Interesting Video on ATSC 3.0

Discussion in 'Over the Air TV By RabbitEars.Info' started by Scott Greczkowski, May 13, 2016.

  1. I think some may be ignoring what that means in the short term. With so much talk of combining and the players that haven't already combined sporting 1080i feeds, who will have to cut back to 720p or even 480i to "cooperate"?
    I'm not convinced that ATSC 3.0 is all that far along. There are quite a few of the underpinnings done but video is still open and that is critical to delivering television. The terms used by Chairman Friedel were "virtually all are essentially done". Sounds to me like double-talk. A/300 itself (the ATSC 3.0 standard) is in Candidate Standard status until the end of July and A/341 was advanced to balloting as a Proposed Standard just five days ago. If the usual track from Proposed to Candidate to Adopted Standard is followed, I'd be surprised if they ratify the video standard portion by September. MPEG-H sound that will be used outside of the US was approved just four days ago. AC-4 sound has been approved for less than two weeks.

    Even if the broadcasters are given the green light, they won't be able to sunset the DTV channels for years and there will be a few years where inertia demands that they maintain their DTV feeds in HD. While all of this is happening, what's in it for us in terms of advanced sound and video? How can networks "share" a UHD station where there an be only one channel?
     
    jamesjimcie likes this.
  2. You can see the various standards which have yet to be adopted/finalized here:
    http://atsc.org/standards/candidate-standards/

    Of those, yes, standard A/341 for using the HEVC video codec is the most important but I don't think there's any question that ATSC 3.0 will be HEVC-based. That's been the expectation for years now and there's no other real possibility except for VP9 (Google's royalty-free codec), although I haven't seen any discussion recently about that potentially being adopted as a standard for ATSC 3.0. (Over in the UK, the BBC is a big backer of VP9, so perhaps it will get incorporated into whatever next-gen broadcast standard they adopt in the future.) Final adoption of A/341 is a formality, I'd say. The codec train has left the station and it's HEVC.

    As for the repercussions to ATSC 1.0 signals that will come when broadcasters cooperate, yes, I expect that there will be some loss of HD picture quality and/or complete loss of some SD subchannels. I don't see how that doesn't happen. In some markets, expect to see 1080i signals get downgraded to 720p, with two or possibly even three channels' 720p signals emanating from the same ATSC 1.0 tower (although upgrading to the latest generation of MPEG-2 encoders will help soften the blow to picture quality).

    As for stations sharing the same ATSC 3.0 tower, yeah, I don't know that two simultaneous UHD broadcasts could fit in the available bandwidth, which would mean that cooperating stations would have to take turns as to when they could air content in UHD. (That said, I don't expect there to be a whole lot of UHD content being broadcast for the next few years anyhow.) However, ATSC 3.0 does allow for the video stream to be divided into different layers that can travel over different paths and be combined by the tuner into the final payload, so it's possible that a station's main high-power ATSC 3.0 signal could be 1080p and then a supplemental UHD "upgrade" signal could be delivered from a cooperating low-power station or maybe even over a broadband internet connection. ATSC 3.0 is pretty flexible like that. It'll be interesting to see what kind of solutions local broadcasters can engineer.
     
    jamesjimcie and osu1991 like this.
  3. This doesn't deserve to be a parenthetical statement. It may represent the entire case for adoption of ATSC 3.0.

    If they don't deliver a viewing experience that is perceptibly superior to DTV with the new standard, there's precious little motivation for the consumer to upgrade. I maintain that the superiority must be a quantum leap and even then, adoption will be slow. What they're promising is nothing like the transitions from black-and-white to color or from SD to HD and the addition of color was the last time we went through a non-mandatory transition. I believe the proponents are fooling themselves in a most dangerous way if they think they can get this done without a mandate.
     
    jamesjimcie likes this.
  4. Yeah, no doubt that's a cause for concern in terms of whether or not ATSC 3.0 will succeed. Some people can't tell the difference between SD and HD while others of us can tell the difference between the picture quality of the average OTA 720p signal and the average 1080p stream from Netflix. If local broadcasters can advertise they offer superior picture quality with "the higher resolution of full HD 1080p, plus improved color and contrast from HDR," as the norm for most shows, then it may be OK that there are only a few big shows, movies and major events here and there broadcast in UHD for the first few years. (I bought my first HDTV back in 2003, I believe, and there was still a lot of primetime stuff in SD then, including everything on Fox.)

