Frontend audio capture for video conferencing applications

A non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations including:

• receiving an input signal;
• determining a signal level of the input signal;
• estimating a first noise level in the input signal based on a first global minimum frequency of one or more minimum frequencies determined for one or more respective portions of a first search window of the input signal, the first search window corresponding to a silent period of the input signal;
• generating a pre-suppression signal comprising applying a first gain factor or a first compression factor to the input signal based on the signal level of the input signal;
• generating a post-suppression signal comprising applying a second compression factor to the pre-suppression signal based on a signal from a voice activity detector and a selected noise level, the selected noise level being the first noise level;
• generating an output signal comprising applying a third gain factor or a third compression factor based on the post-suppression signal;
• estimating a second noise level based on a second global minimum frequency of one or more second minimum frequencies determined for one or more respective portions of a second search window of the input signal, the second search window corresponding to a voice-on period of the input signal;
• and
• responsive to the first global minimum frequency exceeding the second global minimum frequency, updating the selected noise level to use the second noise level.

A system comprising:

• one or more processors;
• and
• one or more computer-readable storage media storing instructions which, when executed by the one or more processors, cause the one or more processors to perform operations including:
• receiving an input signal;
• determining a signal level of the input signal;
• estimating a first noise level in the input signal based on a first global minimum frequency of one or more minimum frequencies determined for one or more respective portions of a first search window of the input signal, the first search window corresponding to a silent period of the input signal;
• generating a pre-suppression signal comprising applying a first gain factor or a first compression factor to the input signal based on the signal level of the input signal;
• generating a post-suppression signal comprising applying a second compression factor to the pre-suppression signal based on a signal from a voice activity detector and a selected noise level, the selected noise level being the first noise level;
• generating an output signal comprising applying a third gain factor or a third compression factor based on the post-suppression signal;
• estimating a second noise level based on a second global minimum frequency of one or more second minimum frequencies determined for one or more respective portions of a second search window of the input signal, the second search window corresponding to a voice-on period of the input signal;
• and
• responsive to the first global minimum frequency exceeding the second global minimum frequency, updating the selected noise level to use the second noise level.