For most of the versions from which I read the code (I didn't take a look at all of the implementations), the default alphabet contains 2 'f', 2 'C' and 2 'H'.
The consequence is that if you call Encode with the default params you have a value, and if you call it with the default alphabet you have another value. The uniqueness of chars should be ensured even if the default params of Encode are used.
Hi Remi, you are absolutely right. I will fix this in the next few days. Thank you for finding this!
Still have to fix for other languages, but PHP should be fixed ;)