JPG Compression — Part 2

Iterative Compression, With Changes
posted April 30, 2006

This is an attempt to address the issue Johno raised in his Comment to my previous article “JPEG Compression — Devil or Angel?” If you haven’t read that article it may make it difficult to understand this one, so I urge you to read it first.

In the earlier article we established that repeatedly saving the same image at the same or less compression does not further degrade the image. But, when changes are made to the image between saves the effects of the compression will be somewhat different, raising the question of whether, and by how much, this causes further observable degradation. For this test I repeatedly made a modest (not small and not major) Curves adjustment to the image. Though each adjustment was modest, the cumulative effect of these adjustments was a major and unacceptable increase in image contrast. For the reference/control I repeated the adjustment to the image in PhotoShop four times and then saved it with the same moderately aggressive compression I used in the first article (50 in PhotoShop’s Save For Web). For the test image I started with the same image used for the control (the original from the earlier article). After making the first adjustment, I saved it (50 in PhotoShop’s Save For Web). I then loaded this first generation JPG image in PhotoShop, repeated the adjustment, and repeated the JPG Save For Web; creating a second generation JPG. I repeated this procedure two more times creating the third and fourth generation JPG.

To be clear: for the control, the adjustment was applied four times, ending with one JPEG save. For the test image, after each adjustment the image was saved as JPG — this JPG was then used to make and save the next JPG, etc., such that the final test image included four adjustments and four JPEG compressions. The two images below show the result. Though the two are different I’m not sure which is better. In some areas the control looks better, in others the test image looks better. Neither seems obviously superior to me overall. Both seem about the same amount different/distant from the original — your opinion may differ.

Images below are magnified by three.



Test Image — 4th Generation JPEG

Below you can compare the single JPEG compression from the original article with the control image in this test. Each image has been subjected to only one JPEG compression, but the contrastier reference/control image has more conspicuous JPG artifact. I included this comparison in case you were thinking, ‘Wait! Something else is different — the artifact are more prominent in both of the images in this test. They are; apparently because the image is more contrasty.


1st generation JPG of original



Original X3

I have included the original image so you can see where we started before any adjustments or JPEG compressions.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: