Voice detection method, voice detection device, and computer device

A voice detection method, comprising steps:

• acquiring an audio sequence;
• extracting a first audio feature from the audio sequence, and performing voice detection on the audio sequence according to the first audio feature to obtain a first voice detection result;
• extracting a second audio feature from the audio sequence, and performing the voice detection on the audio sequence according to the second audio feature to obtain a second voice detection result;
• and
• determining a voice detection result of the audio sequence according to the first voice detection result and the second voice detection result;
• wherein
• the first audio feature comprises an average energy an energy ratio, and a zero-crossing rate of an audio signal;
• the step of extracting the first audio feature from the audio sequence and performing the voice detection on the audio sequence according to the first audio feature to obtain the first voice detection result comprises steps:
• performing sampling frequency conversion and framing processing on the audio sequence to obtain frames of audio sub-signals;
• calculating an average energy and a zero-crossing rate of each of the frames of the audio sub-signals according to each of the frames of the audio sub-signals to obtain the average energy and the zero-crossing rate of the audio signal;
• obtaining energy spectra of the audio sub-signals, obtaining low-frequency band energy and high-band energy according to the energy spectra, and calculating a ratio between an average energy of the low-frequency band energy and an average energy of the high-band energy to obtain the energy ratio of the audio signal;
• and
• performing the voice detection on the audio sequence according to the average energy, the zero-crossing rate, and the energy ratio of the audio signal to obtain the first voice detection result;
• the step of performing the voice detection on the audio sequence according to the average energy, the zero-crossing rate, and the energy ratio of the audio signal to obtain the first voice detection result comprises steps:
• comparing the average energy of the audio signal with a first predetermined threshold;
• comparing the energy ratio of the audio signal with a second predetermined threshold;
• comparing the zero-crossing rate of the audio signal with a third predetermined threshold;
• and
• determining that the first voice detection result is a voice when the average energy of the audio signal is greater than the first predetermined threshold, the energy ratio of the audio signal is greater than the second predetermined threshold, and the zero-crossing rate of the audio signal is greater than the third predetermined threshold;
• the second feature comprises a spectral modulation energy;
• the step of extracting the second audio feature from the audio sequence and performing the voice detection on the audio sequence according to the second audio feature to obtain the second voice detection result comprises steps:
• performing the sampling frequency conversion and segmentation processing on the audio sequence to obtain audio segments;
• calculating a Mel spectrum for each of the audio segments to obtain a Mel spectrogram containing channels;
• performing the fast Fourier transform on each of the channels in the Mel spectrogram, and calculating a normalized modulation energy of each of the channels;
• and
• performing the voice detection on the audio sequence according to the normalized modulation energy of each of the channels to obtain the second voice detection result;
• the step of performing the voice detection on the audio sequence according to the normalized modulation energy of each of the channels to obtain the second voice detection result comprises steps:
• calculating a sum of the normalized modulation energy of each of the channels;
• comparing the sum of the normalized modulation energy of each of the channels with a fourth predetermined threshold;
• if the sum of the normalized modulation energy of each of the channels is greater than the fourth predetermined threshold, determining that the second voice detection result is the voice;
• and
• if the sum of the normalized modulation energy of each of the channels not greater than the fourth predetermined threshold, determining that the second voice detection result is non-voice.