4K over H.264? 4 Sure – and More!

As consumers who binge hundreds of video streams from cable, DirecTV, Netflix, Amazon Prime and other services – we assume that the MPEG H.264 streams we’re watching are limited to 1080p. Surprise! Those clever cooks at MPEG LA (the entity that collects MPEG fees and royalties) always had a lot more in mind. The fact is, H.264 can crunch 4K and 8K video just as easily as 1080p!

The catch is, the streams would be much bigger, way to big to travel over the average network and WiFi. So about 95% of the content you watch at home is H.264-driven 1080p. 4K streams are encoded with HEVC H.265 or Google’s VP9 or AVI technology (all You Tube content is VP9 format) – all able to scrunch the 4K stream into a size closer to 1080p dimensions.

That’s fine for 4K programming in the home, where billion-dollar technology makes it possible to send content to $30 dongles. The math is different in the commercial world, where we need moderately-priced encoders to send in-house content and signage to lobby. lunchroom, classroom, and meeting room TVs.

While 4K H.264 streams are too big for home distribution, they’re fine for dedicated commercial media networks.

About That Spinning Wheel…

Remember the spinning wheel of Netflix that tried our patience in the days of 1.5 M Internet? That’s caused by buffering – the streaming box has to pull in a number of frames into memory so it can figure out how to decode the video. Not a big problem these days, but H.264 (and HEVC/VP9 and so on) will always need buffering time. The encoding format is called Inter-Frame, so called because the stream is made up of Groups of Pictures. Each group starts with an actual picture called the I frame, essentially a JPEG. The rest of the “pictures” are just data that describe what stayed the same, what moved and what colors changed. That keeps the streams small. The decoder does the same in reverse, grabbing the first picture in the group, then storing a few more data frames, then “reading” the information to reconstruct the original video.

As consumers, we’re used to that delay, happy to trade off a little time to see a great video. If you’re giving a presentation in a conference room, you want zero delay, or a close to zero as possible. So you wouldn’t want H.264 streams – or so you thought.

Those clever H.264 gnomes had a solution, called H.264-Intra. Instead of encoding groups of pictures and data, Intra creates a stream made of individual I-Frames, compressed images. For our AV Geek readers, this is similar to Motion JPEG 2000, used in many IP video switching systems. It’s great for fast switching, as the decoder doesn’t have to buffer frames and calculate, it simply uncompresses each frame as they arrive.

An H.264-Intra decoder in a presentation room context could be the best of both worlds, able to quickly switch between Intra-generated content, but also able to playback standard Inter-frame content as well.

The Catch Is….

An existing MPEG decoder likely isn’t expecting to process 4K or intra-frame content. It’s a forgivable oversight. Vendors thought like consumers and didn’t realize that other options are part of the H.264 standard and have value for commercial applications.

But it’s an interesting concept – a more universal streaming format for presentation and distribution content, Is it better? Haven’t seen a live demonstration as yet. Is it zero latency? Of course, nothing is zero – any amount of processing adds delay, so that’s a “we’ll see” as well. Another benefit is that MPEG supports captioning data for ADA requirements – presentation codecs like Motion JPEG 2000 don’t support that.