- EBXfwYH.png
- - 720.15 KB
- (708x911)
JPG optimization is magical
I hope you still make the original version available, otherwise it's following the bad example of mainstream websites where everything is compressed and recompressed over and over again.
But anyways, no plan to alter original images aside from exif stripping on .jpg.
But that does mean that in a hypothetical future of limited disk space, filesizes of uploads could end up being minimized.
Out of many possible resolutions to this issue, it's up to what's perceived to be the lesser evil at that point.
>>8218
https://ezgif.com/
35MB limit
If you have the effort to learn and just want to optimize then you can probably use ImageMagick instead
>>8210
It is lossy. The WebP definitely looks softer, but it's a fool's errand to compare an already lossy image to a lossy image of the original which is also a lossy image. A fair comparison would only be if there was an original lossless image and lossy versions made using both JPG and WebP or whatever is trying to be compared.
>>8220
The artist uploaded it in jpg https://www.pixiv.net/en/artworks/83061315
probably because a 120MB image is a bit excessive.
>>8222
well if the artist thought that a JPG was the right file format for their art, then it's not fair to discard a scenario because it doesn't suit you. Compression can be done before you notice any differences. Saying anything else is falling into the typical case of people attempting to differentiate between MP3 and WAV.
>>8223
Huh? All I said was that it's not valid to evaluate the compression of a lossy format against a lossy format if the source image is also lossy...
If the source image was a lossless image, it would be a fair comparison. Pointing out that there was image degradation between the JPG and WebP is moot because there is nothing to compare the JPG to, especially since it is the source image.
It'd be like saying it's valid to evaluate an OPUS file's compression from re-encoding a 128kbps MP3. Re-encoding artifacts are very much a real thing, so it's impossible to accurately assess the strengths of WebP in this circumstance.