Acrobat replaces single-byte character codes with octal strings in content stream, how to prevent?

Question

I have thousands of PDFs, mainly text, all fonts are subsets, indexes starting from 1. In page content stream I see character indexes represented as single bytes: binary 1, 2, 3, etc. But, after optimizing files with Acrobat (I use it to save to 1.6 version with compressed object streams etc.) those single bytes become octal strings like \001, \002, ... . Raw stream size is increased by factor 1.5, and after a flate filter there is still 1-2 kB increase for each page. As there are thousands of pages, total (and totally useless) increase is considerable and doesn't go well with my goal of creating as small package as possible.
Can this behavior be prevented (with something like registry setting)?

Answer 1

what are you calling getBytes() on?he's calling getBytes() on his String object.
@ puneet_k,
I don't have a solution but note that on Unix-like systems a newline is a single byte 0x0a and on DOS-like systems (Windows) it is a sequence of bytes 0x0d and 0x0a.
Some text editors handle both.
I'm not sure but I think on Unix,
new String("\n").getBytes() equals new byte[] { 0x0a }
and on DOS
new String("\n").getBytes() equals new byte[] { 0x0d, 0x0a }