I checked the attached file, and do see the corruption. I see it with 1.5 SW, 2.0 SW,and the Microsoft Codec as well.
I am interested in the case where you DID see it play cleanly. You say its Ok in 2.0? How was this file encoded?
As for it being slower. If you are using 1.5 GOLD with HW that support it, then switching to 2.0 Beta its expected that the performance would be different. The samples are using the 2.0 API and theres no HW to support it out yet, so its falling back to decode using the CPU.