Progress
Week 1
什么才算是好的作品？31/8/2021的我仍然觉得应该是关于潜在的可能性。同时艺术家是有社会责任的，好的作品的出发点应该是去质疑社会相关的问题，通过艺术寻求一种可能性。有趣的是艺术可以逃脱功能、利益性的目的，做一些“毫无用处”的实验。这是不是最能逃离资本的一种方式？
设计作品的过程充满了各种选择，想到和一个老师的聊天。我问，选择的本质是不是知识和认识型？他沉默了很久。我的知识还是太少，难以果断地做出选择，每一步设计都举步维艰，突然明白了前期研究的一点意义...
想了很久，感觉终于找到了真正感兴趣的方向，虽然可能只是暂时的。但是有很多有趣的想法在排队等我的感觉，好棒啊!
Week 3
“We sense and produce the environment through the materiality of language, but being able to re-visit the detailed observations collected decades or even centuries ago guarantees a new perspective of the laborious knowledge construction process.”
“the current informational culture [is] one of light, marked by the weightlessness of fiber optics and the speed of digital transactions”, and yet “it is also one of weight – of minerals, metals, energy consumption, and entropy”
“The ontology of water has to do with the time evolution, contamination, mixing, fusion of materials and permanent recording memories. The water is transformation, metamorphosis, formal regeneration, articulation of life and death.” (Abernú 2013)
“emphasizes an ethics of the encounter that encourages more sensitivity to alternative (more-than-human) time scales and environmental futures: “When the ongoingness of life on Earth is threatened by climate disruption and ecological collapse, new forms of engagement with non-human beings become a matter of life or death."
Timeline of meteorology
云是我们，我们是云，云是一切。云的有知觉的本质。
人类世这个词的形成时间？为什么在这个时候形成？
伴随着科技政治...？（人类整体性的自我认知发生了变化，
与世界的关系的概念发生转变）
Week 2
“巴别塔之前“，”纯语言“，好理想的概念，很动人。人类所追求的相互理解是共通的。
比较纠结的点在于，人类意义上的语言对于其他物种和物质总是包含着磨灭不掉的主体性。可能作品还是不要太个人化吧，还是积极一点比较好啊。如果用GPT-3翻译图像生成的文字，应该很有趣，不过是对人类来说，好像太人类中心主义了。非常不舍得抛弃了这个想法。很想尝试一下各种有趣的技术，不知道能不能用到...
查资料会一直查到不想关但是很有趣的东西。感觉做一个 Internet Web 监测浏览路径的生成视觉也很有趣诶。
可能有用的想法：不同翻译之间的latent space是什么，透露出的内容差异和思维差异是什么，理解的“理解”？好有趣啊...
Week 4
2D Analyse + points / Flow field + OF
OCR (tesseract.js)：
| o
Ry
f k i
S o
OCR：
3 (?)
3D binaryImg overlay / Find points + min Spanning Tree / Planes / depth with FF + connection
OCR：
R
X S AN
o O O
e
Wa
UG
v o
LJJ x
o
生成3d的方式/效果可不可以啊？
For Each Layer
Detailed tech plan:
Collected satellite videos (more clear one?)
Processing video : binary + gaussian blurry + threshold
Analyse points + connection (processed video frame)
Optical flow + 2D flow field (processed video or original one)
Video processing for moving image （hand-draw animation or genearted？如何统一风格！？）
3D depth
OCR detection （processed video frame / 2D）
- 单词排序 or 乱序？
-中文偏旁detection？
Flipbook layout （binary.js）
Composite layers into 3D models
Week 5
CV + ofw + tracking
Original Video
Result with longer steps
Tracking with bg / shorter step
Strokes from the movements of clouds as dataset.
Generate words based on oracle bone script.
Possible presentation (interaction): Input - Chinese ; Output - Clouds words
Week 6
Digital Sky
3D environment
Origin Text (clouds):
made by ancient people
New generated clouds:
Interaction - talk to the clouds / speak to the program and form a new cloud
(e.g. 1 min / per cloud)
communication
AS a response : What's the shape of the cloud? Based on content?
Movement:
Text Clouds: moving and morphing and disappearing
Immersive Interaction:
Navigation? People in the clouds
(like secret garden through the time)
https://secretgarden.stephaniedinkins.com/
Go further... what is REAL Communication
https://xlab.iii.u-tokyo.ac.jp/projects/algaphon/
Realtime Artificial Cloud Machine
Input: speech from audience (through web or onsite)
Output: change temperature (heat) or condition to create artificial cloud (or fog), then let it go to the sky
（Berndnaut Smilde / like scientific experiement : https://www.youtube.com/watch?v=LHjDT9pYxRA）
Week 7 & 8
Process of Environment Changing
& Interaction
(old text: 50 pieces | enviro limit: 50 clouds)
Original: 30 pieces of clouds
Input: <=4 input, put in a same new cloud (record the time)
Maximum: 20 input clouds at the same time
Moving and changing: out of the box - back to text pool - supplement another one (Same type: original / new input)
Unfold Text Reference:
Problems:
1. sky design find reference
2. the texture of clouds - not enough — try impression way of point cloud?
3. unfold text prob
4. set the environment size & cam rotation range
5. speech recognition
6. text resource - more meaningful?
7. installation - how to be 3d????
Waiting List：
1. wind movement - more scientific
2. when clouds get together （lightning e.g. text flashing）- raining - text falling down
3. speech input sending get/post
4. starting page
Week 9 & 10
Get & Post - connect to server
Speech Recognition - form a new cloud
Developing....
Weather, somehow better than language to express / convey feelings
How we connect to the weather to share feelings.
Convey to strangers... Narrative
Machine sometimes can act as understanding us better than human being. Machine listening & Machine company - communication - What's the better communication?
What people says and feels can be saved in "clouds"
"I am..." Every sentence in one cloud. All acceptable
Digital sky - weather - rainy (different levels) / windy / sunny / lightning / cloudy (sounds)
Pass to another stranger
Problems:
cloud visual - test 3d particles & lights
Week 11 & 12
Artwork Analysis:
Writing - CH:
Writing - EN: