In the upper reaches of the Minjiang River, both the Xiaoheishui Tibetans and the Xiaoxing Township Qiang people preserve a form of multi-part folk singing. What are the similarities and differences between their musical characteristics, structural formation, and generative contexts? Using modern sound analysis software and notational transcription, this study draws the following conclusions: Tibetan multi-part folk singing features micro-multiphonics and macro-vibrato, whereas Qiang multi-part folk singing exhibits macro-multiphonics and micro- vibrato. The vibrato in both traditions is formed through a combination of “pitch fluctuation” and “volume fluctuation.” When examining vibrato within the multiphonic structure, it is found that the macro-vibrato in Tibetan singing serves to expand its relatively compressed micro-multiphonics, thereby alleviating dissonant effects. Conversely, the micro-vibrato in Qiang singing fills its more spacious macro-multiphonics to enrich an otherwise hollow sound. The generative context for the dissonant effects in both traditions lies in the communal need that encompasses both survival and social cohesion, manifesting concretely in the production of substantial sonic energy to drive away fierce animals and to strengthen group identity.