Rss

Oh Look, an H.265 Use Case

Streaming Media has a nice article from a few weeks back on how HEVC Innovation Has Been Fast, But Evolution Will Be Slow, that while it’s great that H.265 can cut bandwidth use by about 40%, it will take a long time to get adoption, since that’ll require encoders, end-user hardware, and so on. It doesn’t work until every step in the chain can use H.265. One frequently-cited application of H.265 is in delivery of 4K video, which will have hideous bandwidth requirements, but that deserves ample skepticism, and risks tying H.265 to the 4K rock… if people realize they don’t need a resolution difference they can’t actually see on home TVs, what’s the point of 4K?

That said, I did realize the other day one scenario where H.265 could really make a difference right now: livestream uploading. When I livestream from home, I have only 1.5 Mbps up (U-verse is the best I can do; the cable modem bandwidth on our cul-de-sac is really sketchy), so I run around 1 Mbps to account for overhead, and end up with a sub-SD picture. When I livestream from public wifi at a conference, I’m sharing the bandwidth with everyone and I’m lucky to get a stream out at all (this is why I’m crazy skeptical of wifi-only livestream cams — when can you ever count on having adequate wifi bandwidth in public?). So the prospect of nearly halving the bandwidth requirements has an immediate payoff: with my 1 Mbps ceiling, H.265 would effectively double my picture quality, so I’d be easily capable of doing 480 SD, and could get into HD the next time U-verse bandwidth improves.

Another advantage of this scenario is that it doesn’t require nearly the rollout of equipment that a mass public move to H.265 would. Seemingly, I’d just need Wirecast to be able to encode in 265 and send that up to the streaming server, and then Wowza to then down-transcode that to H.264 (presumably at a higher bitrate) and package it as HLS / MPEG-DASH / FLV for existing mobile, set-top, and PC clients. No, it’s not trivial, but this scenario only requires two parties to adopt 265 (and Wowza has already demoed 265 streaming support), and suits the asymmetric nature of our broadband connections pretty well: a lot of us have way less upstream bandwidth than down (I’m at 18 down / 1.5 up), so any place we’re maxing out the bandwidth is a place where a new codec or compression technology can be put to good use.

No, I don’t think this is going to happen, at least not before a wider rollout of H.265 in viewer devices, but it’s fun to dream.

Speaking of which… do you suppose the next Apple TV might be waiting on H.265? Apple doesn’t usually lead adoption of new technologies, but this is one that might do them a lot of good: Apple TVs and iOS devices could boast better picture quality at the same bandwidth, and for all we know, Apple might be able to pay their development costs just with saved bandwidth costs from their data centers.

It’s also kind of interesting to look back 10 years at Apple’s role in the H.264 rollout, when Microsoft was still pushing WMV 9 (which they’d gotten adopted as an official SMPTE standard, VC-1, and which is part of the Blu-Ray spec, alongside 264 and MPEG-2). I was at the O’Reilly Mac OS X Conference in 2004 when Apple showed off a demo of their H.264 QuickTime decoder, which was finished and produced beautiful results, but which they wouldn’t release because of a licensing spat with MPEG-LA. Once that was resolved, Apple started using H.264 in iTunes and video iPods, which really helped the format (it didn’t hurt that cable, satellite, and Blu-Ray all went for H.264, which made for a healthy ecosystem of competitive encoders and knowledgeable compressionists).

Oh, why H.265 and not Google’s VP9 as the next major distribution video codec? Here’s why. I don’t have anything against VP9, I’m just highly skeptical of its competitive chances, and the idea that has no patent encumbrances (beyond what Google is covering). With as profligate as the US Patent Office has been in granting software patents, I’m not at all sure it’s possible to create a competitive non-infringing codec, and I surely wouldn’t want to wager that even if genuinely non-infringing, it would survive a court challenge in front of idiot jurists in the East Texas district.

Comment (1)

  1. zeze

    Hi! I’ve been following your web site for some time now and finally got the
    bravery to go ahead and give you a shout out from Austin Tx!
    Just wanted to say keep up the excellent job!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.