Image Quality Test
If you celebrate it, Happy Thanksgiving!
I used my first day off to write something I've been meaning to do for a while. I sometimes see people request uncompressed images in our games, but I'm skeptical that they would actually be able to tell the difference between uncompressed and reasonably compressed images. At the same time, uncompressed images would make the game ten times larger.
So I wrote some code to let you folks test this for yourselves. I've got the images from Rain God in uncompressed PNG, Full Quality JPEG, and 80% Quality JPEG. Pick the two you'd like to compare, and it will randomly choose between them. You get to pick if you think the image you're looking at is the higher quality one or the lower one. If you guess randomly, you're likely to get around 50% right. If you can really see the difference, you should be able to repeatedly do over 85%.
You can download the test here: http://www.mediafire.com/file/d2wo11wvxffuu0e/qualtest.zip/file
Give it a try, and let me know how it goes.
Tlaero
Comments (7)
Sorry for being somewhat late, but today I checked JPEG-80 against PNG and I really can't tell a difference. 3 attempts: 7/16, 6/16 and 10/16. I think Lazy Turtle is very right: if someone can think about compression algorithms when playing Tlaero&Mortze games, he/she needs to check his/her literacy and sexual hormones level. It seems that person is both asexual and can't enjoy a good story. :)
Personally, I wasn't able to tell PNG from JPG80 (highest vs lowest), so from my perspective, stick with the compression.
Did just a quick run last night without paying too much attention, scored way below 50%. It's fair to say Mortze's style makes that differences more difficult to notice (maybe I'm saying a nosense here but also the water element in those pictures can deceive a bit, as you can't really tell if it's a lousy compression or just an intended effect). Anyway, I think people usually complain about that quality loss in those games with plastic-shine 3D dolls as characters and plain backgrounds. The saddest thing is when you see people actually linking filesize with game's quality and amount of available content: "if it's below 3GB it doesn't worth it"
I think you nailed the key element here: Mortze puts SO much effort and QUALITY into his photos - be it the character designs, the environments or THE LIGHTING! - that a little compression isn't fatal in any way. Lesser work would tend to "show the seams" more easily.
welp, I can't seem to tell the difference, so just make what is convenient and I will enjoy the work
Obi-Wan was right: 'Your eyes can deceive you, don't trust them.' I tried it several times, kept my score around 50-55%, 68% at best. My conclusion: While some people might notice differences in quality, most members of general public likely can’t. Or they don’t care. As long as the action on screen is well-made (and that is basically given, considering the quality of Mortze’s work and Tlaero's writing) these little differences become irrelevant. And frankly, if your first thoughts after playing game with 1000s of pictures of naked people having hot, steamy sex are about picture compression, than you should think really hard about your priorities.
Despite all the shenanigans that 2020 hath wrought, I'm still thankful for Tora Productions. Happy Holidays to Tlaero, Mortze and the whole gang here.