For RTL text this export method is not changeable, otherwise you can select another font mode. The best support for unicode text, the CIDFont mode provides.
But in your EMF file something is weird. I insert it into Word and tried to print to Destiller - that crashed. I opened the EMF in a graphic program and only saw random characters.
When you say, you can select the text in the PDF file created by another tool, did you try copy and paste? After pasting did the correct text show up?
Inside wPDF I do not get any usable character information from emftext.emf. The text is printed using index values which point to glyphs of the font being used. Unfortunately I know now way to convert those index values back into text.
But like I said, when rendering the very same metafile on Windows XPS Printer or Windows 10 PDF Printer, the text is selectable. It looks like it's only with wPDF that the text is rasterized. If we can't resolve this, we will have to use another PDF component (i.e.: PDFsharp).
When we draw the same metafile on another PDF Printer (iCutePDF Printer, Windows 10 PDF Printer) or Windows XPS document printer, the text is selectable. It seems that this happens only with wPDF Control.
Maybe we're not doing something right when calling the wPDF control component? Would you like me to send you the metafile in question?
wPDF will print text as text and not rasterize it. This can only happen on some text which was print using a font which cannot be found or embedded.
Regular text will be exported "as text". It is also possible that the text is already rasterized (or outlined) in the input data - this can happen with GDI+.
