Request Encoding impact on Response

Dear all,

I hope I'm in the right location asking this question, if not, please point me in the right direction.

I'm debugging a problem we're having with some middleware uploading PDF's to Salesforce.

I am able to reproduce the problem using SoapUI, but there is some behavior I cannot explain.

I'm doing a simple HTTP GET request to retrieve a PDF, using a SoapUI REST request.
On the left hand side in SoapUI, you can set some request properties.
Encoding is one of them.
If I leave this blank, and I look at the XML version of the response (which is Base64 encoded), the string is correct, meaning that when uploading it to Salesforce the PDF files are displayed correctly.
If I select UTF-8 or iso-8859-1 (which are the 2 proposed values), the result is not correct, meaning that when uploading to SF, I get PDF files with blank pages (The blank PDF files is exactly the bevavior we're seeing with the middleware)
If I enter manually any other value than the 2 proposed values, even garbage like blablabla, the result is also correct.

Now my question: what impact does the Encoding request property have on the Response?
What encoding is implied when leaving blank?
If I look at the Raw request, I see no reference to any Encoding.
If I trace the request using Wireshark, I see no reference to any Encoding.

Thank you for any clues.

Kind regards,
Bob.