These days I encounter this type of situation less and less, but I've always wondered...
If your source image is not ideal and you're given a lossy compressed JPEG, when you edit and export the edited version, how can you match the quality setting of the source file?
If you choose a quality setting lower than the source, you've made it even lossier. But if you choose a compression quality higher than the source, you've added unnecessary bloat to the file size. In plenty of cases it would be ideal to match whatever the compression quality was set to when the source file was saved - but is there any way to derive what setting was used when saving the source image?
It should be acknowledged that there are also many cases when one would want to save with lower or higher quality based on how the image is to be used or delivered: Highest quality possible when performance is not a consideration; and and lowest forgivable quality when performance is the highest priority. But I mean in cases when we don't want to significantly alter the file size, only part of the image content.
Then there is the side issue of: Does saving a compressed JPEG compress the compression? Is block noise getting added on top of existing block noise, or is it just spreading to other areas that previously were not noisy looking? And how much of this depends on other factors (like if the software used to save the source differs from the software used to export the edited image)? There is plenty to analyze if we want to get deep into it.
BTW, in such a situation, I typically "eyeball it", while aiming for a file size close to that of the source. But the geek in me wonders...
