Export format differences

Full format

Data continuity is ensured by filling the lost data packages (if any) with null values. The signal values are aligned to a common starting point (padding might be applied at the beginning and end using null values).

Timestamps and sequence numbers per sample are not provided, note however that metadata contains the initial time point estimate (t0) so that the time correspondence of every sample can be computed.

One CSV file for each signal:

• Col 1..n_chans: signal values

One CSV file for each event:

• Col 1: timestamp (high res., us)
• Col 2..n_val: event values

Metadata file (info.json)

• Initial timestamp (t0) is computed, common t0 for all signals

Lite format

Data continuity is not ensured. Timestamps and sequence numbers per sample are provided. In addition, it also provides the corrected timestamps based on the initial time point estimate (t0).

One CSV file for each signal:

• Col 1..n_chans: signal values
• Col (n_chans+1): sequence number
• Col (n_chans+2): reception timestamp (high res., us), might be null if lost online
• Col (n_chans+3): corrected timestamp (high res., us), using the t0 computation

One CSV file for each event:

• Col 1: timestamp (high res., us)
• Col 2..n_val: event values

Metadata file

• Initial timestamp (t0) is computed for each signal