I'm wondering if any you that have worked with CAT tools and have used translation Memory have received consistent feedback about the end results being too literal. Some of our customers have complained that our translations sound similar to Google translate. How have you gone about fixing this issue? I welcome any comments or suggestions.
Yes. I also received such complaints. I explained them that the contexts required literal translation e.g. legal or technical paragraphs. No more idiomatic translations were necessary.
[Google might learned out of translation for its MT processes as well.]