7 Data formats

The acoustic data is stored in binary formatted files external to the DBMS. The files are in the format produced by the CREST echosounder system - see [Coombs et al. 2003] for a discussion of this echosounder.  The acoustic data is stored in the files in a ’message’ format that consists of a header that defines the type of data that follows in the body part of the message. Multiple messages are stored in one file.

7.1 Header format

The message format files consist of 16 bit integers grouped into messages. Each message begins with a 12 byte (6 x 16 bit unsigned integers) header, defined as follows:

The next message header begins ’length’ bytes past the end of the current message header. This repeats until the end of the file. The format of the message body is determined by the value in the ’type’ field of the header.

7.2 Header type 32

The vast majority of the data collected is processed acoustic data and has always been stored in a ’bundled’ format. The header ’type’ value that indicates ’bundled’ format is 32. Acoustic data is collected via an analog-to-digital converter and the term used for the value obtained from the a/d converter is ’sample’. When acoustic data is collected a threshold is applied and samples below the threshold are set to zero. The bundled data format just records the non-zero samples. A series of samples that do not contain a zero valued sample is called an ’echo’. Hence, the bundled data format stores samples in groups of echoes thus: