I tried with Windows 7 (Firefox 115) and it reports Windows 7.
It seems though that it cannot distinguish between Windows 10 and Windows 11, so, without looking further, I suppose the detection is based on the User-Agent string? (The OS version browsers report on Windows is frozen, so Windows 10 and Windows 11 have the same version there.)
I was working on a carousel library a few months ago. I had made a few stress-test demos so that I could catch obvious issues while I was adding things and tweaking things.
One carousel there had 16K slides.
On Windows both Chrome and Firefox managed that fine. They scrolled from start to end and back without issue and you could see, I think, all the frames in my 60Hz screen.
On GNOME and X11 (dual boot, so same hardware) Chrome was fine but there were issues with Firefox. I was curious so I logged out and logged in with Wayland. On Wayland Firefox was fine too, indistinguishable from Chrome.
I don’t understand hardware, compositors, etc., so I have no idea why that was, but it was interesting to see.
Firefox remains very conservative on enabling modern features on X11. Some distributions force them on, but otherwise it's up to the user to figure out how to do that.
It's likely that some hwaccel flag in about:config wasn't turned on by default. Similarly, if you want smooth touchpad scrolling, you need to set MOZ_USE_XINPUT2
My main Firefox in that setup is from the Mozilla repos, rather than the ESR version that is the default in Debian stable. So, it could very well be that. I will have to check to see what the ESR Firefox from the Debian repos does.
> Firefox remains very conservative on enabling modern features on X11.
So old school throthling if you don't use the "right" version (Apple batterygate, Microsoft wordperfectgate). They could blame it on testing though (we only use Wayland and we are too lazy to test the X11 version)
to prove I'm wrong. It's true that the AVIF image is small doesn't have the obvious blocking artifacts that JPEG and WebP but if you look really close at the reflections on the upper wing of the car it looks like the AVIF just made up some probable-looking blobs of light that don't look that much like the original if you look really closely.
The thing is a video codec doesn't have to be good for images. For instance a single frame of a VHS video looks atrocious but an actual video on VHS isn't that bad.
When I tried to use AVIF to make files of the quality I wanted, I didn't see a clear benefit over WebP and to the contrary I came to the conclusion that WebP was a good drop-in replacement for JPEF for my application. If I wanted to make a big splash image for my web site that didn't have to hold up to close inspection though, AVIF's compression ratio is really high.
My interest in doing the benchmarks was the other thing:
Seeing what the options are these days for the types of image PNG was designed for.
As the results started accumulating, I wasn’t sure I should include all formats in the post and in the TXT file and the spreadsheet, because testing them at what they were not designed for did not seem fair.
Do you think I should add something stronger or more prominent to my intro to explain this?
Here's my take as a web developer from that article, who primarily cares about formats widely supported by web browsers.
For most purposes where I might use a PNG I might use a lossless WebP now because it seems like lossless WebP beats PNG pretty solidly. My take also is that WebP is a good JPEG replacement.
JPEG XL usually does better but practically that doesn't matter much because I think the only web browser that supports it is Safari
Of course there is a lot of politics around JPEG XL, specifically Google doesn't want us to have it and they're a monopolist so we can't have it. If there is any chance we're going to change that we're going to have document that JPEG XL really is better than the alternatives and your article does that.
I think the universities themselves are a problem, regardless of the societal aspect.
How I think about it:
Let us say, as a though experiment, that two things are true.
1. Outside of the university orgs, absolutely no one cares if the universities have enough resources to teach the engineers, scientists, teachers, doctors, etc. that society needs.
2. Outside of the university orgs, absolutely no one cares if half the people equipped to be good university teachers cannot find work.
In that hypothetical scenario, would the universities change? Saying no the Chinese etc. means that they would have to downsize.
I did some reading recently, for a benchmark I was setting up, to try and understand what the situation is. It seems things have started changing in the last year or so.
No, the situation about image compression has not changed. The Grand Poster you were replying to was writing about typical web usage, that is "medium-to-heavily compressed images", while your benchmark is about lossless compression.
BTW, I don't see how Mozilla's interest in a jpegxl _decoder_ (your first link) has anything to do with the performance of jpegxl encoders compared to avif's encoders. In case you're really interested in the former, Firefox now has more than intentions, but it's still not at production level: https://bugzilla.mozilla.org/show_bug.cgi?id=1986393
No. demetris’ benchmark of lossless image compression is not a sign that the situation may be changing. :-D
That was just the context for some reading I did to understand where we are now.
> BTW, I don't see how Mozilla's interest in a jpegxl _decoder_ (your first link) has anything to do with the performance of jpegxl encoders compared to avif's encoders. In case you're really interested in the former, Firefox now has more than intentions, but it's still not at production level: https://bugzilla.mozilla.org/show_bug.cgi?id=1986393
That is one of the links I shared in my comment (along with the bug title in parenthesis). :-)
The Rust rewrites can become tiresome, they have become a meme at this point, but there are really good tools there too.
An example from my personal experience: I used to think that oxipng was just a faster optipng. I took a closer look recently and saw that it is more than that.
If a new tool has actual performance or feature advantages, then that's the answer to "what problem does it solve", regardless of what language it's in.
I tried with Windows 7 (Firefox 115) and it reports Windows 7.
It seems though that it cannot distinguish between Windows 10 and Windows 11, so, without looking further, I suppose the detection is based on the User-Agent string? (The OS version browsers report on Windows is frozen, so Windows 10 and Windows 11 have the same version there.)