Export quality and subsampling

Software RawTherapee
1

RT 5.8 on Manjaro. I’m doing a small gallery (no printing) that will be viewed by 2/3 of the people on the phone and the other third by people on their computers.

I only have 1080p monitors. My export settings are jpeg, 65% quality and best compression.

Will people on 4k monitors see the difference between 65% quality and 85% quality with a balanced compression?

Thanks

2

65% is too low, I’d do at least 80%. People on 4k monitors will likely see the photo at its 100% pixel size, unless you’re explicitly up sampling in the browser, which I would discourage.

3

I don’t control the browsers.

I’m trying to find a balance between smartphones and desktops…

Thanks

4

One way to find out! Show me two jpegs without metadata or names that could give away which is which and I’ll tell you if I can see the difference

1 Like
5

But you do, you can use CSS, JavaScript, or width/height tags to control the size in the browser.

1 Like
6

If those 4K displays are half decent, people will see (and tell) the difference between 65% and 85%.

I’m exporting a lot of jpegs lately, and sometimes 80% is not enough. I would even say that with certain images 85% is barely enough: if you have smooth gradients between pastel tones, anything lower than 85% and you will easily see artifacts in those gradients.

If size is not a huge concern, I would go for at least 90% quality. In the end it’s a gallery, so quality should be the main concern, shouldn’t it?

7

Thank you for the offer. I will do that tomorrow.

8

Actually, that’s a question that I’m not sure about. The majority are on smartphones. I’m trying to balance speed vs quality. It’s always: quality, speed, cost: pick any 2…

I will think about it and make some tests.

Thanks

1 Like
9

I’m not using the image upload because I do not know what pixel.us does with the image: compression, resize…

Here it is, I have 4 images with different settings, v1/v2/v3/v4:

https://drive.google.com/file/d/16myF79HxP3Pb6Zzf5Jm0g0nFJpD3vlPS/view?usp=sharing

https://drive.google.com/file/d/19jjNBXxxsQBJuNSC3iTOW6vXl4Eq-woW/view?usp=sharing

https://drive.google.com/file/d/1VJfKp9lRSuOQknBkdJloPADIDA-3CL8i/view?usp=sharing

https://drive.google.com/file/d/1sS6ERuZdSJtfBBVN7TazzJUTFeDCIrud/view?usp=sharing

Thank you for taking the time

10

they are distinguishable from one another due to slightly different artefacting on some details when you flick between them, but in terms of absolute quality there’s no discernible difference in perception on my end @ 100%.

11

@foto I completely agree with Dave here: at a comfortable viewing distance from my 4K monitor (~40 cm) and without enlarging the images there is virtually no difference between the images.
Only when looking up close, or when stretched to fit my display, I can spot some encoding differences (in particular some busier color in the darker areas), but even then it’s extremely minor and I wouldn’t dare placing a bet on telling you which is which in terms of export quality.

Cool, didn’t expect that!

12

I agree with Dave and Roel and I would say that from a casual observer point of view you probably wouldn’t have to worry about losing quality with your settings.

Even then, I clearly see artifacts around the whiskers over the green background in all images but image v4 (no wonder it’s the biggest size image). And if you enlarge the image and go above 100%, then you will find lots of artifacts, but that’s not the point here, I guess.

13
  • 1st (v1) was 65% quality and balanced compression 206k
  • 2nd (v2) was 75% quality and best compression 242k
  • 3rd (v3) was 85% quality and balanced compression 342k
  • 4th (v4) was 95% quality and best compression 518k

Thank you for helping

14

Perhaps ensure that your hosting is set up such that smartphone users get sent lower-resolution images by detecting their browser?