Technical guide

When should IVR audio be converted before upload?

Compare strict canonical IVR uploads, vendor-accepted sources that platforms transcode, and Twilio's already-prepared raw stream media contract.

By IVRReady · Published August 28, 2026 · Last updated August 28, 2026

Direct answer

Convert audio before upload only when the source is not target-ready, IVRReady has proven capability for that source layout, and a mechanical transformation can satisfy the contract. Some platforms demand one strict upload layout. Others accept broader WAV or file inputs and transcode them. Twilio Media Streams is different again: it expects already-prepared raw media rather than an uploaded prompt file.

Family 1 — strict canonical uploads

Amazon Connect, Webex Contact Center, NICE CXone use executable IVRReady profiles that require their canonical μ-law WAV fields before package admission. A verified noncanonical source is Fixable; transformed bytes are reinspected before the WAV can enter prompts/.

Family 2 — broader accepted WAV matrix plus platform conversion

Genesys Cloud Architect documents multiple accepted WAV layouts and platform conversion. IVRReady copies an accepted source byte-for-byte, while separately recording whether platform conversion is expected and whether the source already matches the preferred canonical layout.

Family 3 — broad file/container acceptance

Talkdesk Greetings documents greeting file acceptance and platform transcoding at a broader boundary. IVRReady’s current input coverage remains WAV. Valid accepted WAVs are copied unchanged instead of being normalized merely because IVRReady could generate a telephony WAV.

Stream family — Twilio Media Streams

Twilio Media Streams is not an uploaded WAV prompt. The delivered artifact is headerless raw 8 kHz mono μ-law and a base64 message template. A canonical source skips audio transcoding but still needs structural data extraction; other verified sources first pass through the canonical WAV proof boundary.

Why IVRReady does not normalize every accepted file

If the vendor accepts the inspected source, changing its audio can add generation loss and make audit trails harder to explain. IVRReady therefore preserves accepted source bytes whenever the delivery contract permits it. Correction is reserved for sources that are not target-ready but can be transformed by a source layout proven with the real browser FFmpeg engine.

Why platform conversion can matter

Re-encoding, channel conversion, sample-rate conversion, and level handling can affect quality and predictability. IVRReady only attributes a specific conversion behavior to a vendor when its cited evidence says so. Otherwise, platform behavior is not inferred from a file being accepted.

Compare the executable target contracts

Use the contact-center audio requirements comparison for the current source-backed matrix, or read the G.711 μ-law WAV guide for the canonical byte-level foundation.

Sources and evidence

Vendor-specific statements follow IVRReady’s executable target profiles and their recorded official evidence. Technical context uses the primary sources below.