Cisco Cisco Unified MeetingPlace Audio Server Quick Setup Guide

Topics in this section include:

Recording environment—Make sure that the environment is extremely quiet. Do some sample 
recordings beforehand, and listen to the results. The recorded passages should be free of hissing, 
pops, frequency distortion, and other types of noise.

Silence before and after each prompt—Each prompt should have between 50 to 100 milliseconds 
of silence at the beginning and end. Too little silence creates pops when the prompt is concatenated 
with other prompts (played just after or before other prompts). Too much silence creates awkward 
gaps in overall speech when several prompts are played together. In some cases, 100 milliseconds 
might not be enough to eliminate pops at the beginning and end of concatenated prompts. In this 
case, try increasing the silence to 150 milliseconds. To measure the amount of silence at the 
beginning or end of a prompt (or to see if there is silence), bring up the prompt in Adobe Audition 
or a similar voice editing program. Open the file as a 16-bit, 8-kHz .wav or raw PCM file. Adobe 
Audition allows accurate measurement of silences passages.

Format—Record all prompts in 8 kHz sample rate, u-law, WAV format. Any recording that is not 
in this format must be processed in Adobe Audition or an equivalent audio editor to convert the 
recording to the required 8 kHz, mu-law, WAV format.

Level—The average recording level for non-silence sections should be -24 dBm. Avoid very loud 
recording levels because they can cause clipping, which creates clicks and pops in the middle of the 
prompts.

Because music prompts are subset of voice prompts, the 

 also applies to custom music prompts.

1495

3 minutes

Music heard by users in a waiting room.

1497

3 seconds

Welcome music heard by users who dial in.