Research on advanced L2 adult learners and viable classroom instruction for them has become increasingly important along with increasing global connections. This study investigated the effects of different kinds of spoken input modification on listening comprehension and incidental vocabulary learning by 106 advanced Chinese speakers of Japanese in Japan. The participants were randomly assigned to four types of input (genuine, simplified, elaborated, modified elaborated) used in four short academic talks by Japanese professionals. Each talk contained eight low-frequency nouns, each appearing three times. Learning outcomes were assessed using three different measures: form-recognition, meaning recognition with contextual information, and meaning recognition via L2 definitions. Participants responded to three types of comprehension questions (replication, synthesis, inference) while listening to the talks. Scores from an online proficiency test and two working memory (WM) tasks served as moderator variables.Results showed that elaborated input was the most effective of the four types for both comprehension and incidental vocabulary learning. Results also showed that modified elaborated input, a novel input modification type that contained the same elaboration but with shorter sentences, was more effective when higher WM was available. In contrast, elaborated input was least influenced by WM capacities. Regarding the relationships between input modification and type of comprehension questions, modified elaborated input had a marginally significant effect on replication items. For synthesis and inference items, statistically significant effects for input type were not found, contradicting previous results in the literature. Proficiency showed significant effects on all tests, whereas WM showed interaction effects with simplified and modified elaborated input. In light of these findings, the study concludes that (a) elaborated input is more beneficial for advanced L2 learners than genuine input regardless of WM, (b) modified elaborated input with short sentences requires WM, (c) input elaboration is more effective than input enhancement for incidental vocabulary learning for both form and meaning recognition, and (d) enhanced incidental vocabulary conditions using greater input elaboration are likely to provide L2 learners with better input and opportunities to learn more lexical items incidentally.