What I’m working on now is a further step in my library project, trying to get closer to my ultimate goal.
I wanted to show a new way of converting images into musical notes. Not just by pixel colour and light and dark, but by adding a way of simulating how images are read in reality – by discerning their content. This involves complex deep learning and for now I just want to show the possibilities of this approach.
And what I’m trying to do is to [roughly] record the image with notes, a process more like serving to read the image stream than for the sake of reading the image, i.e.: reading the image more quickly. If the poor image is seen as the result of image dissemination (multiple downloads and uploads of images), can this process be seen as a compressed and efficient dissemination? (Of course not, and by the way the carbon footprint of the internet, which requires energy to transport and store.)
以及我试图做的是用音符去【大概地】记录图像，这个过程比起为了阅读图像，更像是为了阅读图像流服务，即：更快速的阅读图像。这个过程是不可逆转的，或者是会产生差异的逆转过程，因为这实际是将一张图像直接变成了poor image。如果把poor image看作是图像传播的结果（图像的多次下载与上传），那么这个过程可以看作是一种压缩的高效率的传播吗？（当然不是，顺便一提互联网的碳排放，网络是需要能量来运输和储存的。）