The soundscape becomes a research object only by maintaining the difference between an acoustic signal and someone's situated response to it. The Soundscape Indices protocol couples audiovisual recording with in-situ questionnaires and contextual information, setting out complementary collection stages designed to enable datasets from different research groups to be meaningfully compared. It distinguishes descriptors of perceived soundscape quality from acoustic indicators and from derived indices, thereby resisting the collapse of listening into sound-pressure measurement. The operation is infrastructural as much as methodological: it specifies where recordings, observations, participant judgments and circumstances meet, and which transformations occur before a summary number appears. Acoustical engineering, environmental psychology and data standardisation become collaborators rather than interchangeable accounts of a street. Yet protocol portability introduces its own tension. Synchronising an evaluation with a recording makes the research record more accountable, while the prescribed prompts and spatial sampling decisions also delimit what will become audible as evidence. An index can carry a comparison only when the contextual and perceptual inputs from which it was constructed remain recoverable and contestable.