Wikimedia Foundation accepted the grant that paid for the usability initiative, it came with restrictions. The money was to make the English Wikipedia more usable. It does and it did make significant improvements, the "vector" skin is a vast improvement; I use it where I can. The en.wikinews made it its default skin and I expect that many projects will follow.
While the new developments were not targeted to other projects and languages, they did benefit. Test projects were started for other languages and people were invited to test and to comment. As I understand it, the "vector" skin is better at supporting right to left languages like Arabic.
The multimedia usability project being next, does not have any language restrictions and arguably, if the quality and quantity of Commons is to improve, it is the "other" languages and cultures that are not covered as well as the Western languages like English, German, French and Dutch.
To improve Commons, I would concentrate on one western language; English. One language with a right to left script and one language with a Cyrillic script. The choice for Arabic and Russian would be obvious. Given the expertise at the WMF, I might add Japanese to make it a really nice mix.
Each of these languages pose their issues. Each of these languages represent big vibrant projects that show a healthy growth in traffic. When money is to be spend on usability tests, research in these languages and cultures will have more impact then testing additional western languages. Personally I would consider looking for a partner in a country like Indonesia, they are more likely to have expertise in Asian languages then companies in San Francisco.
In her letter to the board on strategy, Sue argues that our community is to diversify. It seems to me that the multimedia usability project has the potential to have a big impact in reaching out to those places where we want to do better.