What is the VOX file format?
.vox files contain raw OKI/Dialogic ADPCM audio used in interactive voice response (IVR) systems, auto-attendants, and on-hold messaging since the late 1980s. The format was developed around Dialogic voice-processing boards - PC expansion cards for telephony server environments.
The defining characteristic of .vox is the complete absence of a file header. There are no magic bytes, no metadata, and no stored sample rate or channel count. The entire file from byte 0 onwards is a raw stream of 4-bit ADPCM samples packed two per byte (one per nibble). Each nibble encodes a delta from the previous audio sample using the OKI ADPCM algorithm, achieving a 2:1 compression ratio over 8-bit PCM. To decode correctly, the player must be told the sample rate externally - typically 8000 Hz mono for telephony, though 6000 Hz and 11025 Hz also occur. Playing a .vox file without these parameters produces garbled audio.
At 8000 Hz, one second of audio occupies roughly 4 KB. The format is distinct from the Microsoft IMA ADPCM codec used inside .wav files, though both use ADPCM techniques. Modern editors such as Audacity, Adobe Audition, and the sox command-line tool can open .vox files when given the correct encoding parameters.
Security & safety
RISK: LOWRaw binary audio data; cannot execute code. No known malware vector. Standard caution for files from telephony systems (may contain recorded conversations - privacy consideration).
Format details
in a nutshell- MagicVox / generic .vox containers - Some audio tools use .vox for other raw audio or ADPCM variants unrelated to Dialogic; context-dependent.
- Quake engine voice file - .vox used by id Software Quake engine for ambient voice/sound samples; different binary format, not Dialogic ADPCM.
Programs that open VOX files
Technical details
deep spec| Format type | Raw headerless telephony audio (OKI ADPCM) |
| Encoding | OKI/Dialogic ADPCM - 4-bit adaptive delta pulse code modulation |
| Container | None - pure raw audio stream from byte 0 (no header or footer) |
| Magic bytes | None (headerless; no format identifier in file) |
| Sample packing | Two 4-bit ADPCM samples per byte (nibble-packed) |
| Compression | 4-bit ADPCM; 2:1 ratio versus 8-bit PCM |
| Byte order | N/A (raw 4-bit nibble pairs; no multi-byte integer fields) |
| Typical sample rate | 8000 Hz mono (telephony standard); also 6000 Hz and 11025 Hz used |
| Metadata | None stored in file; sample rate and channel count must be specified externally |
| Typical file size | 8 KB - 10 MB (1 second at 8000 Hz ≈ 4 KB) |
| Primary use case | IVR systems, auto-attendants, on-hold messaging, telephony server environments |
| Platform | Windows; Linux (telephony server environments) |
| ADPCM variant | OKI ADPCM (distinct from Microsoft IMA ADPCM used in WAV files) |
| Released | 1980s (Dialogic voice-processing hardware cards era) |
| Latest version | No versioning - static raw format; unchanged since introduction |
| Specification | www.dialogic.com |
VOX conversions
Community Q&A
asked by usersNo questions yet - be the first to ask about VOX files.