# US 12,225,317 B2 · Front-end clipping using visual cues > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,225,317 B2 - Document ID: US12225317B2 - Status: Granted - Kind: B2 - Application number: 17/653,287 - Filed: 2022-03-03 - Published: 2023-09-07 - Granted: 2025-02-11 - Earliest claimed: 2022-03-03 - Assignees: International Business Machines Corporation - Applicants: International Business Machines Corporation - Inventors: Joseph Sayer; Andrew David Lyell; Benjamin David Cox - Web page: https://patentgrep.com/patent/US12225317B2 ## CPC - G10L 25/78: Detection of presence or absence of voice signals (switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10) - H04N 5/067: Arrangements or circuits at the transmitter end - G10L 25/57: for processing of video signals - H04N 7/04: Systems for the transmission of one television signal, i.e. both picture and sound, by a single carrier {(H04N7/084, H04N7/087 take precedence)} - H04N 7/147: {Communication arrangements, e.g. identifying the communication as a video-communication, intermediate storage of the signals (selecting H04Q)} - H04N 7/152: {Multipoint control units therefor} - H04N 21/8547: involving timestamps for synchronizing content - G06V 40/20: Movements or behaviour, e.g. gesture recognition (recognition of facial expressions G06V40/16) - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING - H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract According to one embodiment, a method, computer system, and computer program product for front-end clipping reduction is provided. The embodiment may include capturing input, including at least one visual input and at least one audio input. The embodiment may also include modeling data regarding visual cues based on a visual input from the at least one visual input. The embodiment may further include marking one or more timestamps which, in light of the modeled data, correspond to speech in the at least one audio input. The embodiment may also include transmitting an audio input from within the at least one audio input corresponding to the one or more marked timestamps. ## Claims ### Claim 1 A processor-implemented method, the method comprising: capturing input, including at least one visual input and at least one audio input, to a first device; training a machine learning model to recognize a visual cue indicative of a user desire to speak and predict a volume at which a user will speak based on a visual input from the at least one visual input and an audio input, synchronized to the visual input, from the at least one audio input; marking one or more timestamps which are determined, using the model, to correspond to speech in the at least one audio input; and transmitting an audio input from within the at least one audio input corresponding to the one or more marked timestamps from the first device to a second device. ### Claim 2 (depends on 1) The method of claim 1, further comprising: transmitting a visual input. ### Claim 3 (depends on 2) The method of claim 2, wherein transmitting the visual input further comprises: synchronizing a visual input in time with a corresponding audio input. ### Claim 4 (depends on 1) The method of claim 1, wherein marking a timestamp from within the one or more timestamps further comprises: measuring a likelihood that a timestamp within the one or more timestamps corresponds to speech in an audio input from within the at least one audio input. ### Claim 5 (depends on 1) The method of claim 1, wherein the one or more marked timestamps correspond to a present moment at which the marking occurs. ### Claim 6 (depends on 1) The method of claim 1, wherein the one or more marked timestamps correspond to future moments in time from a point at which the marking occurs. ### Claim 7 (depends on 1) The method of claim 1, wherein one timestamp from within the marked timestamps is marked more than once. ### Claim 8 A computer system, the computer system comprising: capturing input, including at least one visual input and at least one audio input, to a first device; training a machine learning model to recognize a visual cue indicative of a user desire to speak and predict a volume at which a user will speak based on a visual input from the at least one visual input and an audio input, synchronized to the visual input, from the at least one audio input; marking one or more timestamps which are determined, using the model, to correspond to speech in the at least one audio input; and transmitting an audio input from within the at least one audio input corresponding to the one or more marked timestamps from the first device to a second device. ### Claim 9 (depends on 8) The computer system of claim 8, further comprising: transmitting a visual input. ### Claim 10 (depends on 9) The computer system of claim 9, wherein transmitting the visual input further comprises: synchronizing a visual input in time with a corresponding audio input. ### Claim 11 (depends on 8) The computer system of claim 8, wherein marking a timestamp from within the one or more timestamps further comprises: measuring a likelihood that a timestamp within the one or more timestamps corresponds to speech in an audio input from within the at least one audio input. ### Claim 12 (depends on 8) The computer system of claim 8, wherein the one or more marked timestamps correspond to a present moment at which the marking occurs. ### Claim 13 (depends on 8) The computer system of claim 8, wherein the one or more marked timestamps correspond to future moments in time from a point at which the marking occurs. ### Claim 14 (depends on 8) The computer system of claim 8, wherein one timestamp from within the marked timestamps is marked more than once. ### Claim 15 A computer program product, the computer program product comprising: capturing input, including at least one visual input and at least one audio input, to a first device; training a machine learning model to recognize a visual cue indicative of a user desire to speak and predict a volume at which a user will speak based on a visual input from the at least one visual input and an audio input, synchronized to the visual input, from the at least one audio input; marking one or more timestamps which are determined, using the model, to correspond to speech in the at least one audio input; and transmitting an audio input from within the at least one audio input corresponding to the one or more marked timestamps from the first device to a second device. ### Claim 16 (depends on 15) The computer program product of claim 15, further comprising: transmitting a visual input. ### Claim 17 (depends on 16) The computer program product of claim 16, wherein transmitting the visual input further comprises: synchronizing a visual input in time with a corresponding audio input. ### Claim 18 (depends on 15) The computer program product of claim 15, wherein marking a timestamp from within the one or more timestamps further comprises: measuring a likelihood that a timestamp within the one or more timestamps corresponds to speech in an audio input from within the at least one audio input. ### Claim 19 (depends on 15) The computer program product of claim 15, wherein the one or more marked timestamps correspond to a present moment at which the marking occurs. ### Claim 20 (depends on 15) The computer program product of claim 15, wherein the one or more marked timestamps correspond to future moments in time from a point at which the marking occurs.