Part of the JPEG image format standard includes a "Quantization Table" (an 8*8 matrix/array of values), this essentially helps compress an image by going over each 8*8 block in a given image with it (lower values causing less compression). Below is the table often used based on psycho-visual experiments, but by using the sliders next to the table, you can manipulate it, then click the Quantize button to see the results to compress the image how you deem fit.
⬆ Click the image to zoom in (sample below) ⬆
_
Low
Low-Mid
Mid
Mid-High
High
Sample size 16*16:
Upload your own JPEG:
⬆ Click the image to zoom in (sample below) ⬆