I have seen these blurs all the time with various compression algorithms, and thanks to the lecture I finally know in detail where this comes from. Yet at the same time, I am wondering is this artifact specific to JPEG style encoding or does it generalize to most of compression artifacts?
o0WeiyuFeng0o
The blur came from compression or lossing quality. This slide shows the significance of thinking images in terms of frequency. High-frequency will be filtered out when doing the compression.In addition, since there are so many different image formats, I am wondering how image format affects the artifact. Does different image format process the low and high frequency regions of images similarly?
aramk-hub
May be a dumb question, but how does copy pasting an image work? For example, copying an image and pasting it into a pdf (not adjusting the pasted size). Suppose original image was a high quality JPEG, do we lose information or does it stay the same when pasted into a pdf for example?
AadithSrinivasan
I was working on doing some opencv with some jpeg screenshots and I definitely noticed the compression artifacts when I zoomed in closely on small parts of the images like numbers and people's faces. At that time, I was confused since it looked perfectly clear when not zoomed in but now I understand why those are there.
rubywerman
https://squoosh.app/ this is a great tool for better understanding different image compression algorithms
melodysifry
Is there a way to extend the idea of filtering out high frequencies to reduce the appearance of the 8x8 pixel block boundaries?
I have seen these blurs all the time with various compression algorithms, and thanks to the lecture I finally know in detail where this comes from. Yet at the same time, I am wondering is this artifact specific to JPEG style encoding or does it generalize to most of compression artifacts?
The blur came from compression or lossing quality. This slide shows the significance of thinking images in terms of frequency. High-frequency will be filtered out when doing the compression.In addition, since there are so many different image formats, I am wondering how image format affects the artifact. Does different image format process the low and high frequency regions of images similarly?
May be a dumb question, but how does copy pasting an image work? For example, copying an image and pasting it into a pdf (not adjusting the pasted size). Suppose original image was a high quality JPEG, do we lose information or does it stay the same when pasted into a pdf for example?
I was working on doing some opencv with some jpeg screenshots and I definitely noticed the compression artifacts when I zoomed in closely on small parts of the images like numbers and people's faces. At that time, I was confused since it looked perfectly clear when not zoomed in but now I understand why those are there.
https://squoosh.app/ this is a great tool for better understanding different image compression algorithms
Is there a way to extend the idea of filtering out high frequencies to reduce the appearance of the 8x8 pixel block boundaries?