Speech recognition effect test method, device, equipment and product

• Main Author: BAN ZHIHUA
• Format: Patent
• Language: Chinese; English
• Published: 30.04.2024
• Subjects: ACOUSTICS; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

The invention discloses a voice recognition effect test method, device, equipment and product, and relates to the technical field of voice recognition, and the method comprises the steps: determining first test audios and label texts corresponding to the first test audios, and combining at least two first test audios to obtain a combined audio; the combined audio is played to the target device, recording is carried out, and a playback recorded audio is obtained; segmenting the playback recording audio to obtain second test audios matched with the first test audios in number, and endowing the tag texts corresponding to the first test audios with the corresponding second test audios based on the combination sequence of the first test audios and the segmentation sequence of the second test audios; and determining a voice recognition effect of the to-be-tested terminal device based on the second test audio. According to the invention, the test passing rate of the speech recognition algorithm is improved, the numb