It’s just a loop that runs the command for every single text file. That’s all.
Hmm, I’m really not sure what this SOHO notes app does, but the file you sent looked like a UTF-16 file with BOM, although it was not even a real text file. At least the mime header was off and classified the file as data and not text.
The iconv command I used just converted the utf-16 to utf-8.
The Chinese characters are a side effect of wrong endianess conversion. UTF-16 has an endianess: Little or Big Endian format. UTF-8 does not. So, if the file is UTF-16LE but you convert from UTF-16BE to UTF-8, the characters look like Chinese characters, because the byte swapping shifted the characters to a higher unicode table, which happens to be one that holds Chinese characters.
You could try to use
-f utf-16le instead of
-f utf-16, but I can’t promise that it will help. I suspect that the SOHO export is a bit wonky.