Convert Audio to 8 kHz Mono WAV for 3CX, Asterisk & IVR

Phone systems are the one place where a perfectly good audio file gets rejected for being too good. A 44.1 kHz stereo MP3 that plays fine everywhere else will be refused by 3CX, sound like chipmunks on Asterisk, or silently fail as an IVR prompt. The reason is simple: telephone audio has been 8,000 samples per second, mono, since the 1960s, and most PBX software still expects exactly that.

What phone systems actually want

Almost every business phone system, cloud or on-premise, expects the same thing for hold music, greetings and IVR menus:

• Format: WAV, uncompressed PCM (not MP3, not compressed WAV)
• Sample rate: 8 kHz (8,000 Hz). Some newer systems also accept 16 kHz for HD voice.
• Bit depth: 16-bit
• Channels: mono

8 kHz sounds low because it is: the top of the audible range is cut at 4 kHz, which is why phone calls sound the way they do. That is the G.711 codec standard the whole telephone network runs on. Uploading a 44.1 kHz file does not give callers better sound — the system either refuses it or downsamples it badly on the fly, which is where distortion comes from.

Requirements by system

These are the most common documented requirements. Vendors change things, so if an upload fails, check the current spec — but the 8 kHz / 16-bit / mono WAV below works on all of them.

How to convert a file in two steps

1. Make it mono. Open the Mono/Stereo Converter, upload your file, choose Stereo to mono, download.
2. Resample to 8 kHz, 16-bit WAV. Open the Sample Rate Converter, upload the mono file, pick 8 kHz (or 16 kHz if your system supports wideband), set bit depth to 16-bit, output WAV, download.

Both tools run in the browser, no account, no install. The whole thing takes under a minute. If you have a batch of prompts, do the mono step on all of them first, then the resample step — it's faster than alternating.

Getting the level right

Phone audio is quiet by design, and a hold-music file mastered at modern streaming loudness will clip and buzz through the codec. Before converting, aim for peaks around −6 to −3 dBFS and no heavy limiting — the Loudness Normalizer set to around −16 LUFS is a safe target. Voice prompts should be a touch louder than music beds so they cut through.

Common errors and what they mean

• “Invalid file format” / “File not supported” — usually a compressed WAV (ADPCM, MP3-in-WAV) or wrong sample rate. Re-export as plain PCM at 8 kHz.
• Sped-up, chipmunk audio — a 44.1 or 48 kHz file being played as 8 kHz. Resample it.
• Slow, deep audio — the reverse: an 8 kHz file tagged as 44.1 kHz. Re-convert from the original.
• Only one side plays / sounds thin — stereo file on a system that reads the left channel only. Convert to mono first.
• Crackle or buzz on loud parts— file is too hot for the codec. Lower the level, don't limit harder.

8 kHz or 16 kHz?

8 kHz works everywhere and is what callers on the public phone network hear regardless. 16 kHz (wideband, G.722, “HD voice”) only helps for calls that stay inside your own system — desk phone to desk phone, or a softphone app — and only if the system is configured for it. When in doubt, use 8 kHz.