I'm having trouble with weird results when indexing old .doc Word documents. docx works fine, but the doc files all appear like this in the search results:

���� � ���NAMN PŠDOKUMENT�Datum�Beteckning�����2015-01-01� �Sida � PAGE \* MERGEFORMAT �1� (� NUMPAGES \* MERGEFORMAT �1�) ���� ���� ���� ��

Has anyone got this to work? We're on EPiServer 7.13 and Find 8.8 if it matters. The Excerpt field is set with

conventions.ProjectExcerptUsing(spec =>
    doc => doc.SearchAttachment().AsCropped(spec.ExcerptLength)
);

Any help appreciated. Cheers.