[quote=“Jeff-Foliage, post:1, topic:9894, full:true”]
[…] I would like to know how much I might be losing by going from Max Qual to High Qual? (What are the diminishing returns for Max versus High)[/quote]
Hi,
JPEG compression is lossy, that’s how it can create much smaller filesizes. But the amount of compression also depends on image structure and amount of detail (or noise). The more detail, the less compression.
The lower the quality settings, the more aggressive information will be discarded before doing another compression step. The first to go is color accuracy in fine detail, and then ultimately also luminosity will start to suffer. It’s a gradual deterioration, while attempting to first discard information that human vision is less sensitive to notice.
It’s not really possible to predict that amount of compression accurately without actually doing it. So for each quality setting the program would have to actually compress the image in order to report it’s size. That would take time and result in a not very responsive updating of file size values.
So for the quality, one would ideally have a preview at 100% or larger to allow a before/after comparison, where one could select a region of interest with lots of fine color detail, or also smooth gradients which may start developing posterization artifacts.
For an update of expected file size, one could have an option for more accurate calcuation, based on an actual compression without really saving it yet, and doing it again when actually saving it.
For a desired maximum file size one could have an option, to have a slow under-the-hood kind of trial and error for a given target file size.
Cheers,
Bart
P.S. For those seeking to maximize filesize reduction without visual degradation, Google (amongst others) have developed a utility called Guetzli, that will analyze the image extensively, replaces some high detail areas with slightly better to compress detail, reduces color accuracy within the limits of getting too obvious, tries different compression rates, and finally produces a significantly reduced output size JPEG that’s visually hard to distinguish from a 100% quality JPEG. It also discards meta data.
Starting with a (recommended) 100% quality JPEG as input, it can often reduce the file size to between 10 to 30% of the original size (depending on image content). But it takes many minutes to do so, and larger sized images can take a very very long time.
