Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

It's time to replace GIFs with AV1 video

singhkays.com

Post reply on HN
Print view View original

251–260 of 327 posts

Re: It's time to replace GIFs with AV1 video

#251

Post by harshreality »

echelon wrote: post #183
paavoova wrote: post #175

Earlier quoted context omitted.

> You can simply save the file and expect it to work everywhere I don't quite follow. This is because the gif is decoded and played. No different than a video. You don't need a proprietary player to loop a video, you just go back to the start of the video. For streaming, this is only problematic for large videos that can't be cached, but the same applies to large gifs. Browsers can loop video, it's just a right-click…

Video isn't superior for sharable, loopable images. Try downloading a video from a popular social network. Can you easily do it without inspecting the source? If it's a two second clip, does it loop on your system? Or does the video player exit / end the stream? This is absolutely a problem that GIF doesn't have. > You don't need a proprietary player to loop a video > You can save any video file just like a gif. Exce…

Your complaints about true video formats (vp8/vp9/x264/x265/av1 with html5 video) replacing true gifs are misplaced.

Inability to save videos/gifs isn't a format problem, it's a problem due to javascript obfuscation or DASH/HLS making it difficult because you have to find the video chunk url(s), fetch them, and if there are multiple pieces, piece the full video back together. Campaigning for a return to gif isn't going to make those sites switch back. The only sites that might switch back would be sites that already allow right-click download. They won't switch back either, though, because gifs are vastly less efficient.

Looping is a html5 video tag attribute. If a video isn't looping, it's because the site serving the video didn't add that tag attribute. Many gif-style video upload sites automatically loop videos.

Also, gifs in browsers don't support seeking which is annoying for gifs more than a few seconds long.

HTML5 video is perfectly fine, loops fine, and is easy to save if the site doesn't obfuscate it with javascript or turn it into a chunked HLS or DASH mess.

Re: It's time to replace GIFs with AV1 video

#252

Post by iainmerrick »

Dylan16807 wrote: post #209
pjc50 wrote: post #119

Earlier quoted context omitted.

So is animated gif, though.

What I'm saying is, in uses akin to this article, apng is not a meaningful upgrade over gif. It's a huge upgrade for certain kinds of content, especially animated icons with aliased edges. But that narrowness means there's only a moderately small push toward adoption.

In principle it has the same advantages PNG has over GIF for static images -- 24-bit color and 8-bit alpha, both of which are a big deal.

APNG or something similar should have replaced GIF completely, but the delivery was badly fumbled. Browsers now mostly support it, but authoring tools mostly don’t.

Re: It's time to replace GIFs with AV1 video

#253

Post by 6gvONxR4sf7o »

yoavm wrote: post #77
izzydata wrote: post #33

The most important part of GIFs for me is that they behave like images in browsers. They are always auto-playing with no concept of play and pause. You can drag and drop them from a browser to your desktop to save them. You can save an entire page and have all the image files save with it. I've never had this work for webm or other video formats. You could even argue that GIFs not being a video with no video decoding…

> It may take longer to load, but I can have 100+ GIFs playing at once with no impact to my CPU. I don't know what you're talking about really... I was just using Google Slides today and both Firefox and Chrome went to 250% CPU usage, only because there was one slide with a 5 seconds full screen recording GIF. I begged to the person who added it to turn that to a video.

In my experience that's a google slides problem, not a gif problem.

Re: It's time to replace GIFs with AV1 video

#254

Post by soulofmischief »

ryanwhitney wrote: post #236
twhb wrote: post #203

Earlier quoted context omitted.

The UX for saving and sharing videos is the same as for images, including right click to save. You and I don't get that UX, though, because Twitter, Google, Facebook, Netflix, and most websites write code to fight off-site sharing and deploy it on their videos but not their images. They skip images presumably because it's easier to bypass there and because everybody knows you're user-hostile when you break image UX,…

> The UX for saving and sharing videos is the same as for images… Minus drag and drop, in browsers at least.

This comment reminded me of Terry Davis, because that's exactly the kind of feature he would have deemed essential in a browser if he'd ever decided to add networking to TempleOS. RIP brother.

Videos really should be first-class entities, though.

Imagine an OS/browser-level ffmpeg implementation which would allow seamless copy/paste of videos, integrated snipping tools both for cropping and time, color adjustment, etc. with automatic imgur-style hosting of your newly edited video just a button-click away.

Re: It's time to replace GIFs with AV1 video

#255

Post by kkarakk »

harshreality wrote: post #251
echelon wrote: post #183

Earlier quoted context omitted.

