100 Things #73: Active Speech Level Control in SoundCheck

SoundCheck’s incorporation of active speech level allows for stimulus to conform with telephony testing standards. This means testing to IEEE and TIA standards can be automated within a sequence, allowing an entire test to be run without operator intervention. Active speech level is incorporated directly into SoundCheck’s stimulus step, and becomes available when selecting dB RMS. Active speech level can also be used for setting voice activation levels or wake word testing for smart devices.

Active Speech Level Control in SoundCheck

Learn more about active speech level in SoundCheck

Our Telephones application page details all of the benefits of using SoundCheck to test your communications devices.

Video Script:

Did you know that SoundCheck offers Active Speech level control in both the stimulus editor and post-processing?

Many telephony standards such as IEEE and TIA specify speech stimulus levels in terms of active speech level, or ASL.  It can also be used for setting voice activation levels or wake word testing for smart devices.

When active speech level is used, signal levels are scaled using only the parts of the waveform where speech is actually present, rather than the overall RMS level. Silent gaps are excluded but utterances that are part of normal speech are included. This removes variability due to different speaking speeds and content when making measurements, and provides better consistency for standardized measurements.

Configuring ASL in SoundCheck is easy. Let’s take a look.

If you’re testing with speech, you’ll be loading a speech file as a WAV file stimulus. Simply select the file, or open it from the memory list. In the level drop-down, select dB RMS and this will enable the ASL checkbox. Make sure the ASL checkbox is selected, and set the level of your stimulus. 

ASL is also available as a post-processing step. Here, you have some additional parameters. You can adjust the Time Constant  – that’s the time constant of the exponential averaging used to measure the level of active speech . You can modify the Hangover Time – that’s the allowable time for silence between utterances, 200ms is the default. And you can also set the Margin  – that’s the  difference between threshold of activity and the active speech level. This is useful because if the background noise is high, you can reduce the margin to exclude the noise. 

When you are using Active speech level in post-processing, The ASL value in the Memory List displays the Percentage of time during the waveform where speech is active, the mean power of speech measured over the aggregate time of activity and the mean power of the waveform measured over its entire duration.

Active Speech Level is an optional module, and can be added to any SoundCheck system.