This delay is intended.
The BeagleBoard audio blocks use ALSA framework for sound capture and playback. On both the capture and playback sides, there are ring buffers that hold audio samples. These ring buffers give a cushion while capturing/playing back audio against buffer overruns / underruns. In the BeagleBoard ALSA Audio Capture / Playback implementation, we try to set the ring buffer size to hold 0.5 seconds of data. The playback starts once this ring buffer is full so the minimum latency will be about 0.5 seconds of playback latency plus capture latency which varies with the frame size.
We arrived at the ring buffer size through a lot of experimentation. At lower frame sizes, buffer overruns / underruns become an issue and manifests itself as glitches in the audio. Reducing the latency to imperceptible levels requires better real-time response out of the kernel that Linux inherently is not capable of.
