4 ms·

I get your point, just to play devil's advocate though: up until a few years ago, there was an often repeated argument that "the human eye can't see more than 3

I get your point, just to play devil's advocate though: up until a few years ago, there was an often repeated argument that "the human eye can't see more than 30fps" (it was used to justify the fact that many console games were capped at 30fps back then)... Today, many gamers use a 144hz monitor and would never go back to a lower refresh rate. I wonder if the same story will happen with regards to lossy video compression, I mean perhaps, once our tech will be so much better that we will be able to use lossless formats for video, people will look back at the formats we have today and think that they look bad. Remember VHS? I used to think they looked alright back then :-) I think that the lower adoption of Blu Ray vs. streaming is purely due to convenience, not quality. Most people don't want to deal with physical media. In the same way that uncompressed audio is better than Spotify and similar services, yet almost no one buys audio CDs any more. In Audio, there are now services that let you download uncompressed and/or lossless audio (Bandcamp, Beatport etc.), they are not as mainstream as Spotify but they are moderately successful and have their place in the market. I wonder if the same thing could happen for movies: offering higher quality downloadable movies, without the physical medium. Perhaps cinephiles could be interested. Maybe it will happen when average bandwidth, at least in the developed world, approaches the equivalent of a 4K Blu Ray bitrate: no one wants to wait for a long download to end, people want just-in-time streaming. (Note: Blu Ray, unlike Audio CD, is not lossless, it's just that it's got much less compression than the typical streaming services).

Youden 5y ago
I don't think the claim that the human eye can't see more than 30fps was ever supported by evidence, it was just "common knowledge" like other falsehoods such as certain parts of the tongue only detecting certain tastes. I think it was actually based on the minimum frame rate necessary for a human to perceive motion instead of a slide show. Scientifically, a quick Google indicates that humans can see up to around 90Hz for more normal cases and 500Hz for a specific edge-case [0]. When it comes to video compression though, industry practice is largely based on the result of user testing. There are well-controlled tests where a bunch of people are put in a specific test environment and shown videos encoded using different parameters and there are larger scale A/B tests where a streaming platform will tweak parameters and see how they affect metrics like time spent consuming content. So I'm confident that the compression of today matches the users and hardware of today. However hardware can change and I expect that's what will drive increases in video quality. For example the encodes of today are biases to some extent by the fact that people watch videos on tiny screens like phones or tablets, which are likely a bottleneck for quality perception. If people start consuming more content on say VR headsets, streaming platforms may see the need to increase quality. 8K is also a possible driver (especially for very large displays). Connectivity is also a big deal. Lossless music is possible because the bitrate is relatively low but streaming 4K blu-ray is out of the question for many people at the moment. In the future though it's definitely a possibility. [0]: https://www.nature.com/articles/srep07861 https://www.nature.com/articles/srep07861

Youden 5y ago

I don't think the claim that the human eye can't see more than 30fps was ever supported by evidence, it was just "common knowledge" like other falsehoods such as certain parts of the tongue only detecting certain tastes. I think it was actually based on the minimum frame rate necessary for a human to perceive motion instead of a slide show. Scientifically, a quick Google indicates that humans can see up to around 90Hz for more normal cases and 500Hz for a specific edge-case [0]. When it comes to video compression though, industry practice is largely based on the result of user testing. There are well-controlled tests where a bunch of people are put in a specific test environment and shown videos encoded using different parameters and there are larger scale A/B tests where a streaming platform will tweak parameters and see how they affect metrics like time spent consuming content. So I'm confident that the compression of today matches the users and hardware of today. However hardware can change and I expect that's what will drive increases in video quality. For example the encodes of today are biases to some extent by the fact that people watch videos on tiny screens like phones or tablets, which are likely a bottleneck for quality perception. If people start consuming more content on say VR headsets, streaming platforms may see the need to increase quality. 8K is also a possible driver (especially for very large displays). Connectivity is also a big deal. Lossless music is possible because the bitrate is relatively low but streaming 4K blu-ray is out of the question for many people at the moment. In the future though it's definitely a possibility. [0]: https://www.nature.com/articles/srep07861 https://www.nature.com/articles/srep07861