Data formats

TOAST programs use SeismicUnix data format to store waveform data.

The TOAST project decided to use SeismicUn*x as the standard format for seismic waveform data. However, the small sampling interval used in ultrasonic data cannot be stored with the standard SeismicUn*x header definition. SeismicUn*x sampling intervals are given as integer values counting in microseconds, whereas, the standard sampling rate for ultrasonic data is 10MHz.


Within TOAST the header field d1 (byte# 181-184) is used to indicate the sampling interval. Since sampling interval values in the order used for microseismic or ultrasonic data cannot be represented by the short integer values used in Seismic Unix, we add an exception to the SeismicUnix header. Small sampling intervals (smaller than 1 microsecond) as used in ultrasonic records are to be stored in SeismicUn*x data format within TOAST as follows:

  1. Time values for ultrasonic data will be given in nanoseconds for dt and microseconds for delrt.
  2. This applies to fields
    • dt (byte# 117-118): sample interval
    • delrt (byte# 109-110): delay recording time
  3. Time units other than microseconds or nanoseconds for dt and milliseconds or microseconds for delrt are not allowed.
  4. Field d1 (byte# 181-184) in TOAST data will be
    • either zero, indicating standard seismic data with a micro seconds time scale
    • or provide the sampling interval in seconds, thus indicating
      • either standard seismic data, if d1 in seconds matches dt if the latter is taken in microseconds
      • or ultrasonic data, if d1 in seconds matches dt if the latter is taken in nanoseconds and delrt is taken in microseconds
In this way it will be possible to use SU tools to sort traces or similar or even waveform filters. The user just has to take care to pass filter frequencies in kHz rather than Hz in the case of ultrasonic data.