Abstract This chapter discusses the preparatory work of creating a collection of corpora, together functioning as a multifaceted corpus of the same sign language, across three data subsets of signing in different modalities and on different learning levels, namely visual signing (deaf/hard-of-hearing/CODA), tactile signing (deafblind), and L2 signing (hearing second-language learners). This work, led by native signers and based on personal experiences in the research field, involved planning and adapting data collection and annotation of the Swedish Sign Language corpora over time to ensure a corpus infrastructure that is uniform across data sets, compatible with a parallel lexical database, and available as a searchable resource for the public.