The EncodingValidatorEx first checks the BOM and, if this is not present, checks whether the text is UTF-8:

The EncodingValidatorEx is designed for fast and memory-efficient UTF-8 detection. It scans the stream and checks whether it contains any UTF-8 characters. If it does not detect any, it uses the configured encoding.

The EncodingValidator, on the other hand, reads data blocks for each encoding to be checked, as well as for UTF-8, and validates them.

For small messages of, say, 10 KB, this difference is negligible. If you mainly process messages of 10 MB or more, the EncodingValidatorEx is the better choice in terms of speed and memory usage.