Audio resampling for media synchronization
AssigneeAmazon Technologies, Inc.
InventorsJordan Gilman, Trevor Wells, Max C. Denton
Life of this patent
Estimating the term…
Abstract
Approaches provided herein can provide for adjustments to audio data. In particular, an audio resampling process can be performed wherein the correct audio frame size is determined. A “target” audio stream duration, as may be indicated by a time-to-sample atom (STTS) atom, can be used to apply one or more pitch-invariant, time-stretching audio filters to the audio data as that data is decoded during a media transcode process. A resulting output media metadata can specify accurate, constant audio frame durations, and the audio stream duration can match the corresponding video stream duration in a media file.
Claims20
A computer-implemented method, comprising:
- receiving a media file to be transcoded, the media file including at least an audio portion and a video portion;
- determining, based at least in part upon metadata associated with the media file, that a total duration of the audio portion is different from an expected duration;
- determining, based at least in part upon the metadata, an average frame duration across a plurality of audio frames of the audio portion;
- determining, for the plurality of audio frames, that individual frame durations for the plurality of audio frames fall within a permissible threshold of the average frame duration;
- applying a resampling filter to the plurality of frames in order to apply a constant amount of duration adjustment across the plurality of audio frames, wherein the total duration of the resampled audio portion equals the expected duration;
- and
- performing transcoding of the media file including the video portion and the resampled audio portion.
A method, comprising:
- determining one or more variations in duration of a plurality of audio frames in an audio track;
- and
- performing, using an audio filter, resampling for one or more of the audio frames to cause the one or more audio frames to have a target audio frame duration, wherein the resampling is performed without deleting one or more frames of input audio, adding one or more frames of silent audio, or adjusting a pitch of the audio frames.
A system, comprising:
- a processor;
- and
- memory including instructions that, when executed by the processor, cause the system to:
- receive a media file to be transcoded;
- determine one or more variations in duration of a plurality of audio frames in an audio track of the media file;
- and
- perform, using an audio filter, resampling for one or more of the audio frames to cause the one or more audio frames to have a target audio frame duration, wherein the resampling is performed without deleting one or more frames of input audio, adding one or more frames of silent audio, or adjusting a pitch of the audio frames.
Dependencies are inferred from the claim text.