converting latin to utf8mb4 causes questionmarks

  • The original format of the data is unknown
  • The new table is in utf8mb4_general_ci

If I do CONVERT(BINARY CONVERT(column USING latin1) USING UTF8) as mentioned here – it fixes all text, but converts something like: © in the original column to ? in the new column.

If it helps to determine what original encoding it was in, the original text renders as e.g. KotaÄići and converts to Kotačići.

Is there a way to both preserve special characters and restore correct utf8 text format?