The post in question: post #678607

From tumblr: http://41.media.tumblr.com/a0622235949250142e53013ecbe7279c/tumblr_npn2csSrQ91tl8xjio1_1280.png
From Twitter: https://pbs.twimg.com/media/Cb_y2EGW4AAZzrv.jpg:orig

The Twitter picture is larger, but it has compression artifacts if you look close. So which is the better version? Can any artists or image experts chime in?

We usually accept the one with the higher file size.

The tumblr one: 1280x1231 (1.10 MB)
The twitter one: 2600x2500 (711 KB)

We'd accept the tumblr post

Okay thanks. And I solved the mystery, the Twitter version is a jpg of the original DA version! http://rainbowscreen.deviantart.com/art/Not-for-you-Penelope-537697775

TheHuskyK9 said:
We usually accept the one with the higher file size.

The tumblr one: 1280x1231 (1.10 MB)
The twitter one: 2600x2500 (711 KB)

We'd accept the tumblr post

Is that really the way we do it? I'd imagine that we've so much more image quality is lost from a 76% downscale than we will with a little bit of JPEG compression.

We've got the deviantart version in this case, but I'd have a strong personal preference for the Twitter image if it wasn't available. It shows little visible compression.

In this case, it seems like the larger resolution one has demonstrably more visual data preserved, despite it being a more compressed file size. If you blow them up to equal size side by side, the larger res one retains many sharper details on lines that simple become blurry in the tumblr version... or is this not a proper way to compare for some reason?

Feels like the jpeg twitter upload was compressed from a much larger version of the starting file than the tumblr post was downscaled from, and in this case it seems like the twitter version would be the "right" one to keep.

Edit: And if I was less tired I would have noticed that Leo said he already found the original post anyway, which makes my whole reply useless. Neat.