Video isn't superior for sharable, loopable images. Try downloading a video from a popular social network. Can you easily do it without inspecting the source? If it's a two second clip, does it loop on your system? Or does the video player exit / end the stream? This is absolutely a problem that GIF doesn't have. > You don't need a proprietary player to loop a video > You can save any video file just like a gif. Exce…

Your complaints about true video formats (vp8/vp9/x264/x265/av1 with html5 video) replacing true gifs are misplaced. Inability to save videos/gifs isn't a format problem, it's a problem due to javascript obfuscation or DASH/HLS making it difficult because you have to find the video chunk url(s), fetch them, and if there are multiple pieces, piece the full video back together. Campaigning for a return to gif isn't goi…

so you're saying the solution to use as a replacement is harder to use and clunkier? right, i'm sure everyone will rush to adopt it and customers won't care at all

Re: It's time to replace GIFs with AV1 video

#256

Post by kkarakk »

yoavm wrote: post #77
izzydata wrote: post #33

The most important part of GIFs for me is that they behave like images in browsers. They are always auto-playing with no concept of play and pause. You can drag and drop them from a browser to your desktop to save them. You can save an entire page and have all the image files save with it. I've never had this work for webm or other video formats. You could even argue that GIFs not being a video with no video decoding…

> It may take longer to load, but I can have 100+ GIFs playing at once with no impact to my CPU. I don't know what you're talking about really... I was just using Google Slides today and both Firefox and Chrome went to 250% CPU usage, only because there was one slide with a 5 seconds full screen recording GIF. I begged to the person who added it to turn that to a video.

Google problem not gif problem, sites like knowyourmeme embed 10 gifs in a page routinely with no issues(apart from the loading time)

Re: It's time to replace GIFs with AV1 video

#257

Post by xster »

Nobody mentioned so I'll add: the one thing holding me back is the current ability to insert autoplaying gifs into GitHub's markdown.

No type formats are supported and being able to add an animated gif showing what you're creating a pull request for is immensely useful.

Re: It's time to replace GIFs with AV1 video

#258

Post by Kaiyou »

Ezhik wrote: post #126

At this point, killing GIF is a UX problem, not a format problem. I've talked about this before [0], but the big problem for me is that video is just difficult as hell. Compared to a GIF, it is just that much harder to save a video on a phone, and then upload it the same way as an image. Try to save a "GIF" from Twitter or from GIPHY - it's a /huge/ pain. Whatever the GIF killer is will need to pass the right click t…

There's already apng (animated png) which is like GIF, but with the quality of PNG. No one really seems to care, though.

Re: It's time to replace GIFs with AV1 video

#259

Post by Kaiyou »

iainmerrick wrote: post #171
afiori wrote: post #153

Earlier quoted context omitted.

wasn't webp supposed to have GIF like animations? As far as i know I have never seen one in the wild, just webm. I think the problem is that GIFs are images and so displayed as such with little concept of play-control. A video format communicate that you might want to pause for example. I believe a GIF-killer will need to be considered an image as a format.

Yes, agreed. That lack of play-control is both a weakness and a strength of GIF. The strength is misunderstood and underrated by those who think videos should replace all usage of GIFs.

Play control isn't impossible for GIF. I'm using sxiv for viewing images, which is as minimal as it gets, but it does have pause, single step forward and single step backward for GIF.

Re: It's time to replace GIFs with AV1 video

#260

Post by niftich »

niftich wrote: post #221
clouddrover wrote: post #218

Earlier quoted context omitted.

Which patents is HEIF itself encumbered by and who runs a patent licensing pool for HEIF?

MPEG tells [1] rightsholders to file Patent Statement and Licensing Declarations to ISO/IEC. ISO/IEC maintains [6] a spreadsheet of received documents and the relevant standards that they concern. Examining that spreadsheet, I found four declarations that concerned the HEIF standard, all four submitted by Nokia Technologies Oy. These documents [2][3][4][5] provide references to patents approved or pending. The patent…

The US patent applications in the above list resolve to the following five:

https://patents.google.com/patent/US20160234144A1/ This one seems to be about specifying a Mimetype paramater that tries to describe the cost of image transformations requested in the target file, so that clients who know they can't perform those transformations can pick a different file.

https://patents.google.com/patent/US20160232939A1/ Seems to be about image sequences and specifying such a concept with coherent metadata in the container.

https://patents.google.com/patent/US20150193494A1/ This seems to talk about the myriad ways that ISOBMFF can assert metadata about data elements within, but there aren't always good ways to ensure the metadata points back to a specific data element. This talks about ways of figuring out whether such loosely-floating metadata in the container is still valid for data items; they also propose using checksums to figure out if the data items changed.

https://patents.google.com/patent/US20180146225A1/ This is a fancy restatement of P-frames for still images, leaving open the possibility that later frames also 'enhance' the first image in various ways e.g. upscale resolution and others.

https://patents.google.com/patent/US20140314148A1/ This defines signalling to allow putting all the I-frames at the start, and all the P-frames at the end.

Post reply on HN