Reverberations

Kai-Luen Liang

1949: Tiananmen Square: The ghosts of a public square filled with bodies caught up in the fervor of victory. The sound of a single voice through multiple loudspeakers echoing throughout Tiananmen Square. The voice bouncing off the surrounding concrete, no longer dampened by human bodies. The founding of the People’s Republic of China was marked by this grand celebration of victory over the Kuomintang and western imperialist powers. Mao looked down at the square from above and let his voice ring out after each phrase as if to allow the space between the echoes to transfer into the bodies and minds of his audience. As this historic speech was uploaded and poorly transferred into digital space, the sounds of the reverb and echo become compressed and augmented. A metallic ring stretches the space in between words and myth.

I remember hearing this famous speech through poorly transferred audio recordings as a teenager. I never remembered or could decipher what was actually being said with Mao’s heavy Hunanese accent. But the sound of the echo and the pauses in between remained etched in my mind as an audio map of a mythologized moment. The space seemed infinite and Mao’s voice perforated through any form of matter. The faces coming from the audience looked as a nebulous forgotten cloud leftover by the last rain. Distant faces in foggy memories all pierced with the knives of a loud voice.

In this piece, the actual words from the audio were cut out leaving only the tails of the amplified and stretched reverberation of Tiananmen Square in a single moment in 1949. The space in between words blur old boundaries. The nameless and faceless coalesce into a digital sound mass.

This short film looks at maintenance of the sound and image of Mao’s China through CCP propaganda posters made from the 1950’s to 1976 and the audio reverberations of Mao’s now iconic Tiananmen Square speech in 1949 . A dataset of over 2000 propaganda posters were collected and carefully prepared as training data into a Generative Adversarial Network, an AI/machine learning model. The machine learning model learns the features of the input dataset and attempts to recreate new images from the features it has learned from the data. The machine thus becomes involved in socialist neural dreaming blurring the boundaries between the real and the fake, history as fiction, engaging itself in false binaries.

Machine labor through clocked human hours left forgotten.