The Red Compression Issue

Red is one of the colours I use most often in my graphic design for marketing. However, despite it being a good colour to use in marketing advertisements, it also suffers the most when compressed via social media platforms. Yes. They all suffer but red more than others. Why?

The red issue was such a problem for one client, that they ended up changing the colour of their logo so that it would do better on social media, their number one place for generating business. So I do think it's an issue that could do with some attention.

These examples below are maybe not the best example, but you can still clearly see that the red version of the same image looks more distorted than the blue variant.

I have tried different variations of red to try and find a solution and found some work better than others, but they still never look as good as non- red spectrum images.

Red vs Blue

These images are identical. The only difference is one is red and one is blue. They both don't look as good as each other. The red suffers more distortion than the blue version. This is the same when compared to other colours, not just blue. Red always loses.

enter image description hereenter image description here

The Question

Is there a reason for this and is there an official red spectrum sweet-spot that vastly reduces the compression artefacts to make it on par with other colours with the same compression?

Note: This question is not about compression as such. It's about why red suffers more than other colours and how to solve this.

Twitter Compressed Images

These are the images that have been compressed by Twitter in this case. It's similar for other social media platforms. All struggle with red colours.

Note: These aren't the exact images use by Twitter, they are a screenshots, but they look almost identical.

Please be friendly and helpful if you can :)

enter image description hereenter image description here

  • "looks more distorted" in what way? The red already looks less sharp than the blue even in the original. The colours are 'within one' on RGB. The banding in the background is present in all 4 examples. – Tetsujin Sep 28 '18 at 7:17
  • @InvariantChange: "It's about why RED suffers more than other colours. The question is not about why images look terrible on social media." Please edit your question then and remove the references to social media, as they only serve to confuse the question. – PieBie Sep 28 '18 at 7:38
  • 1
    Red is a pompous color and deserves to be put in its place. That's why. – Joonas Sep 28 '18 at 9:04
  • 1
    I dont really understand why this question was downvoted? So im upvoting it. – joojaa Sep 28 '18 at 11:15

It's an Optical Illusion

I sent this question to a friend of mine (who is a biologist) and he gave me this answer.

To answer this question, he said you have to look at the science behind it, which is called Tetrachromacy.

There is actually no difference in the image compression of red as there is from other colours. It is actually an illusion because of how the human eye perceives colour information via our colour sensitive cones.

By population, about 64% of the cones in the human eye are red-sensitive. This means that we see red variations (artefacts) more clearly than other colours variation. Second to red is green at 32% and 2% of cones are blue sensitive.

So in my example above, the reason the red and blue look so different is because there is a big gap between the number of red cones in the human eye used to receive this colour information than blue cones. See image.

enter image description here

I have been looking for an answer to this for a few years. Isn't it funny how by simply asking the question, it can lead you on a path to better answer it yourself!

It would be interesting to know if there are variations of how we each see these compression artefacts based on demographics or even variation in epigenetics. And if this plays a roll in this famous but unrelated illusion.

enter image description here

I hope this information is useful to other people. If it's not. Please flag it, so it can be removed.

| improve this answer | |
  • 1
    Usually the person making the compression algorithm knows this and has accounted for it. So whoever made the compressor most likely used a engine that compresses blue more than red. So for example your JPEG compression in Photoshop already does this – joojaa Sep 28 '18 at 10:55
  • 1
    Well it is a result of RGB -> YCrCb conversion. So while the jpeg compressor has some leevay into implementing the actual compression what happens is that Y gets compressed least and Cb gets compressed a bit more tha Cr, but not necceserily enough to compensate for how your eye works. – joojaa Sep 28 '18 at 11:02
  • 1
    Thanks for rewording. I was wondering if this was taken into account. Maybe it is to some degree but there must be limits to what they can compensate for. I suspect this is a very complex subject that requires a lot of peer-reviewed studies to read through! Thanks for your input and time - Very much appreciated :) – Invariant Change Sep 28 '18 at 11:23
  • 1
    Actually im now looking at the images in yCrCb space and i get the feeling its just simply because the red image has more entropy in yCrCb due to not being pure red while the ble is much closer to pure. So this may be just a case of bad choice of red combined with your vision system – joojaa Sep 28 '18 at 11:26
  • So it has nothing to do with social media? – PieBie Sep 28 '18 at 14:05

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