I am importing a
.mov where the media information is recorded at 30 fps with h.264 and the audio is AAC sampled at 32khz with 32bps.
Rendering is done at 30 fps and AAC at 192khz bitrate. When rendering the whole sequence including previous clips, audio and video inside the clip at some time $t$ are way out of sync (almost a second), but if I choose to render just 5 seconds before and after $t$ using , the audio and video are rendereded correctly in sync (or at least is below the threshold of perception). This is maddening since it fails the test of Least Astonishment, as rendering a segment of the video will provide no testing mechanism of how the full video will be rendered. Another way to put it is that the rendering process fails to satisfy an equivalence between concatenating rendering parts in separate. Because of the above I'm pretty biased against this being "expected functionality" and it is probably a bug. I am using v2.79 on Mac OS X 10.11.6