Just in case I've found a problem.
My SEGY file used to have elevation and coordinate scaler header = 0.
Once I changed it to 1 it started to work properly.

Anyway it is strange that headers given from container differ from headers given by seis block. Or maybe it is for made intensionnally for the speed purpose...

The scaling is used yes, but you can turn it off with scale=false. I'll double check the consistency

I just figured out that get_header(SeisBlock) return one column vector of values for each trace while get_header(SeisCon) return two column matrix: 1st col min value for the shot and 2nd max value.

Maybe in the future get_header(SeisCon) could be renamed to something like get_header_limits(SeisCon).
Though it is not really important.

That's a fair point. We wanted to be consistent while keeping the lookup table light (storing all the headers of a SeisCon would be quite big). Not sure what a good alternative would be but will think about it