If ABBYY is unable to recognize text, why is it returning a DOCX (we request for docx in the request) that's empty?  Shouldn't it be returning a null response?