Hey there @ErikaGanda!
I think the safest way to figure it out is to ask the sequencing center, or the source of the sequence data. The problem is that the two ranges of phred scores overlap, so really low quality phred 64 data might look like high quality phred 33 data (or vice versa). Depending on the nature of the scores, you might be able to figure it out based on the signal (for example, if there are quality scores outside of the range of overlap, then that can point you in the right direction).
@Robert_Edgar has a nice write-up on this, here (as well as some suggested tools to assist you): https://drive5.com/usearch/manual/quality_score.html
As well, q2-demux’s
summarize visualization will also attempt to warn you if it thinks the wrong phred offset was selected on import.