You've got a nice-sounding greeting, you upload it to the phone system, and… silence. Or an error. Or a voice that sounds like a chipmunk. Nine times out of ten the recording is fine and the file format is wrong. Here's what each popular phone system expects, and how to convert a file yourself.
Phone calls carry narrowband audio, so most systems want WAV, 8 kHz, mono. The difference is the encoding: µ-law (Amazon Connect, Cisco UCCX and many PBXs) or 16-bit PCM (3CX). Microsoft Teams is the exception and also takes MP3. A 44.1 kHz stereo MP3 from a music app is the most common reason a prompt won't play.
The public phone network was built around G.711: 8,000 samples per second, one channel. Higher-quality audio doesn't sound better to the caller, because it gets squeezed down to 8 kHz anyway, and some systems simply reject or mis-play files that aren't already in their native format. Matching the format up front means no conversion surprises and the cleanest possible sound.
Three settings matter: sample rate (8 kHz for most systems), channels (mono), and encoding (µ-law, A-law or 16-bit PCM). µ-law is the North American G.711 standard; A-law is used in most other countries.
| Phone system | Format | Notes |
|---|---|---|
| Amazon Connect | WAV, 8 kHz, mono, µ-law | Under 50 MB and under 5 minutes, per AWS documentation |
| 3CX | WAV, 8 kHz, mono, 16-bit PCM | 3CX's documented prompt format |
| Microsoft Teams (auto attendant / call queue) | WAV (uncompressed PCM, 8/16/32-bit), WMA (mono) or MP3 | Maximum 5 MB |
| Cisco UCCX | WAV, 8 kHz, 8-bit, mono, µ-law (CCITT) | The format Cisco engineers recommend for prompts |
| FreePBX / Asterisk | WAV, 8 kHz, mono (16-bit PCM or µ-law) | System Recordings can convert uploads to the formats you select |
| Avaya, Mitel, Genesys, NICE CXone | Usually WAV, 8 kHz, mono, µ-law | Varies by model and version; check your admin guide |
| RingCentral, Zoom Phone | MP3 or WAV | Hosted platforms convert for you; watch the upload size limit |
Not sure which you have? Look for an “upload greeting” or “prompts” page in your admin portal; it usually lists the accepted formats right next to the button.
Upload rejected: wrong container or encoding, often an MP3 or a stereo WAV on a system that wants µ-law. Silence when it plays: the system accepted the file but can't decode the encoding. Fast, slow or high-pitched voice: the sample rate doesn't match what the system assumes. Hiss or harsh “s” sounds: a loud recording or heavy background music squeezed down to 8 kHz. Keep prompts clean, with no music bed unless it's mixed for phone audio.
Open the file, choose Tracks → Mix → Mix Stereo Down to Mono, set Project Rate to 8000 Hz (bottom left), then File → Export → Export as WAV and choose the encoding: U-Law for Amazon Connect and Cisco, or Signed 16-bit PCM for 3CX.
µ-law, 8 kHz, mono:sox greeting.wav -r 8000 -c 1 -e u-law greeting-ulaw.wav
16-bit PCM, 8 kHz, mono:sox greeting.wav -r 8000 -c 1 -b 16 -e signed-integer greeting-pcm.wav
Name files so anyone can tell what they are: 01_main_greeting.wav, 02_after_hours.wav, 03_holiday.wav. It saves a lot of guesswork when someone else uploads them later.
Tell us your phone system and we'll deliver every prompt already in the right format — greeting, menu, after-hours, holiday and hold messages, 10 prompts for $39, in natural English or Spanish voices.
See IVR Prompts →For most phone systems, a WAV file at 8 kHz, mono. Use µ-law encoding for Amazon Connect, Cisco UCCX and many PBXs, and 16-bit PCM for 3CX. Microsoft Teams also accepts MP3.
The file's sample rate doesn't match what the phone system expects. Convert it to 8 kHz mono before uploading.
Some hosted systems such as Microsoft Teams accept MP3, but many PBXs and contact center platforms need WAV. Converting to the native format avoids errors and quality loss.
Both are G.711 phone audio encodings. µ-law is used in North America and Japan; A-law is used in most other countries. Use the one your phone system and carrier expect.