VESA ratifies DisplayPort 1.4 standard

The VESA standards body has finalized the specification for the DisplayPort 1.4 standard. The association calls this the first major update to DisplayPort since the publication of DisplayPort 1.3, which became a specification back in September 2014.

The most important new feature in this version of the spec is Display Stream Compression 1.2. VESA says DSC 1.2 can allow for up to 3:1 compression of the video stream with end results that its members claim are "visually lossless."

The video stream compression has an immediate benefit. DisplayPort 1.3's Alt Mode already allowed for simultaneous transmission of video and data over USB Type-C and Thunderbolt interfaces. DisplayPort 1.4 goes a few steps further, enabling SuperSpeed USB (USB 3.0) data in tandem with high-definition video. Furthermore, the new standard supports 8K HDR video (7680 x 4320) at 60 Hz and 4K HDR video at 120 Hz.

There are also a few other improvements, too. In VESA's own words:

  • Forward Error Correction – FEC, which overlays the DSC 1.2 transport, addresses the transport error resiliency needed for compressed video transport to external displays.

     

  • HDR meta transport – HDR meta transport uses the "secondary data packet" transport inherent in the DisplayPort standard to provide support for the current CTA 861.3 standard, which is useful for DP to HDMI 2.0a protocol conversion, among other examples. It also offers a flexible metadata packet transport to support future dynamic HDR standards.

     

  • Expanded audio transport – This spec extension covers capabilities such as 32 audio channels, 1536kHz sample rate, and inclusion of all known audio formats.

Who cares about crummy 4K anymore? 8K is where it's going to be.

Comments closed
    • sparkman
    • 4 years ago

    [quote<]with end results that its members claim are "visually lossless."[/quote<] So it uses [url=https://en.wikipedia.org/wiki/Lossy_compression<]lossy[/url<] compression?

    • anotherengineer
    • 4 years ago

    Well I’m not buying a new monitor until it comes with at least DP 1.3 and freesync and

    [url<]http://cdn.meme.am/instances/27183696.jpg[/url<]

    • Tristan
    • 4 years ago

    Why VESA can’t increase real bandwitch over 32Gbps ?
    Super MHL will deliver 270Gbps over single cable, with much better quality uncompressed
    [url<]http://www.pcworld.com/article/3019418/new-supermhl-8k-connector-could-be-in-pcs-by-2018.html[/url<]

      • Parallax
      • 4 years ago

      This. Lossy compression does not belong in digital display connectors.

        • dyrdak
        • 4 years ago

        Now all that Monster Cables needed to do is to ship “No DSC” ™ pure copper gold plated DP1.4 cables.
        Really this is backwards. I’d understand lossless but not this joke. Hopefully this can be disabled in the drivers/device setup menu. Otherwise, what’s good is DP1.4 monitor for any graphics job if all it could present was a JPEG of the source RAW material.

        • Deanjo
        • 4 years ago

        Technically, all digital is lossy compression to some degree.

          • brucethemoose
          • 4 years ago

          True, but in between the video’s production and your screen, you want to recompress the content as infrequently as possible.

          • BobbinThreadbare
          • 4 years ago

          Not if the original image is digitally created (ie a computer game).

            • Deanjo
            • 4 years ago

            Even then there is lossy compression at some point (such as reducing polygon count on the model or movement of a line to a diagonal on a grid matrix, not rendering pixels that are not visible, etc).

        • xeridea
        • 4 years ago

        “Visually Lossless”
        A form or manner of lossy compression where the data that is lost after the file is compressed and decompressed is not detectable to the eye; the compressed data appearing identical to the uncompressed data.

        If it is not detectable by the human eye by definition, and on a 4k or 8k video, with crazy pixel density, I don’t see it mattering too much. Perhaps some graphics artists with bionic eyes may be able to tell, but then you just do 4k @60hz or 8k @ 30hz and be fine for your graphic work.

        Typical lossy compression such as jpeg or video, you can tell the difference if you look really close (perhaps zoom in), have a low quality setting, or certain imagery that shows artifacts.

      • Krogoth
      • 4 years ago

      It is because of two reasons.

      Maintaining backwards support for legacy interfaces (VGA and DVI) via adapters and there are electrical and EMI issues that arise with cabling when you try to crank up the frequency in order to get more bandwidth over the same cable.

      • meerkt
      • 4 years ago

      I don’t get MHL. It’s hard enough to understand why there’s both HDMI and DP (though I can come up with a reason), but a third standard?

      I also don’t think I’ve ever seen anyone use it in practice. Does anyone use it? Is it even viable if you wanted to use it?

      • mczak
      • 4 years ago

      There’s no way Super MHL is going to deliver 270Gbps (I think that number came up because the mhl consortium press releases doesn’t talk about hard facts, that is what the actual bandwidth is going to be, and these numbers thrown around are mere calculations based on incomplete information).
      Fact is, current MHL 3.0 has a single lane, with a data rate of 6 Gbps. For reference, DP 1.3 is at 8 Gbps per lane (with up to 4 lanes).
      From what I could gather, SuperMHL should have configurations ranging from 1 to 6 lanes, and link speed is most likely staying the same (some sources say it’s again doubled) – 6 lanes at 6 Gbps would give it a minimal edge over 4 lanes at 8 Gbps of DP only. It does however support – you wouldn’t guess it – something called (drumroll…) VESA Display Stream Compression (in contrast to DP, it looks like only 1.1 version not 1.2, but I don’t know if the differences are significant).
      Note that just about all those standards (be it HDMI, DP, MHL, Slimport,…) all use the exact same underlying signaling technology (TMDS), it is therefore not plausible some would be able to achieve significantly higher rates per link within a similar timeframe (at least not without compromises, which could be things like higher quality and/or shorter cables).