    I think local broadcasters will need to offer more carrots to spur adoption than just better picture and sound, though. That's why I think integrated on-demand streaming of recent shows -- like the UK's Freeview system offers -- will be important. (Although while ATSC 3.0 can certainly support such a thing, it remains to be seen whether it will happen.)

    A mandate still may happen, we'll see. If 2019 rolls around and lots of new TVs sold here still don't have a built-in 3.0 tuner, well, that would be a problem. But given the expected marginal incremental cost to install a dual 3.0/1.0 tuner versus a simple 1.0 tuner in a new TV, I don't think many manufacturers will want to risk omitting that bullet point from their marketing claims.
     
    jamesjimcie likes this.
  5. Therein lies the egregious omission: ATSC 3.0 doesn't support 1080p nor does it support HDR on HD content!
    You seem to be assuming that most everyone is on a path to upgrade most, if not all of their televisions in the next five years. The economy hasn't bounced back that much.
     
    jamesjimcie likes this.
  6. It would appear that this NPRM doesn't meet the first two bullet points in that article:
    • According to Trip, Congressional legislation mandating a new modulation standard doesn't exist.
    • I'm pretty sure there were no "comments from the public" that drove this NPRM.
     
  7. Number 23 is interesting. (Page 12)
    http://transition.fcc.gov/Daily_Releases/Daily_Business/2017/db0224/FCC-17-13A1.pdf
     
    jamesjimcie likes this.
  8. jamesjimcie likes this.
  9. I'm not sure where you're getting that info. ATSC 3.0 definitely does support 1080p and it also supports HDR (although I don't believe one particular HDR format, such as HDR10, Dolby Vision or HLG, has been endorsed).
    https://en.wikipedia.org/wiki/ATSC_3.0#Video
    http://atsc.org/newsletter/high-dynamic-range-planned-for-atsc-3-0/

    This article from a little over a week ago is an interesting read, including this bit:

    When it comes to higher-quality content, broadcasters said they see HDR as a first step before 4K/UHD. Vecchi said HDR is more cost-effective for distributors than UHD, and Pizzi argued that it will have a bigger impact on viewers, as well. "If you're using 1080p, it's the biggest bang for the bit," he said. Fox Networks EVP/GM Richard Friedel said Fox plans to offer HDR with the launch of ATSC 3.0 broadcasts.

    Sinclair Broadcast Group SVP/CTO Del Parks agreed that UHD is expensive, but noted that most of Sinclair's stations could handle 1080p60 broadcasts with HDR. "That's our target," he said, "and we've already started looking at how to do it." CBS & CW TV Networks VP of Engineering and Advanced Technology Bob Seidel said CBS has been active in HDR tests and demos but hasn't made any decisions on how to move forward.
     
    jamesjimcie likes this.
  10. The NPRM seems to be dominated with fundamental questions about how this transition is going to work. All the while Pai appears to be all in. Maybe it would have more impact if the other members of the committee seemed more "onboard". His predecessor, Tom Wheeler, took a similar tack and I'm not convinced it was all that fruitful; remember "Net Neutrality"?.

    The writing seems to be on the wall with OTA reception going forward: nothing is guaranteed. Even if a station doesn't have to repack, they may choose to move their DTV operations to another channel or add their ATSC 3.0 channels on a different channel.

    Advertisers had better be on the lookout for changing customer numbers.
     
