Embodiments of the invention describe a method of generating an enhanced audio stream. The enhanced audio stream is generated from live audio received and accompanies a live video broadcast. This live audio is buffered and combined with a visual description of the live video broadcast. Together these form an enhanced audio stream. The visual description is entered by a visual describer while they are viewing a an instantaneously updated visual buffer size display. The visual buffer size display is calibrated to a pre-set maximum buffer size, and shows the pre-set maximum buffer size and the current buffer size in a visual manner. Further, live audio from the buffer that is inserted into the enhanced audio stream is compressed in time.