Disclosed are various embodiments directed to transcoding a bit stream comprising audio and video frames that contains discontinuities. The video and/or audio portion of the bit stream may be divided into segments, where some of the segments are transcoded in parallel using a plurality of transcoder devices. In some embodiments, a discontinuity is detected in the bit stream before encoding, where the discontinuity is associated with a shift in a set of time stamps associated with the bit stream. A magnitude of the discontinuity is determined. The transcoding system aligns at least a portion of the audio frames to the series of video frames according to at least the discontinuity and the magnitude of the discontinuity. In other embodiments, the discontinuity is detected after encoding by analyzing time stamps that have been established during the transcoding process.