6 ms·
Codec Royalties on Content and the Jaws Moment
- Silhouette 3y agoI think governments should simply legislate that patents are unenforceable to the extent that they restrict the free exchange of information. The status quo represents a threat to communication itself and potentially leaves everyone incapable of safely offering information to others. This is "patently" nonsensical and contrary to any moral or economic argument I've ever heard for the existence of IP rights.
- clouddrover 3y ago> Cloud and streaming services are big beneficiaries of codec technologies. Newer, better codecs dramatically reduce their storage requirements and provide ultra-high resolution videos without buffering or latency issues, enhancing the services that they charge users to watch or sell advertisements around. At present, these companies are not contributing to the cost of the codec technologies they rely on. This is, of course, wrong. These companies are contributing to the cost of the codec technologies they rely on by now building those technologies. Through, for example, contributing to the development of AV1: https://aomedia.org/membership/members/ https://aomedia.org/membership/members/ The web (and the internet generally) is premised on open, royalty-free formats and protocols that everyone is free to implement. It's how it works and how it has always worked. There is nothing at all special about audio and video formats. Audio and video formats on the web can and should be royalty-free just like everything else.
- charcircuit 3y ago>There is nothing at all special about audio and video formats Some formats are able to compresses better without losing details. If there is a superior codec, not having open licensing is not a deal breaker.
- kevingadd 3y agoIf the codec lacks hardware decode support and lacks a supremely efficient software/gpu accelerated decoder on end user devices, it's not actually a useful codec even if it's superior. And you won't get widespread availability if the codec is expensive.
- charcircuit 3y agoHardware accelerated H.264 is widespread and not openly licensed
- kevingadd 3y agoIt's still the case that web browsers have to ship software H264 decoders because it's 2023 and you still can't rely on being able to HW decode it on every device Also, h264 predated the wide availability of HW video codecs in general, so software decoding was the thing for quite a while and that provided a motivation for HW decoding. It's an ancient codec, it shouldn't be surprising that eventually everyone got decode support for it. In 2023 if you're proposing people adopt a new codec with sw decode only, why should they eat significantly reduced battery life + higher patent fees when they already have hardware decoders for h264 and vp9? It's a losing argument. I also can't speak to this with confidence but I suspect Apple choosing to use H264 single-handedly advanced support for it. If they were to do the same thing with H265 or AV1 in a strong way we'd probably see the same effect.
- clouddrover 3y ago> not having open licensing is not a deal breaker Which is what everyone says right up until the point it does become a deal breaker, such as with the rumblings about content royalties. The better approach is to stick with the royalty-free licensing principles the web is built on from the start: https://www.w3.org/Consortium/Patent/ https://www.w3.org/Consortium/Patent/ Royalty-free internet and web standards have unlocked massive value (and massive profits) for all participants on the web. Audio and video formats can and should stick to these same principles.
- wheybags 3y agoAs someone who works with low level video processing code fairly regularly - codec patents are demonically awful. I can't even begin to imagine how much better everything would be for everyone if they just died.
- throw0101b 3y ago> As someone who works with low level video processing code fairly regularly - codec patents are demonically awful. Perhaps software patents in general?
- est31 3y agoThe best marketing for AV1 is the horrible mess of HEVC licensing (and the way higher price). It's hard to lower your price when you are free but when your paid alternative increases its price (actually paid price plus the risk) it is even better for you :).
- ninjin 3y agoAlthough not an expert in video processing, I am somewhat surprised that it has taken until AV1 for the conversation on moving "past" H.264 and H.265 to really get started. Sure, VP8 was not great on a technical level, but my own experience was that it was on par with H.264 which arguably is still the industry standard. With VP9 there was no question that it was superior to anything with a proprietary licensing structure. AV1 of course improves upon this (greatly at that), but both VP8 and VP9 have had hardware decode and encode support for at least half a decade at this point and I am not sure why we did not already start having these kinds of discussions a few years ago once VP9 had proven to be solid and received hardware support (Google have served VP9 on YouTube for at least a few years).
- lern_too_spel 3y ago> both VP8 and VP9 have had hardware decode and encode support for at least half a decade at this point It's only recently that either VP9 or AV1 could be efficiently decoded on iPhones (or Macs). YouTube can afford to serve lower quality experiences to these devices, but many services can't.
- ninjin 3y agoThank you, that is probably what I was missing as I do not frequently deal with the Apple ecosystem. On the note of YouTube, my understanding is that they gracefully fall back on H.264 if you lack support for VP9 on your hardware (they probably are or will do the same for AV1). Google is most likely promoting VP9 this way for their own long-term benefit. HLS has also been an issue as the RFC mandates proprietary codecs, but Apple have silently added VP9 support recently in their implementation [1]. The HLS container formats are still proprietary though from what I can understand. MPEG-DASH should support VP9 and AV1, but I have not read up on it much as MPEG-DASH is proprietary as a protocol [2]. [1]: https://github.com/video-dev/hls.js/issues/5684 https://github.com/video-dev/hls.js/issues/5684 [2]: https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_over_HTTP#Patent_holders https://en.wikipedia.org/wiki/Dynamic_Adaptive_Streaming_ove...