It's a similar argument, but as with all engineering questions the details matter. The human auditory system can be entirely fooled with 2 channels of 40 kSamp

It's a similar argument, but as with all engineering questions the details matter. The human auditory system can be entirely fooled with 2 channels of 40 kSamp/s 16-bit data: 1.3E6 bits per second. The human visual system is much more difficult. It's hard to put a hard number on it due to the stochastic nature of "pixel" sampling in the eye, but generally 1000 Hz is accepted [1] (pulled from blurbusters). I'm also going to gloss over foveated rendering since I'm assuming a one-way "trick the system" box. The human visual system would need 2 channels of 7500 x 7500 (120 degree / 0.016 degrees per pixel) x 1000 Hz x 3x16-bit color: 2.7E12 bits per second. To get to my point: it is technologically trivial to work with audio compared to video. Audio simulation and reconstruction has been a mostly dead area of research for decades while video is still flourishing. Audio did not stop at the actual technical best place it could have, but when it was good enough and diminishing returns hit. 1. https://www.lrc.rpi.edu/programs/solidstate/assist/pdf/AR-Flicker.pdf#page=6 https://www.lrc.rpi.edu/programs/solidstate/assist/pdf/AR-Fl...