Preventing audio delay-induced miscommunication in audio/video conferences

    公开(公告)号:US11581007B2

    公开(公告)日:2023-02-14

    申请号:US17302191

    申请日:2021-04-27

    Applicant: Kyndryl, Inc.

    Abstract: Embodiments for delay-induced miscommunication reduction are provided. The embodiment may include capturing data streams transmitted between participants in an A/V exchange; translating, on a sender device prior to transmission to a recipient device, an audio stream within the data streams to text; timestamping, on a sender device prior to transmission to the recipient device, each word in the translated audio stream; transmitting the audio stream and the sender-side translated and timestamped audio stream to the recipient device; translating, on the recipient device, the transmitted audio stream to text; timestamping, on the recipient device, each word in the translated audio stream; determining a lag exists in the A/V exchange based on a comparison of each timestamp for corresponding words on the sender-side translated and timestamped audio stream and the recipient-side translated and timestamped audio stream; and generating a true transcript of an intended exchange between the participants based on the comparison.

    PREVENTING AUDIO DELAY-INDUCED MISCOMMUNICATION IN AUDIO/VIDEO CONFERENCES

    公开(公告)号:US20220343938A1

    公开(公告)日:2022-10-27

    申请号:US17302191

    申请日:2021-04-27

    Applicant: Kyndryl, Inc.

    Abstract: Embodiments for delay-induced miscommunication reduction are provided. The embodiment may include capturing data streams transmitted between participants in an A/V exchange; translating, on a sender device prior to transmission to a recipient device, an audio stream within the data streams to text; timestamping, on a sender device prior to transmission to the recipient device, each word in the translated audio stream; transmitting the audio stream and the sender-side translated and timestamped audio stream to the recipient device; translating, on the recipient device, the transmitted audio stream to text; timestamping, on the recipient device, each word in the translated audio stream; determining a lag exists in the A/V exchange based on a comparison of each timestamp for corresponding words on the sender-side translated and timestamped audio stream and the recipient-side translated and timestamped audio stream; and generating a true transcript of an intended exchange between the participants based on the comparison.

Patent Agency Ranking