I have my camera set so when I shoot it saves both raw and jpeg
My early p&s or bridge cameras saved jpeg only
What I am finding now that my skill level and experience has increased ( photoshop and camera use) I am want to go back and edit/re-edit some of my early images
The j peg files are limited in what I can do with them .
Someone said jpeg was brought in for archiving of images
Yes this is probably true and was back in the days when mem storage was limited in the days when we had 2 digit year code yep pre y2k
I move save edited work as Tiff files ok it eats more storage mem but I can go back in “n” amount of time and re edit if I want to
And as for the jpeg is universally accepted most pc/Mac can read Tiff
Ok your phone and tablet may not, but this is where planning your shot/photography session comes in
And you still have the option of converting Tiff to jpeg
Ok I have waffled on
Like all advice: use the bits that will help YOU and feel free to disregard the rest
So I think there's enough interest that we should take a closer look and understand what's going on. Let's work through an example.
First how JPEG functions: JPEG can achieve extreme compression rates of beyond 90%. Digital data compression works by removing redundancy. Look at language as an example and consider what I'm writing right now. I'm typing a lot of spaces. I can't replace the spaces with some other place holder and come out ahead but look just in this paragraph at how many times the character pair [space a] occurs. I count seven times. So that's 14 characters that could be replaced with 7 characters. Multiple instances of [space a] is redundant.
JPEG does it's job by creating redundancy. The JPEG algorithm places an 8x8 pixel grid over the image. That's 64 pixels in a grid cell. In a photo the data is dense (not much redundancy) and odds are those 64 pixels are all unique. JPEG goes to work altering pixels so that a grid cell will contain redundant pixels. A grid cell can be transformed from 64 unique pixels to only 24 unique pixels. NOTE: This is not a bad thing. We need JPEG and it works great.
Once JPEG does it's job it's not reversible. The pixels that are changed can't be changed back. The compression grid in a JPEG is invisible to us when we view a JPEG because it's so small and because the algorithm did a great job blending it into the image.
However if we make any substantial changes to the tone/color of the image what basically happens is those changes interact with the compression grid and begin to make the grid visible. We undermine the job the algorithm did blending the grid into the image -- nothing can be done to avoid this.
Over time this is a diminishing problem for multiple reasons. The damage done editing a JPEG was much more visible when the original was from an 8 or 10 megapixel camera. Today I have 26 and 45 megapixel cameras. If I edit JPEGs from those cameras the same damage occurs but I can't see it because it's so small. We've swamped the problem under increased resolution. Another reason the problem has diminished is our viewing habits. We continue to look at smaller and smaller images and again the damage done editing JPEGs is increasingly invisible. In that sense it's completely fair to say, who cares I can't see it.
The example: I selected an older photo from a 12 megapixel camera so we can see what's happening. First here's the three photos so you can see the need for the editing in the first place. NOTE: In the camera's defense I made no effort to get a good JPEG from the camera. I shoot only for raw files.
So the SOOC JPEG is flat, too light and cyan/green. That's a substantial alteration of tone/color to produce the third image -- the edited JPEG.
Next here's a 100% comparison of a section of the sky.
The overall blotchy effect you see in the edited JPEG is due to the compression, the editing that interacts with the compression grid and also to some degree the fact that the camera sharpened the image. In my processed raw file I sharpened selectively.
In an edited JPEG from a 12 megapixel camera the damage created by editing looks a lot like noise added to the photo.
One more thing to see: Here's a 200% view of the sky and I circled a spot where you can actually see corners of the compression grid cells. The editing done to the JPEG is making the compression grid visible and at normal viewing sizes that starts to look like blotchy noise in the photo.
Bottomline: The JPEG contained sufficient data for me to successfully edit and improve the image. With modern high-res cameras we're not going to see or barely see the damage that's caused by that editing so big deal. However you must be certain that you will never want to print the image or ever display it large enough to where the JPEG damage will become a visible big deal. The JPEG damage is not reversible.
Other concerns still apply. If the SOOC JPEG has clipped highlights for example you're screwed. If the SOOC JPEG lacks data in the shadows that you need you're screwed. And etc.