You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently the slider "compresses" the image by zero-ing out DCT components in order of highest weight.
This is not quite how JPEG works in practice. To apply compression, you divide each component by a specific constant factor. These constants for each component are specified in a quantization table.
The values in the quantization table are chosen to preserve low-frequency information and discard high-frequency (noise-like) detail as humans are less critical to the loss of information in this area.
So it could be interesting to compare compressing using different quantization tables. Or simply as someone suggested on Twitter: zero-out the components by the zig zag order (highest frequency to lowest) which is generally how the quantization tables are designed.
The text was updated successfully, but these errors were encountered:
Currently the slider "compresses" the image by zero-ing out DCT components in order of highest weight.
This is not quite how JPEG works in practice. To apply compression, you divide each component by a specific constant factor. These constants for each component are specified in a quantization table.
From: http://www.robertstocker.co.uk/jpeg/jpeg_new_10.htm
So it could be interesting to compare compressing using different quantization tables. Or simply as someone suggested on Twitter: zero-out the components by the zig zag order (highest frequency to lowest) which is generally how the quantization tables are designed.
The text was updated successfully, but these errors were encountered: