Home News Contact
IVRSoft Logo

CT Developer Studio
Features
Screen Shots
System Requirements
Versions & Pricing
Online Manual
Q & A
CT Messenger
FastIVR for Cisco
Call Alert! free
Order
The Company
Support
Downloads

Resources

Fax Modems
Voice Modems
Telephony Boards
Credit Card Processing
Recording Voice Prompts





   Recording Voice Prompts

Tips on recording quality IVR voice prompts

Voice Prompt Recording

To record high quality voice prompts for your voice mail or interactive voice response (IVR) system, you should not use the telephone handset microphone. The handset microphone is by far less sensitive than even a cheap studio microphone, and has dynamic range that is not enough for a quality narration.

Selecting a microphone

Every type of microphone is different. You will need a studio microphone designed specifically to record voice. These microphones have a specific dynamic range to only record voice frequencies, and cut off all other frequencies (e.g. low-frequency noises) automatically.

Do not use the microphone supplied with your sound card, they are usually on the lower end of quality. There are a number of microphones available that were designed specifically for sound card use.

On the higher end there are high-quality studio microphones. Most high quality microphones, however, cannot be connected directly to the sound card. They must be connected first to a separate microphone 'preamplifier' ("preamp") that is connected to the 'Line Input' of your sound card. With this setup the recording sound quality will be greatly improved.

The microphone you use should be "unidirectional" (picks up sound in a single direction). "Omnidirectional" microphones pick up sounds from all directions, resulting in all noise in the room being recorded along with the voice.

Most microphone preamplifiers built into sound cards are of poor quality. As a result, recording the sound through the Line input, instead of the Mic input, may also improve the results.

Placing the microphone

Make sure you don't move the microphone around as this will create microphone handling noise on your recording. Using a microphone stand is a good idea; you can simply place a microphone on a pile of books if you don't have a microphone stand.

Place the microphone in about 15 - 30 cm (0.5 - 1 ft) from your mouth when recording (of course, the mic must be sensitive enough to support this distance. If the mic is closer than 1/2 ft, it will most probably record bubbling and other unpleasant artifacts.

Sound card

Recording sound quality varies greatly on different sound cards, especially the noise level of the sound cards. Most modern sound cards report a respectable SNR (signal to noise ratio, the measure of noise added by the card). In reality, the noise level is often higher than the specifications because of the interference between the electrical components inside your computer. This causes the recording to have more background noise, especially when recording using the built-in microphone input.

To improve the SNR of your recordings, avoid using sound cards that are built into the motherboard (usually called 'integrated sound card'). Get a good card in a price range of $50-100 US.

Software for sound recording and editing

You may use any of the available sound editing applications. Avoid using MS Sound Recorded that comes with Windows; its implementation of downsampling provides unacceptable sound quality.

Try using Sound Forge or CoolEdit. These tools provide both quality and powerful mixing and audio effects.

Audio format

When you initially record the voice prompts, use the highest available recording format. 44100 Hz, 16-bit, mono is recommended. Only downsample the recording sound when you are surely done with the processing and mixing.

Sound post-processing

Make sure to save your original voice prompt before starting editing the file.

If you want to filter out noise, equalize, add some music background or audio effects to your voice prompts, do it now. The number of tools and effects available depends on the software you used to record the voice prompt.

Converting to telephony-supported format

As a final stage of recording your voice prompts, you will need to downsample and save the recording to a format supported by telephony hardware. The final audio format depends on the telephony hardware you use. Dialogic cards (www.dialogic.com) require 11 kHz sound; all other telephony devices, including voice modems and other telephony boards, such as Brooktrout require 8000 Hz, 16-bit, mono format.

Summary

  • Pick the best unidirectional microphone allowed by your budget
  • Position and properly fix the microphone
  • Set the level of recording
  • Select software you are going to use for recording
  • Select best available recording format (44100 PCM / 16 bit / mono)
  • Record your voice prompts
  • Save the original, uncompressed PCM WAV file
  • Only do sound post-processing, mixing, equalizing and noise filtering with a copy of the file
  • After you are done with the post-processing and audio mixing, re-sample the file to the quality appropriate for telephony (11025 Hz for Dialogic, 8000 Hz for all other cards; 16-bit PCM is recommended)
  • Save the resulting file in a supported format (11 kHz or 8 kHz, 16 bit, mono).

Order professionally recorded voice prompts for your IVR

If you prefer your IVR voice prompts recorded professionally in the studio, please contact us for more information.

More IVR Software: 0..9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

IVR Software Development - CTI Experts


Copyright © 2002-2010 IVR Software Development