Noise Machine

Software, based on Max8,

Infinite generated noise using only oscillators, from pure numbers.

If you want to download the patch, click here.


2021, Software

Visit The Work

Using Firefox or Chrome is recommended, for Safari version lower than 15.2, fatal bugs might occur.

Visit the NFT of This Work

E-Shredder is a webpage/piece of code, which shows you copies of the most recent sold NFTs, and gradually “destroy” them.
More interestingly, itself is an NFT too, so if you buy it, you can witness it destroying itself.

More details: here

Leung Man Tao: “Naai Gong” or “Gang Du”

2021, webpage/software


Visit the original page.

Below is a embedded version of the work, might need a few seconds to load.

This is a work related to the social movement in Hong Kong in 2019. It is also about the subtle effects brought by the fact that there is different spoken languages and writing script of Chinese, especially in the context of Hong Kong and Mainland China.

The source text is written by Leung Man Tao (梁文道). A writer and critic who is active in Mainland China, Hong Kong, and Taiwan. He was attacked by some people online from both Hong Kong and Mainland China for his neutral stand in the social movement in 2019. He’s the only public figure I know that received attacks from both regions, and after reading posts from the attackers, I noticed that, although the people who attacked him from Hong Kong and Mainland have very different ‘political stands’, behind the stands, they have very similar logics and ways of thinking. This work, to some degree, tries to show, explore, and raise questions about these similarities.

I collected Leung’s essays published in 2019 (the ones in traditional Chinese are published in Hong Kong, and the ones in simplified Chinese are published in Mainland China), tokenized the essays into sequences of words with a customized Chinese tokenizer (which is based on CC-CEDICT dictionary) and put them into a simple Markov model to generate new paragraphs.

Most of Leung’s articles in 2019 are about serious topics like politics, culture, and art, but after going through the Markov model, the generated text appears to be, somehow, meaningless, absurd, confusing, and ridiculous. I like this contrast because it somehow reminds me of the fact that the Internet (which is also the space where the attacks happened) is deconstructing discussions. That is not necessarily a good thing or necessarily a bad thing. It all depends on how we react. With the deconstruction, can we jump out of the existing perceptions and stereotypes? And after deconstructing, can we find a way back to the original question and settle down on an answer?

Reflect on one essay in the input of Leung about Cantonese, and the fact that in recent years Cantonese has been given a kind of political meaning. I wrote an algorithm to randomly choose two words in the text, replace one of them with a word that sounds similar to it in Mandarin (in simplified Chinese), and do the same to the other but with Cantonese (and traditional Chinese), then repeat the process every few seconds. Due to historical reasons Mandarin, simplified Chinese, and Mainland China are closely bound together, and so are Cantonese, traditional Chinese, and Hong Kong, although there is a population in Mainland China that speak Cantonese and write in simplified script and vice versa in Hong Kong. This impression, or stereotype, of spoken languages and writing scripts, creates boundaries for people to discuss and understand. 

The algorithm makes the text change slightly over time. After a while, because of the replacements, the text becomes more and more senseless in terms of the literal meaning of the characters, but since the replaced words share the same or a similar pronunciation with the original ones, the visitors can still guess what words were in the places before and then get a fuzzy idea of the unreplaced text.

What happened to the text is a metaphor for what happened to Leung (and others) in real life: people quote, then ‘interpret’ and twist, what he (and they) wrote/said to support their opinions on him (and them), then A’s twisted quotation is quoted by B, and B’s by C… The meaning of the words in their quotation goes further and further away from the original one. Thus, it is harder and harder for people to get to the original idea of the text. 

It is also a metaphor for the more and more divided society. With algorithms on social media and the instigation from some individuals and groups, the space for neutral opinions is lost. You are either part of us or the enemy. There is no place in between, and there is nothing in common between “us” and the enemy. Thus, the possibility of discussion and negotiation is gone. Only hate and attack are left.

Pronunciation data collected from open Cantonese dictionary, CC-Canto database and CC-CEDICT database, pre-processed by me.

ScreenShots (Click on the picture to view the original resolution version)

To my bio

To the GitHub Repo of this work


2021, 網頁/軟件

English version






回應源文本中一篇討論廣東話和普通話的文章以及以及近年來廣東話被賦予了某種政治意義的這一事實,我寫了一個算法,每隔數秒,將生成的文本中的兩個詞,一個替換為在普通話中同音或近音的詞(替換為簡體中文), 一個替換為廣東話中同音或近音的詞(替換為繁體中文)。由於某些歷史原因,普通話,簡體字和大陸綑綁在了一起,廣東話,繁體字和香港亦然(雖然在大陸有大量講廣東話寫簡體字的人口,在香港亦有大量講普通話/國語寫繁體字的人口)。這種語言和文字的固有觀念或是偏見,實際上為兩地人的互相溝通和理解帶來阻礙。



這也是對這個越來越分裂,“二極管化”的社會的隱喻。 社交媒體的算法以及一些個人和團體的煽動,使得中立意見的空間丟失了。 你要么是“我們”的一部分,要么是“敵人”的一部分。 不存在兩者之間的位置,“我們”和“敵人”之間也沒有任何共同點。 這樣,討論和談判的可能性就沒有了。 只剩下仇恨和攻擊。