU1452 platforms only. For details about the Microsoft Teams Certified ADSP-21569 based solution, please refer here . AEC User Guide RES User Guide Ready to get started? Find all the information needed to start your evaluation journey here... Please contact SigmaStudioLicensing@analog.com for licensing information.

What is Noise Reduction? The Noise Reduction algorithm is used to eliminate persistent background noises. The examples of such noises maybe fan noise, AC noise, machine noise, wind noise, engine noise, noise introduced due to the data and electrical interfaces in the system, etc. This Noise Reduction block is optimized for speech over a communication channel. The most common application is hands-free telephony, but the adaptive filtering is applicable in a wide range of systems. It is often used with Acoustic Echo Cancellation (AEC) to remove noise as well as echoes and reverberation. For applications that require low latency, such as communications in which both a direct acoustic path and an amplified path are present, the Low Latency Noise Reduction block is more suitable. Why do I need Noise Reduction? Its often the case that the environment around us has some kind of persistent background noises, no matter where we are, due to the multiple sources listed above. Such background noise, if not eliminated can hamper the audio experience very badly. An efficient speech communication system should be able to eliminate these persistent background noises to enable clear and crisp speech audio. The Noise Reduction block will remove these unwanted background noises. What are the design considerations? Tunable Parameters Algorithm Parameter Range Optimal NR Smoothing factor 0.01 to 0.5 0.2 Note that the standard compliance testing was performed with the above mentioned optimal parameters and there is no guarantee of performance if this parameter is changed. Resource Requirement Algorithm MIPS (16kHz) Data Memory (16kHz) NR 4 10k More information about Noise Reduction available on ADI wiki Note that the above information is for NR on the ADAU1467 and ADAU1452 platforms only. For details about the Microsoft Teams Certified ADSP-21569 based solution, please refer here . Noise Reduction User Guide Ready to get started? Find all the information needed to start your evaluation journey here... Please contact SigmaStudioLicensing@analog.com for licensing information.

Speech Processing Solutions

Application Areas

Basic Communications

Walkie talkies, nurse call systems, intercoms, etc., though considered as basic communicating systems, clarity of speech is paramount. Our offerings help in suppressing unwanted and background noises, echoes and allows the user to hear crystal clear voice.

Gaming

You wouldn’t want to restrict a gamer from using his gaming speakers while playing live multiplayer games due to the echo caused by the game sounds while interacting with the players. Fast convergence for audio solutions becomes a critical factor in gaming applications to enable a good gaming experience. Hassle free gaming experience made possible with our speech processing solutions to eliminate echo and enhance audio quality in multiplayer gaming interactions.

Conferencing Solutions

Clear speech is at the heart of any conferencing system. Tabletop speaker phones, ceiling microphones, unified communication sound bars and a plethora of other options for conferencing can make use of the speech processing solutions from ADI to eliminate echo, noise, and unwanted sounds in the environment to enable a clear speech interaction. ADI offers both basic and Microsoft Teams Certified conferencing solutions. Please visit the Getting Started page for more information on the solutions.

Voice UI

It’s the age of voice recognition, and anything that can make audio more intelligible for the voice recognition algorithms is utmost critical. Faster and more efficient voice UI, made possible with the speech processing solutions from ADI.

Be part of a superior hearing experience with our offerings,

Ready to get started? Find all the information needed to start your evaluation journey here...

Please contact SigmaStudioLicensing@analog.com for licensing information.

Tags: beamformer speech processing teams certified zoom speech teams noise echo skype certified cancellation echo cancellation noise reduction aec zoom certified audio skype Show More

Recommended

Partner Forums

Partner Libraries

Speech Processing Solutions

Ready to get started? Find all the information needed to start your evaluation journey here...