    Comptech and jamesjimcie like this.
  11. From the A/341 document
    There is mention of how 1080p video is to be handled but it appears under the "Progressive Video" heading and it isn't clear what the ultimate output will be.

    HLG is one of the HDR schemes supported. The other scheme, Perceptual Quantizer, is part of both HDR10 and DV but for some reason, they don't specify which, if either, is going to be supported or if PQ can operate on its own.

    It is possible that 1080p is supported under the class of progressive formats but it isn't clear how. The candidate standard seems a little loosey-goosey in the Progressive Video discussion as compared with the SD and HD run-downs.
     
    jamesjimcie likes this.
  12. What's unclear? The document you referenced clearly states that HEVC-encoded progressive scan video will be allowed up to a max resolution of 3840 x 2160 (UHD), but that, of course includes 1920 x 1080 HD. Here's the key excerpt (bolded emphasis is mine):

    6.2.3.1 Progressive Video Formats • The spatial resolution shall be constrained to not more than 2160 lines and 3840 horizontal pixels. • The spatial resolution in both dimensions shall be evenly divisible by 8. • The picture rate in 60 Hz regions shall be one of the following in Hz: 24/1.001, 24, 30/1.001, 30, 60/1.001, 60, 120/1.001, 120. • The picture rate in 50 Hz regions shall be one of the following in Hz: 25, 50, 100. ATSC S34-168r9 Video – HEVC 2 March 2017 10 • The scan shall be progressive. • The pixel aspect ratio shall be 1:1 (square pixels). Coded representation of video with 1080 lines (e.g., 1920x1080) may be coded either as 1080 lines or as 1088 lines.
    When the video is coded as 1088 lines, the bottom 8 lines shall be black.


    And as the 6.2 section you quotes states, HDR, WCG, and HFR features won't be available for 1080i or any SD signals. They will, however, be available for 1080p and UHD signals.
     
    jamesjimcie likes this.
  13. So the question is whether or not they'll blow their bandwidth budget on such a format. I would imagine that a fully enhanced 1080p stream would consume as much as half of the available bandwidth of a channel.

    There's also the issue of finding content that is encoded for HDR and WCG at 1080p. Most of what I've read suggests that real-time encoding is still well into the future, much less real-time transcoding.

    Remember, that in order to get people to adopt the standard, they're going to have to make it attractive right away and they won't be afforded the luxury of a mandatory DTV sunset.
     
    jamesjimcie likes this.
  14. Adding HDR, etc. to a video stream doesn't require much extra bandwidth. And while UHD would probably take over half of a station's available bandwidth, a good-looking 1080p stream, even with HDR, should require no more than 20% of an ATSC 3.0's approximately 25 Mbps of available bandwidth, i.e. 5 Mbps. Netflix streams very nice 1080p at 5.7 Mbps in the less efficient AVC h.264 format. HEVC h.265, which ATSC 3.0 will use, reduces the bitrate requirement considerably. Granted, that's for pre-encoded material, not real-time encoded. So if we're talking about live sports at 1080p60, that would likely mean somewhat more bandwidth required, but not twice as much.

    If the article I linked above, about Fox and Sinclair's plans to use 1080p60 HDR as their main format under ATSC 3.0 is any indication (or this one about new TV cameras), I don't think there will be an issue in terms of new content being available in HDR. I would guess pretty much all scripted TV shows will move to being shot in UHD HDR (some already have), although networks and local affiliates may choose to down-res shows to 1080p HDR for broadcast, particularly in the first few years of ATSC 3.0. As for older existing content, Technicolor has already developed a solution that allows for SDR content to be artificially upscaled to HDR. I'm sure that the effect is not as good as if the content was actually filmed in HDR but it's still an improvement. So perhaps we'll see that used for some popular older shows and movies, the way we saw Seinfeld get remastered in HD for rerun syndication.

