This article investigates Oshii Mamoru's experiments with voice and vision in his film Ghost in the Shell (1995). The audio-visual inversion articulated by the disembodied voice in the film dissolves the conventional image-voice conformity. The inorganic gaze adopted by Oshii breaks out of the human observer's body to spread into space, unlike the anthropocentric gaze standard in cinema. While this depersonalized sight expresses the subject's dissemination, it also echoes the film's motif of mankind's ontological opening up towards the environment. Oshii's audio-visual experiment can be considered a critique of Cartesian optics: whereas the bodiless voice undermines the Cartesian domination of vision over other senses, the inorganic gaze produces a non-perspectival space and non-human vision supplementing Renaissance perspectival systems. Consequently, Oshii's tendency to separate the voice from vision renders his animated bodies as heterogeneous, discrete agents distributed through multiple spatiotemporal dimensions beyond classical constructions of subject--object boundaries. The result of his challenge to Cartesian optics produces in the audience an intense affectivity reminiscent of Eisenstein's ecstasy' of animation.