Yes, there's a general voicing assimilation to the preceding phoneme for both -s and -ed! (plus an additional epenthetic vowel insertion if the suffix consonant is too similar to a preceding consonant) It took me a long time to learn this because I am a native speaker and did it automatically without thinking or initially without even noticing that it was realized differently in different contexts in the first place. It's a really fun example of how we can do fieldwork on ourselves (gather a lot of data and try to derive the rule).