Towards the goal of using passive acoustics to monitor biodiversity and ecosystem health, the need for more extensive fish sound libraries is increasingly clear. Diverse habitats, such as coral reefs, where fish communities are complex, present a particular challenge: most recorded sounds remain unidentified, limiting ecological interpretation of soundscape patterns. An estimated 96% of fish sounds remain undescribed, and existing characterizations are often based on single observations, instruments, or locations. Moreover, while machine learning and AI-based approaches are emerging to automate soundscape analyses, their utility is constrained until more fish sounds are reliably identified and catalogued. Expanding sound libraries requires more replicated field observations, even for species already described. Conventional SCUBA-based surveys, however, hinder acoustic data due to exhalant bubble noise and behavioral disturbance. Autonomous video-audio systems with hydrophone arrays offer effective localization and identification of sound producers, but they remain limited in scale and availability, particularly in low-infrastructure settings. This chapter describes “bubble-less” observation approaches designed to improve fish sound documentation in understudied tropical coastal habitats, with examples from East Africa, the Caribbean, and Southeast Asia. No single method can meet all needs; instead, a suite of complementary approaches is essential for building globally robust and diverse fish sound libraries.