Audio Support #124
Description
We may rethink of sampling and make some re-design based on the current Eval<T> system.
For video and audio, we feel like they are continious, but they are actually both discrete.
For video, we have frame rate, which is the static frames we have per second.
For audio, we have sample rate, which is the samples we have per second.
And the latter is much larger than the former. That means, we should do some design to batch them or even make some simd accelerate on it to make it fast.