In your example document a character encoding is declared as UTF-8, but the barcode string itself is defined with windows-1251.

That inconsistence results some double byte sequences are interpreted wrong way (as Chinese glyphs, not present in the TTF font file).

There are two solutions:
1. to change the document encoding to windows-1251. As I see, it is not accepteble by you.

2. to represent the barcode string as correct UTF-8: to change 153 (0x99) -> 0x2122, 151 (0x97) -> 0x2014, etc
(See the bottom part of the code table on http://freebarcodefonts.dobsonsw.com/Code128Transformation.htm )

The codes can be written using HTML entities i.e. ™ — …