It appears that the internal strings encoding between Skyrim and SkyrimSE is not the same.
Skyrim used a different codepage for each language while Skyrim SE seems to use only UTF8

Skyrim SE seems to be able to read some ANSI set (at least the 1252 one), but this is not consistent. It's a possibility that the game tries to detect the right encoding for each string, but since such detection can't be 100% accurate, the process can fail, and falls back on utf8.
So it is better to use utf8, no matter what.


Anyway, this can cause some issues if you try to load a set of strings from the original skyrim (because you modified it for any reason) or if you try to compare an old esp over a new one: because the encoding will not match. (you will get some strange characters for any special char)


Here is how you can fix this issue:

Let say you want to use an old set of strings for Skyrim.esm:

-open the translator in sseTranslator workspace and set the source and destination language (1) as usual
then  load Skyrim.esm (version SE)
-go to menu -> tool -> load String as translation (2) , choose your old set of Skyrim strings
-choose to overwrite everything (3)
-then, this is the important part: change the codepage (4)  by the one you used in the original skyrim (see below for a list of codepage vs Language)
-click OK

Now you can finalize your old set of string, it will be automatically converted with the right codepage.
you can also save your dictionary for Skyrim. (ctrl-S)


The process is exactly the same with the espCompare Tool (5): 
-load a esp Special Edition
-choose menu->tool->espCompare Tool (5) and load its equivalent version Original Skyrim
-choose the old codepage.

Note: if you try to translate an esp that has not been correctly updated for SkyrimSE and ask for the SkyrimWorkSpace in sseTranslator, you can translate it though the process explained above, but you need to use the 'Force Utf8 on export' option ( menu -> Options -> utf8)  when you finalize the translation