Slashdot Mirror
Netflix To Re-Encode Entire 1 Petabyte Video Catalogue In 2016 To Save Bandwidth (variety.com)
An anonymous reader writes: Netflix has spent four years developing a new and more efficient video-encoding process that can shave off 20% in terms of space and bandwidth without reducing the quality of streamed video. With streaming video accounting for 70% of broadband use, the saving is much-needed, although the advent of 4K streaming, higher frame rates and HDR are likely to account for it all soon after. Netflix video algorithms manager Anne Aaron explained to Variety that certain types of video benefit little from the one-size-fits-all compression approach that Netflix has been using until now: "You shouldn't allocate the same amount of bits for My Little Pony as for The Avengers."
Shaving 20% off seems pretty optimistic to me. Unless they've suddenly discovered some whole new realm of compression mathematics I'd be surprised if thats anything more than a peak compression in some rare edge cases.
I watch YouTube a lot, on average about 2-3 hours a day. As of late, I live in a country where there is a bandwidth cap of 40 GB/month. And I have no option but to YouTube at 144p to avoid extra bandwidth charges.
I applaud all efforts by tech companies to reduce bandwidth usage (and not to forget, making inter-webs more exciting). Then again, none of those efforts matter, if bandwidth caps are forcing consumers to use internet like back in 90s.
Actually I doubt they'd reduce the color precision from 24 bit (it's not 32 bit).
In fact, content with lots of synthetic content can sometimes be smaller by being 30 bit instead of 24 bit (think how often synthetic content puts in gradients, with 24 bit those gradients are more dithered than 30 bit, and the compression algorithms struggle a bit more with what appears to be 'noisy' content from dithering compared to less noisy undithered content).
Of course this is using general purpose algorithms that are used for both animated and photorealistic alike. There may be some gains in theory from a codec focused exclusively on animated content, but in practice no one seems to think it's worth the trouble to pursue.
XML is like violence. If it doesn't solve the problem, use more.
Netflix probably aren't too keen on the idea of paying people to puzzle over what compression would best suit each and every item in their 1-Petabyte video library.
The summary says they spent four years developing the new approach. I suspect that paying people to puzzle over (in layman's terms: do research) how to improve the encoding across their Petabyte video library was exactly what they did.
On a serious note, animated content is much harder for the 8-bit encoding. It's the hard edges with high contrast cell shading. You get a lot more compression artifacts than a typical movie. You can resolve this by using 10-bit encoding, but there's a lot of Netflix devices with embedding video codecs. They really can't change, and almost none of the chipsets out there support 10-bit decoding. So that leaves option two, which is to increase the bitrate.
I recode a lot of video (don't ask), and in my experience, animation is a lot easier on the resulting files for the same quality. The absolute worst you can do if you want small high quality files is "film grain", whether it is from a bad source or artificially added for artistic reasons. Second worst is a badly compressed source with lots of artifacts. Then there's video with lots of small objects moving across a detailed background. The hard contrasts at the edges of cell shaded videos are only problematic if you don't have a good quality source, i.e. if your source already has lots of compression artifacts.