What this miniSEED 3 inspector checks
The parser walks every variable-length record from its exact byte offset. It checks the MS indicator, format version, little-endian length fields, physical bounds, UTC field ranges, sample rate or period, reserved flags, source-identifier bytes, and selected fixed-width payload sizes. CRC-32C is recalculated across the complete record with the checksum field zeroed, as required by the specification.
Continuity and extra headers
Records are grouped by exact source identifier and publication version. Declared rates and sample counts produce an exclusive expected end time; sorted records then expose gaps and overlaps, while physical file order is audited separately. Extra headers are decoded as strict UTF-8 JSON and must have an object root. Their root namespaces and a bounded preview are shown, but extension-specific semantics are not inferred.
Why miniSEED 2 stops explicitly
Version 2 uses fixed record lengths, a different fixed-section header, blockette chains, multiple possible byte orders, and timing corrections. Treating those bytes as version 3 would create misleading results. Convert with current FDSN tooling or inspect with a libmseed-based application instead of relying on a partial guess.
Validation boundary
This is not the official miniSEED 3 validator or a waveform viewer. It does not decompress Steim data, decode or plot samples, validate every FDSN reserved extra-header field against JSON Schema, apply StationXML response metadata, resolve leap-second tables, reconcile archive filenames, repair records, or establish scientific fitness. Unknown future encodings remain structurally visible. Use FDSN’s current validator and libmseed tools for authoritative processing.
Frequently asked questions
Does a passing CRC prove the data are authentic?
No. CRC-32C detects accidental record corruption; it is not a digital signature and says nothing about provenance, calibration, or scientific correctness.
Why group publication versions separately?
Revised records can intentionally cover the same source and time range as earlier publications. Combining versions would turn expected revisions into false overlap findings.
Are tiny floating-point differences reported as gaps?
The comparison allows a very small tolerance: the larger of one microsecond and one millionth of the preceding sample interval. The exact declared timing still remains visible in the record table.
Authoritative references
Use the FDSN miniSEED 3 record definition, extra-header rules, and validation and conversion software for authoritative implementation and archive workflows.
Inspect a NetCDF classic file · Inspect a DDI Codebook · Browse scientific-data tools