The soundtrack for the 2012 cult indie hit Dear Esther has been praised for its emotional power and innovative fusion of digital effects and live instrumentation. This talk will explain the process of evolving the soundtrack from the original 2007 mod, which was built using only digital samples, to the remake's lush orchestration. The talk will focus on how to create rich and deep soundtracks on a tight budget, and our process of design, placement, and optimization of the music as a device to focus the player's emotional journey. The talk will begin by explaining the compositional philosophy of Dear Esther's music (finalist in Excellence in Audio at the IGF 2012), and how this related to the vision and design of the game, from story to environment art. This will feed into a discussion of the compositional process itself, particularly, the fusion of digital samples, effects, and live instrumentation. The relationship between music and environmental audio, including the use of quirky and strange semi-musical audio cues that offset the landscape and soundtrack, will be explored, as well as how distortions, splices, and manipulations can bring a unique tone and flavor to game music. The talk will conclude with thoughts, based on the experience of Dear Esther and other thechineseroom games, on how composers can communicate with developers and designers, and vice versa. The success of Dear Esther is frequently ascribed to the deep, holistic fusion of art, story, and music. This core of the talk is how this was achieved, and how studios can create powerful, integrated scores, even on a low budget.