Let's say I have a 1 mb JPG image file that I need to make a minor edit to and re-save.

For the editing, I can use photshop or gimp.

For compression, I can use photoshop, gimp, and any of those online tools such as tinyjpg, jpegmini, etc.

I'm wondering what is the best process to save and compress a JPG edited in photoshop/gimp for the web to preserve the most amount of quality while lowering the filesize to a reasonable size (which in my opinion would probably be around 200kb or lower for a 1mb jpg image)?

For example, am I better off saving/compressing the JPG image in Gimp/photoshop at the highest quality number that doesn't cause an actual increase over the original jpg filesize -- which would be about 95+ in GIMP and 100 in Photoshop) and then running the image afterwards through an presumably more optimized/lossless compression tool like TinyJPG, or is it better to simply reduce the image to a quality number that creates the desired filesize I want in Photoshop/Gimp?

In other words, is it more efficient to save a JPG image in Gimp/Photoshop to your desired filesize or to save the JPG file at a maximum quality setting in GIMP/Photoshop and to then run it through an online tool that is presumably better at compression-optimzation than Gimp/Photoshop?

The latter seems like it would be the better option, but I know every time you tamper with a JPG you get degradation, so that extra step of using an online tool might actually be the worse option.

  • 4
    "Best" is rather relative. Each image may or may not benefit from any one of the methods you cite. There's no "universal" tool since all images are unique. – Scott Sep 15 '20 at 7:20
  • have you checked this Q? graphicdesign.stackexchange.com/questions/31638/… – Luciano Sep 15 '20 at 11:31
  • "to save the JPG file at a maximum quality setting" if you give the image to another JPEG encoder, at that point you can save it in a lossless format (PNG or TIFF) so you avoid the losses caused by the double encoding. – xenoid Sep 16 '20 at 10:16

One thing to try is Google's guetzli. Basically it tries the various JPEG encoding parameters for you and checks the resulting images.

The other side of the coin is that it is a memory and CPU hog. But if you have many similar images, you can try a few and if the parameters are similar you can reuse them directly on the whole set using a simpler encoder.


You have a few contradictory requirements.

If you need to preserve the information, make a backup, and use a lossless compression method. PNG, TIF, PSD, or XCF.

But if you need to compress it to be published use a 4:4:4 subsampling. On Gimp it is under the "Advanced" options when exporting to JPG.

Do not smooth it and click the "Optimize" checkbox and play with the compression level.

enter image description here

The problem is that the relationship between "quality" and file size is subjective.

One option is that if you see it ok, then it is ok.

But if you need to process digitally the file, or the file is meant to be processed then you need to understand what is this compression doing.

I made some tests a few years back (I really need to update that page) with a method to compare different settings on different programs and see the difference in the information of a compressed file vs the uncompressed version.

It is in Spanish but you could use google translate.

It probably will give you a hint on what to expect when compressing a file.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.