Reading Response 4- Dara Hou

“On Compression” by Cory Arcangel greatly describe what is happening behind the typical file of JPEGs. After reading his article, I came to the conclusion that the purpose of this writing is not just based on compression of files itself but rather how Discrete Cosine Transform (DCT) is a big component to how the JPEGs looks and works. He introduced the readers to varieties of terms; these terms come with explanations and examples to give the readers a clear understanding of how everything works and how it relates back to file compression.

Throughout this article, I liked his explanations of the two kinds of compression, Lossy and Lossless, not simply because it is easy to grasp the information but also how he use that to nicely transition to the Discrete Cosine Transform. Before going into the DCT, he gives a brief summary of what the two compressions do. For example, he explained that “Lossless compression does not lose any information from the original source” (Arcangel 221). Because it is “lossless”, this kind is used primarily to send texts. After allowing the readers to get an idea of what it is, he backs it up with an example to further clarify it. He then proceed on to describe the second kind of compression, Lossy. Instead of texts, this is used for images, video, etc. After a brief explanation of the Lossy compression, he transition to DCT, which is a math formula used in Lossy compression (Arcangel 221).

Because his main focus on this paper is about DCT, many formulas and math calculations were thrown across his process of explaining. It is important, however, to know his intention for the readers- that it is important to just understand the basic idea. Like he said, “if you want to bring the noise, the math is in the end notes” (Arcangel 220). From my understanding, the basic idea is to understand the four steps of DCT-based compression, since that is the root of how all this compression process works. Arcangel walk the readers through the four steps, beginning with creating the basis functions. It then compare the basis functions to the input data, making a DCT coefficients. The coefficients then quantized. Last but not least, the final step is to decompress. (Arcangel 222). So what is the point to understanding this process? Perhaps it is useful to know if you want to save tons of space using when compressing a file.

From reading this paper, Arcangel provided me much more information about JPEG and DCT. He goes into detail of what is happening behind JPEG files. When I compress a file, I never thought about what is actually happening. After getting a better understanding of how it works, I will look at it differently next time I compress a file to post on this blog.