Jpeg Compression damage

[William London]

I have 100 jpg images I’ve been working on for wall art printing and have changed their canvas area several times due to how I was going to output. And also dipped in and re-edited a few but always saved at quality 12 as never needed to be compressed and looking back should have immediately made tiffs but didn’t. How much damage have I realistically done? Any idea? As thinking of starting over blowing up the original again out of fear.
Finest
William

 

[JeffK]

I'm sure others will respond - especially @thebestcpu who has the best expertise - but jpg is a lossy format. That means each time you save and reopen the file, you will lose quality. You can make multiple changes within your editing software, even saving it while keep the file open and working on it, and you won't have loss of quality. It's when you save, close, and then reopen that quality is affected.

Constantly resizing also has an effect. One method I've read about is to make your image a smart object and details will be preserved within the editing process. It does get sticky since adjustments can be limited when using smart objects. But again, saving, closing, then reopening does its damage. Also saving it at the same quality level does not prevent loss.

How big is the loss of quality? As most things in life, it depends. There is a loss of detail and also depth of color. It actually might not be that noticeable. I would think that if you open your original, open a copy, and compare side by side, you can get a visual on what the loss is. Zoom in and look at your edges as well.

If you're doing ongoing work with an image, it's always best to save as a tiff or png, both lossless formats. BUT - why are you reworking jpg's? Why aren't you saving as a psd file? You can print from a psd file unless you're sending the file to a print service that demands a jpg. If that's the case, you can alter your psd file, do a "save as" for the jpg, then close the psd file without accepting changes especially if you're just resizing.

Any prints I produce on my own printer are from psd files and I always indicate on the print setup that Photoshop manages the colors.

I'm certainly not the last word on this and hopefully others will weigh in. But it seems you're aware of jpg limitations and follow the workflow that works best for the image.

- Jeff

 

[thebestcpu]

I have 100 jpg images I’ve been working on for wall art printing and have changed their canvas area several times due to how I was going to output. And also dipped in and re-edited a few but always saved at quality 12 as never needed to be compressed and looking back should have immediately made tiffs but didn’t. How much damage have I realistically done? Any idea? As thinking of starting over blowing up the original again out of fear.
Finest
William

Hi @William London
@JeffK has already covered all the salient bits of information.
My advice is don't throw anything away and start over. A very good chance you would never notice the difference especially large for wall art.

I will add a couple more points on top of what Jeff has already stated.

JPEG is 8 bit depth only and TIFFs and PNGs can be 16 bit. This will only matter if the original images were in 16 bit with fine details in shadows or highlights or unless you are doing a high number of more extreme adjustments in your processing.

JPEGs are lossy yet setting 12 is quite good and the compression is mostly done in the color components and not the luminance component of the image. The eye picks up loss of luminance detail and much less so in color detail. Where it can come in to play as an issue is if you are using processing the changes or rotates color which could bring out the compression artifacts from the color. The amount again would not be high with compression setting 12.

JeffK already covered of repeated resizings and some loss of detail when that is done and the use of Smart Objects to help minimize the issue. Smart Objects do experience a loss of detail with resizing for the first resize that is done. With additional resizings, the Smart Ojbect always recalcualtes the image from the original so there is not loss from repeated resizing except for the first one.
Again, this is not a large loss unless you have done it tons of times or do massive enlargements following by massive reductions. This applies for transformations of other types as well.

All of these losses are real yet at a small scale. That is where wall art helps a lot. The eye can barely discern 300 ppi at a 10 to 12 inch distance. Wall art is typically viewed at much further differences. If you typical viewing distance was 6 feet, then you eye could only discern ~50 ppi on your wall print. Most of the issues of which you have concern would be small compared to what you could view at that typical distance.

Thought I doubt is needed, you could always switch now from where you are at and save in a non-lossy format and start using Smart Objects yet I bet it would be overkill for your project

Just my opinion of course and others may differ.
John Wheeler