SEED Manual

The following is taken directly from the official SEED manual. You can download the entire manual in Adobe Acrobat (.pdf) format.

Appendix G: Data Only SEED Volumes (Mini-SEED)

Contributed by Tim Ahern

The SEED format consists of Volume Control Headers, Abbreviation Control Headers, Station Control Headers, Time Span Control Headers and finally Data Records. At the 1991 FDSN meeting in Vienna, Austria the concept of Dataless SEED volumes was introduced and accepted. The structure of SEED data records is simple, straightforward, and much simpler to understand than the control header structure of SEED. Some data loggers offer SEED data records as a method of transferring waveform information. The term Data Only SEED Volumes (Mini-SEED) has come to be used to identify SEED data records without any of the associated control header information. Data Only and Dataless SEED volumes are to a certain extent the two parts of a complete SEED volume. Only Time Span Control Headers are not included in either of these components, however Time Span Control Headers can be derived from the Data Only SEED.

The SEED format standard is defined by the FDSN Working Group on Data Exchange. This working group has recognized the need to more specifically address the definition and use of Data Only SEED as a data exchange format. Data Only SEED also has potential for use as a data analysis format. In the SEED format, much of the information needed to specify the time series in the data records is in the SEED control headers. In fact the data record portion of the SEED format does not contain information about the organization of the data in the Data Only SEED records. Missing information includes

1. specification of the data encoding format as normally specified in the DDL
2. the byte swap order of the data as either VAX like or Motorola like.
3. the data record length

With the inclusion of the above information, the Data Only format can be used to completely decode the time series information in the data records. Of course response information and some other information remains unavailable and the need to retain full SEED volume production is encouraged.

The Data Only SEED data blockette has been designed to include the needed information. The data blockette is defined as follows;

[1000] Data Only SEED Blockette (8 bytes)

1. UWORD : Blockette type (1000): Data Only SEED
   
   
2. UWORD : Byte number of next blockette. (Calculate this as the byte offset from the beginning of the logical record - including the fixed section of the data header; use 0 if no more blockettes will follow.)
   
   
3. BYTE : A code indicating the encoding format. This number is assigned by the FDSN Data Exchange Working Group. To request that a new format be included contact the FDSN through the FDSN Archive at the IRIS Data Management Center. To be supported in Data Only SEED, the data format must be expressible in SEED DDL. A list of valid codes at the time of publication follows.

   CODES 0-9	GENERAL
   0	ASCII text, byte order as specified in field 4
   1	16 bit integers
   2	24 bit integers
   3	32 bit integers
   4	IEEE floating point
   5	IEEE Double precision floating point
   CODES 10 - 29	FDSN Networks
   10	STEIM (1) Compression
   11	STEIM (2) Compression
   12	GEOSCOPE Multiplexed Format 24 bit integer
   13	GEOSCOPE Multiplexed Format 16 bit gain ranged, 3 bit exponent
   14	GEOSCOPE Multiplexed Format 16 bit gain ranged, 4 bit exponent
   15	US National Network compression
   16	CDSN 16 bit gain ranged
   17	Graefenberg 16 bit gain ranged
   18	IPG - Strasbourg 16 bit gain ranged
   19	STEIM (3) Compression
   CODES 30 - 49	OLDER NETWORKS
   30	SRO Format
   31	HGLP Format
   32	DWWSSN Gain Ranged Format
   33	RSTN 16 bit gain ranged

   
   
4. The byte swapping order for 16 bit and 32 bit words. A 0 indicates VAX or 8086 order and a 1 indicates 68000 or SPARC word order. See fields 11 and 12 of blockette 50.
   
   
5. The exponent (as a power of two) of the record length for these data. The data record can be as small as 256 bytes and, in Data Only SEED format as large as 2 raised to the 256 power.

Additional Considerations in Data Only SEED

1. Any SEED data blockette can be included in the Data Only SEED format except those that refer to abbreviation dictionary blockettes. For instance blockette 100 is permitted but blockette 400 is not.
   
   
2. The Data Only SEED data blockette can be present in a full SEED volume. In this case the values in the Data Only SEED blockette take precedence over values in the SEED control headers.
   
   
3. When combining Data Only SEED data records with a Dataless SEED volume to produce a SEED volume, Time Span Control Headers must be constructed. The only other major consideration is that if the Data Only SEED record length exceeds the maximum length of 4096 bytes allowed in SEED, then the longer Data Only SEED data records must be blocked into data records of valid length.
   
   
4. Much of the necessary information needed to make the time series decipherable is already well defined in the fixed section of the data header. These fields remain unchanged.
   
   
5. Each data record must have blockette 1000.
   
   
6. The order of the fixed section of the data header must be the same as field 4 of blockette 1000 implies.

The IRIS SEED reader, RDSEED, is now capable of processing Dataless SEED volumes and Data Only SEED volumes simultaneously and therefore may provide a method of reading Mini-SEED data records that do not include blockette 1000.