File formats & Data Types

Native file format & headers

SPRAAK has its own native file format which is fully described here:

• File formats and headers

FORMAT & TYPE

How the data is stored in the body of the file is described by a combination of the FORMAT and TYPE keys.

• FORMAT:
  • binary: BIN01 for little endian, BIN10 for big endian (default= current operating system)
  • ascii: ASCII
• TYPE:
  • signed integers: I8, I16, I32, I64
  • unsigned integers: U8, U16, U32, U64
  • floats: F32, F64
  • character & string: CHAR, STRING
  • others: ALAW, ULAW, BOOL

Character Sets

SPRAAK can handle a variety of character sets to specify phonetic alphabets, word transcriptions, .. The only constraint is that the character set should be BYTE based and a superset of ASCII, thus encompassing common schemes such as Latin-1, UTF-8, ..

General tools

• spr_addkey.c add/modify/delete a single key in the header
• spr_copy.c copy data files (frame based)
• spr_kdiff.c compare files in key format
• spr_kappend.c concatenate data files (not for string or object TYPES)