When it comes to delivering clear and intelligible audio messages in public spaces and transportation hubs, one of the challenges is to adjust the volume and quality of the sound according to the ambient noise level.

Ambient noise is the background sound that is present in the environment, such as traffic, wind, conversations, etc.  If the ambient noise is too high, the audio message might be drowned out or distorted. If the ambient noise is too low, the audio message might be too loud or intrusive.

One way to deal with this challenge is to use gap sensing, which is a method of measuring the ambient noise level during the gaps or pauses between words or sentences in the audio message. Based on this measurement, the volume of the audio message can be adjusted accordingly.

Gap sensing has some limitations:

• It can only measure the ambient noise level at discrete intervals, not continuously

• It can be affected by the length and frequency of the gaps in the audio message

• It can miss sudden changes or spikes in the ambient noise level that occur during the audio message

• It can cause noticeable fluctuations or variations in the volume of the audio message

A better way to deal with this challenge is to use realtime ambient noise sensing, which is a method of measuring the ambient noise level continuously and simultaneously with the audio message. Based on this measurement, the volume and quality of the audio message can be adjusted dynamically and smoothly.

Realtime ambient noise sensing has advantages:

• It can measure the ambient noise level at any time, not just during the gaps

• It can be independent of the content and structure of the audio message

• It can capture and respond to any changes or variations in the ambient noise level that occur during or between the audio message

• It can provide a consistent and optimal sound experience for the listeners

One of the technologies that enables realtime ambient noise sensing is spectral subtraction, which is a technique of removing or reducing the ambient noise from a signal by subtracting its spectrum from the spectrum of the signal.

Spectral subtraction can improve the quality and intelligibility of the audio message by enhancing its signal-to-noise ratio. The latest developments in spectral subtraction allow for a very accurate and effective ambient noise sensing system that performs better than the older echo cancelling ones.

Echo cancelling is a technique of removing or reducing the echo or feedback that occurs when a signal is reflected back by a surface or object, or picked up from itself via the ANS microphone.

Echo cancelling can improve the quality and clarity of the audio message by mostly cancelling the original signal from the microphone and only leaving the background noise.

Realtime ambient noise sensing is important for delivering clear and intelligible audio messages in public spaces and transportation hubs, as it can adjust the volume and quality of the sound according to the ambient noise level.

Realtime ambient noise sensing can be achieved by using spectral subtraction, which can leave only  ambient noise from a signal by subtracting the original source from the spectrum of the signal.

Spectral subtraction can provide a better performance than gap sensing or echo cancelling, which are older and less effective techniques.

The Smart pi DSP uses a custom spectral subtraction algorithm to ensure the best possible Ambient Noise Sensing and Control system is available even in harsh environments including underground train and bus stations.