    All that said, you're right that consumers won't seek out ATSC 3.0 signals and buy new tuners to receive them unless the public perceives an improvement over the existing broadcast system. And while I plan to do that, frankly, I don't think a whole lot of OTA viewers will and I think broadcasters know that. They have to realize that for ATSC 3.0 to succeed, it will require a "if we built it, they will eventually come" attitude. I think 3.0 tuners will become the norm, with or without a mandate, in new TVs, and as new TVs are bought, viewership of 3.0 signals will slowly increase. It will be years before viewership of 3.0 OTA broadcasts exceeds 1.0 broadcasts.
     
    jamesjimcie likes this.
  15. Most new UHD TV don't have tuners. They are sold as Home Theater Displays. Even the manufactures are tired of government mandates!
     
    jamesjimcie likes this.
  16. That's what someone theorized about multi-channel audio once. Later we came to find that it would overload optical digital connections.
    Your reasoning starts out badly and contradicts your first statement according to projections. One stream of UHD with HDR and WCG is supposed to take up 100% or more of the available bandwidth (when compression is ideal). Obviously, compression will rarely be ideal.
    I think you're giving too much credit to hardware that is just getting started. Again, if the compression does noticeable damage to the content early on because it isn't close to ideal, consumers aren't going to willingly buy in.
    Would you buy a used car from a Fox affiliate?
    TurnerVision all over again?
    The repack doesn't afford them that luxury. They can't start in the small markets and hope to develop a groundswell of support for a new technology that they desperately want to replace DTV. They're going to have to do a whole lot better than the conventional MPVDs have done at deploying UHD. DIRECTV, the self-proclaimed industry leader, is well over two years into their much-ballyhooed efforts and they haven't expanded much at all; still just one demo channel, a once in a while live sports feed and a PPV channel if my information is correct. Back in the early days DIRECTV was claiming they would have "dozens of channels by 2016".

    According to research by Strategy Analytics, they forecast that nearly 50% penetration of UHD may happen by 2020. Not only is the penetration going to have to be 100%, but it is going to have to be pretty much every TV that is UHD before a DTV sunset could be discussed.

    So where will the clear advantages lie for the consumer? They can talk all they want about their plans and hopes, but consumers are going to want to see it in action before they commit to converting everything over and time is something they simply don't have. I can envision most of the OTA TiVotees voting no simply because their lifetime DVR experience would be compromised at best.
     
    jamesjimcie likes this.
  17. jamesjimcie and NashGuy like this.
  18. I'm sorry, dude, but you continually make false and/or misguided assertions about ATSC 3.0. I enjoy a good back-and-forth and sharing of info on a forum as much as the next poster but what's the point if you keep saying stuff that just isn't true? Please provide links that back up what you say. It's fine to have your own opinion (which seems to be that the future of OTA broadcast isn't good, I guess?) but you can't have your own "alternative facts".

    Just to start, one stream of UHD with HDR and WCG does NOT have to take up 100% or more of the available bandwidth in the 6 Mhz spectrum available to an ATSC 3.0 broadcast station. The available bandwidth is usually stated as about 25Mbps, although it actually varies based on how the transmission is configured; the more robust the signal, the lower the bandwidth but the easier it is to receive, the less robust the signal, the higher the bandwidth but the harder it is to receive. Both Netflix and Amazon are already streaming UHD with HDR in HEVC at bitrates of about 15 to 16 Mbps; I watched that content from both providers on my UHD TV and it looks great. (Please note that 16 is a much lower number than 25.)

    As to the preceding point about the feasibility of broadcasting in 1080p HEVC using the available bandwidth in ATSC 3.0, check out this article from an industry news source (emphasis mine), which quotes:

    Even those who are not looking to broadcast UHD will see massive benefits. Up to six 1080p signals with higher dynamic range and higher frame rate can be delivered (or up to eight 720p or 12 1080i signals). Those services can also be mixed and matched with 360p services for mobile devices and much more.
     
    jamesjimcie likes this.