Controlling sounds of individual objects in a video
A method for modifying a sound produced by a sound source in a video includes capturing video and audio of a scene is disclosed. Audio is captured using a microphone array. A sound source is isolated and a direction of arrival of the sound source with respect to a capture location is identified. One or more visual objects in the captured video are identified. One of the isolated sound sources is associated with one of the identified visual objects. An input identifying one of the isolated sound sources is received during playing of the captured video and audio. The input includes a command. Responsive to receiving the input, an attribute of the identified isolated sound source is modified. The input may identify a visual object associated with a sound source. A system and article of manufacture are also disclosed.
1. A computer-implemented method for modifying a sound produced by a sound source in a video, comprising:
capturing video and audio of a scene, wherein the video and audio are simultaneously captured at a capture location;
for each of one or more sound sources in the captured audio, isolating the sound source and identifying a direction of arrival of the sound source with respect to the capture location;
identifying one or more visual objects in the captured video;
associating one of the isolated sound sources with one of the identified visual objects;
isolating continuously a sound produced by the sound source associated with the one or more visual objects as the visual object first moves towards then away from a relative position of a user;
playing the captured video and audio;
receiving an input identifying one of the isolated sound sources during the playing of the captured video and audio, wherein the input includes a command;
responsive to receiving the input, modifying an attribute of the playing of the identified isolated sound source, modifying the attribute including changing pitch of the continuously isolated sound source as the visual object first moves towards then away from the relative position of the user to simulate a Doppler effect.
2. The method of claim 1 , wherein the identifying one or more visual objects in the captured video includes identifying a visual object associated with the isolated sound source.
3. The method of claim 1 , wherein the isolating the sound source and identifying a direction of arrival of the sound source with respect to the capture location further comprises:
applying a beamforming algorithm to audio signals received by two or more microphones at the capture location thereby forming a beam pattern in which a sound source located in the beam pattern is isolated with respect to sound sources outside of the pattern.
4. The method of claim 3 , further comprising:
responsive to the sound source moving from a first position to a second position, changing parameters of the beamforming algorithm to steer the beam so that the sound source remains located in the beam pattern at the second position.
5. The method of claim 1 , wherein the playing the captured video and audio includes playing the captured audio and video on an augmented reality device.
6. A computer system comprising:
a processor; and
a memory communicatively coupled with the processor, wherein the memory is encoded within instructions for modifying a sound produced by a sound source in a video that when executed by the processor perform operations comprising:
capturing video and audio of a scene, wherein the video and audio are simultaneously captured at a capture location;
for each of one or more sound sources in the captured audio, isolating the sound source and identifying a direction of arrival of the sound source with respect to the capture location;
identifying one or more visual objects in the captured video;
associating one of the isolated sound sources with one of the identified visual objects;
isolating continuously a sound produced by the sound source associated with the one or more visual objects as the visual object first moves towards then away from a relative position of a user;
playing the captured video and audio;
receiving an input identifying one of the isolated sound sources during the playing of the captured video and audio, wherein the input includes a command;
responsive to receiving the input, modifying an attribute of the playing of the identified isolated sound source, modifying the attribute including changing pitch of the continuously isolated sound source as the visual object first moves towards then away from the relative position of the user to simulate a Doppler effect.
7. The computer system of claim 6 , wherein the identifying one or more visual objects in the captured video includes identifying a visual object associated with the isolated sound source.
8. The computer system of claim 6 , wherein the isolating the sound source and identifying a direction of arrival of the sound source with respect to the capture location further comprises:
applying a beamforming algorithm to audio signals received by two or more microphones at the capture location thereby forming a beam pattern in which a sound source located in the beam pattern is isolated with respect to sound sources outside of the pattern.
9. The computer system of claim 6 , further comprising:
responsive to the sound source moving from a first position to a second position, changing parameters of the beamforming algorithm to steer the beam so that the sound source remains located in the beam pattern at the second position.
10. A computer program product including a computer readable storage medium having computer readable program instructions stored thereon for causing a processor to perform operations for modifying a sound produced by a sound source in a video comprising:
capturing video and audio of a scene, wherein the video and audio are simultaneously captured at a capture location;
for each of one or more sound sources in the captured audio, isolating the sound source and identifying a direction of arrival of the sound source with respect to the capture location;
identifying one or more visual objects in the captured video;
associating one of the isolated sound sources with one of the identified visual objects;
playing the captured video and audio;
receiving an input identifying one of the isolated sound sources during the playing of the captured video and audio, wherein the input includes a command;
responsive to receiving the input, modifying an attribute of the playing of the identified isolated sound source, modifying the attribute including changing pitch of the continuously isolated sound source as the visual object first moves towards then away from the relative position of the user to simulate a Doppler effect.
11. The computer program product of claim 10 wherein the identifying one or more visual objects in the captured video includes identifying a visual object associated with the isolated sound source.
12. The computer program product of claim 10 , wherein the isolating the sound source and identifying a direction of arrival of the sound source with respect to the capture location further comprises:
applying a beamforming algorithm to audio signals received by two or more microphones at the capture location thereby forming a beam pattern in which a sound source located in the beam pattern is isolated with respect to sound sources outside of the pattern.