Dear Iconem,
our developer looked at this problem.
He found out, that the specification doesn’t define, if the strings should be 7bit ascii or utf8 or latin1.
The User ID field is ASCII character data that identifies the user
Description
Optional text description of the data. Any remaining unused characters must be null
UTF8 is mentioned only in case, if is used OGC, but we are using GeoTiff.
But there was a bug about latin1 info writing into file. It writes into file that there is more VLR blocks as in the reality. This bug was solved and should be implemented in one of the next releases.
Regarding the question, the 1.4 spec says that “string (a variable series of 1 byte characters, ASCII encoded , null terminated)” (page 4).‘’