(Page 1 of 1 in this chapter)

Appendix D

VOX File Format

This appendix presents the VOX file format. Note: Refer to the Microsoft Windows MultiMedia Programmer's Reference Manual for information about the WAVE file format.
A voice file is partitioned into segments. Each message consists of one or more segments in the file. A message can consist of one or more segments if it has been edited, or if, when the message was created, there were some existing re-usable segments.
Each segment of the file is described by an entry in the file header. These segment descriptors are divided into three groups, in order. Segment Descriptor Description Active Describes the messages currently in existence. The active descriptors are ordered by message number. When there are multiple segments for one message, the descriptors' order corresponds to the order of the segments in the message. Freed Describes sections of the file that are available for re-use. Unused These descriptors do not contain anything.
The used descriptors are ordered by their segments' position in the file.

Segment Descriptor	Description
Active	Describes the messages currently in existence. The active descriptors are ordered by message number. When there are multiple segments for one message, the descriptors' order corresponds to the order of the segments in the message.
Freed	Describes sections of the file that are available for re-use.
Unused	These descriptors do not contain anything.

Encoding	`VTYPE`	2 bytes	Header (32 bytes)
Total number of indices	`TOTLIDX`	2
Number of indices in use	`USEDIDX`	2
Number of indices in messages	`ACTVIDX`	2
Number of indices in free pool	`FREEIDX`	2
Highest message number in use	`HIGHMSG`	2
Total number of bytes	`TOTLBYTE`	4
Total number of frames	`TOTLFRM`	4
Reserved		12
Active Indices 10 x `ACTVIDX` bytes	Indices [10 x `TOTLIDX` bytes] (see Index Structure)
Freed Indices 10 x (`USEDIDX` - `ACTVIDX`) bytes
Unused Indices 10 x `FREEIDX` bytes
Actual voice frames in segments of various length where each segment is described by an index in either the Active Index list or the Freed Index list.	Frames [`TOTLFRM` x framesize] (see Summary of Frame Sizes)

This table presents the index structure: Message Number MSGNO 2 bytes Index Structure [10 bytes] Starting byte in file STRTBYTE 4 Number of frames in segment NFRAMES 4

This table presents frame sizes of common encodings: Encoding Description msec/frame bytes/frame 1 NMS ADPCM 16 kbs 20 42 2 NMS ADPCM 24 kbs 20 62 3 NMS ADPCM 32 kbs 20 82 4 framed PCM 64 kbs 20 162 10 mu-law 64 kbs 10 80 11 A-law 64 kbs 10 80 13 PCM 8 k mono 16 bit 10 160 14 OKI ADPCM 24 kbs 10 30 15 OKI ADPCM 32 kbs 10 40 16 PCM 11 khz 8-bit 10 110 17 PCM 11 khz 16-bit 10 220 20 G.726 ADPCM 32 kbs 10 40 22 IMA ADPCM 24 kbits/s 10 34 23 IMA ADPCM 32 kbits/s 10 46

Encoding	Description	msec/frame	bytes/frame
1	NMS ADPCM 16 kbs	20	42
2	NMS ADPCM 24 kbs	20	62
3	NMS ADPCM 32 kbs	20	82
4	framed PCM 64 kbs	20	162
10	mu-law 64 kbs	10	80
11	A-law 64 kbs	10	80
13	PCM 8 k mono 16 bit	10	160
14	OKI ADPCM 24 kbs	10	30
15	OKI ADPCM 32 kbs	10	40
16	PCM 11 khz 8-bit	10	110
17	PCM 11 khz 16-bit	10	220
20	G.726 ADPCM 32 kbs	10	40
22	IMA ADPCM 24 kbits/s	10	34
23	IMA ADPCM 32 kbits/s	10	46

(Page 1 of 1 in this chapter)