Converting PDF charcodes to Unicode (ProcessText in ElementReaderAdv sample)


I have been using ProcessText code sample on PDF files for some time
and it works perfectly. I just ran across a PDF file today that this
routine is not picking up the text/chars as they are displayed in the
PDF file when viewed with the Adobe Reader. It finds the text but the
(char)char_code is all garbage.

Here is my code routine:
       private void ProcessText(ElementReader page_reader)
            Element element;

            while ((element = page_reader.Next()) != null)
                switch (element.GetType())
                    case Element.Type.e_text_end:
                        // Finish the text block

                    case Element.Type.e_text:
                            GState gs = element.GetGState();

                            ColorSpace cs_fill =
                            ColorPt fill = gs.GetFillColor();

                            ColorPt outc = new ColorPt();
                            cs_fill.Convert2RGB(fill, outc);

                            ColorSpace cs_stroke =
                            ColorPt stroke = gs.GetStrokeColor();

                            Font font = gs.GetFont();

                            // Use element.GetCTM() if you are
interested in the CTM
                            // (current transformation matrix).
                            Matrix2D ctm = element.GetCTM();

                            Matrix2D text_mtx =

                            Matrix2D mtx = ctm * text_mtx;
                            double font_sz_scale_factor =
System.Math.Sqrt(mtx.m_b * mtx.m_b + mtx.m_d * mtx.m_d);
                            double font_size = gs.GetFontSize();

                            ColorPt font_color = gs.GetFillColor();
                            ColorSpace cs = gs.GetFillColorSpace();

                            ColorPt rgb = new ColorPt();
                            cs.Convert2RGB(font_color, rgb);

                            double x, y;
                            int char_code;

                            CharIterator end = element.CharEnd();
                            for (CharIterator itr =
element.CharBegin(); itr != end; itr.Next())
                                //Console.Write("Character code: ");
                                char_code = itr.Current().char_code;

                                x = itr.Current().x; // character
positioning information
                                y = itr.Current().y;

                                // To get the exact character
positioning information you need to
                                // concatenate current text matrix
with CTM and then multiply
                                // relative postitioning coordinates
with the resulting matrix.
                                mtx = ctm * text_mtx;
                                mtx.Mult(ref x, ref y);
                                char_code = char_code;
                                    Console.WriteLine("(" +
char_code.ToString() + ") " + (char)char_code + " Position: x={0:f}
y={1:f}", x, y);
                                catch { }


'char_code' does not necessarily correspond to ASCII or Unicode value.
To obtain Unicode value you can use font.MapToUnicode() method:

In C++:
Unicode uni;
int uni_sz;
bool unicode_available = font.MapToUnicode(char_code, uni, uni_sz,

In C#:

char val = 0;
int [] uni = new int[1];
int uni_sz;
if (font.MapToUnicode(char_code, uni, 1, ref uni_sz)) {
   val = uni[0];

Alternatively you can use element.GetTextString() method which will
directly return the Unicode translation.