        • jts888
        • 4 years ago

        As far as I can tell, superMHL will be:
        – 6×6 Gb/s electrical lanes (36 Gb/s)
        – VESA DSC ~3:1 compression for (~108 Gb/s uncompressed equivalent)
        – 4:2:0 subsampling mandatory

        The last one is the real deal killer, infinitely more than DSC could ever be.

          • Aggressor Prime
          • 4 years ago

          Actually, superMHL supports RGB 4:4:4, YCbCr 4:4:4, YCbCr 4:2:2, and YCbCr 4:2:0 (4K resolutions and above). That just means 4:2:0 can’t be used on lower resolutions, but it doesn’t mean 4:4:4 can’t be used on 4K or higher resolutions. Check out the whitepaper on page 4.
          [url<]http://www.mhltech.org/whitepaper_request.aspx[/url<]

      • Aggressor Prime
      • 4 years ago

      PCWorld is on a roll for misreporting.

      Check superMHL’s whitepaper if you want the facts: [url<]http://www.mhltech.org/whitepaper_request.aspx[/url<] Max uncompressed bandwidth is 36Gbps (assuming superMHL 32-pin connectors on Source and Sink). superMHL can use Display Stream Compression v1.1 to achieve 108Gbps (using 3:1 compression) of "visually lossless" bandwidth. Note, this is mathematically lossy data and you can see the difference if you stand closer to your monitor. PCWorld also misreported on DP 1.4 ([url<]http://www.pcworld.com/article/3039964/new-displayport-1-4-standard-to-support-8k-displays-work-with-usb-c.html[/url<]) when they said that DP 1.3 is already in PCs. We only have DP 1.2 in PCs at the moment. Also, they are misleading when they say DP 1.4 is a "big jump." It is only a big jump because Display Stream Compression is supported in the release of the DP 1.4 standard while the technology was not yet ready for DP 1.3's release. VESA talks about DP 1.3 as if DSC v1.1 could be added later. What is the difference between DSC v1.1 (superMHL and maybe DP 1.3) and v1.2 (DP 1.4)? Not much. [url<]http://www.vesa.org/featured-articles/vesa-updates-display-stream-compression-standard-to-support-new-applications-and-richer-display-content/[/url<] *Native 4:2:0 and 4:2:2 coding From what I understand, this means we can get a 6:1 compression with 4:2:0 DSC v1.2 and 4.5:1 compression with 4:2:2 DSC v1.2 vs uncompressed 4:4:4. *Up to 16 bits per color Note, this only is new to the compression pipeline. Uncompressed data already supports 16bpc. *HDR This really isn't new to DSC v1.2 but to the DP 1.4 standard since it can now pass on HDR metadata. ----- Now that I've cleared up the misinformation on the technology, here are some of my thoughts. I hate the idea of compressing data from your video card to your monitor. Movies are already compressed enough and I don't want my video games to be compressed. So what does DP 1.4 add for me? Only HDR stands out. Thus, DP 1.4 is like HDMI 2.0a. Now HDR is important, especially for movie fans, but we will still be limited to 3840x2160@120Hz w/ 24-bit color (96Hz w/ 30-bit color, 80Hz w/ 36-bit color, 60Hz w/ 48-bit color) if we want uncompressed data. The audio changes don't matter to me since anything that can carry DTS-HD MA or Dolby TrueHD can also carry DTS:X or Dolby Atmos. Both are lossless and support object audio in 3D space.

      • ptsant
      • 4 years ago

      Do you even suspect what kind of cable and shielding that would require? This is not a trivial amount of bandwidth and we are approaching the limits of what is cost-effectively feasible (unless of course you want your DP cable to be a fiber and your monitor to cost an extra $200).

      Remember, even if this is digital, it still has to be transported as an analog signal and the laws of physics do apply.

    • Tristan
    • 4 years ago

    I am pretty sure that Pascal will have DP 1.4. Electrical parameters are the same as DP 1.3. This ‘visually loseless’ compressions is very cheap in term of transistors, and can be implemented easily on RAMDAC. It is no coincidence, that VESA release new standard in unexpected way, just before Pascal and Polaris introduction. Who else need such standard ? Probably no one else.

      • jts888
      • 4 years ago

      There is no conceivable way that DP 1.4 will make it into any GPU shipping in 2016 if the spec was just ratified this week.

      DSC was introduced in DP 1.3 over a year ago, and nobody uses it.
      Similarly, FEC transmission does not require a lot of logic, but nobody’s implementing it in GPUs already.

      Power and space budgets on Pascal and Polaris chips were set a long time ago, lithography masks were etched, etc.
      If we’re lucky, TSMC and Samsung/GloFo are already starting to pile up 14/16nm GPUs, and the best we can hope for is that we see DP 1.4 in Polaris/Pascal 2.0 in 2017.

        • Tristan
        • 4 years ago

        Very often standard is last step to complete some features. For example Vulkan, DX12 (and earlier), HBM, GDDR5X and many others are standards made publicly after they were implemented in hardware. Spec for DSC was known publicly years ago, and latest 1.2 is just small update.

          • jts888
          • 4 years ago

          DSC is just the on-wire specification for the compressed line data format.
          Even if a good encoder has been designed yet, nobody is in a rush to add it to their silicon since displays that will require the compression could still be years away.

          Digital display standards shift slowly because of this chicken and egg problem in general, and a lot of people in industry aren’t even sure if DSC is following generally the right path.

          It has only 3:1 compression since it works only on isolated scan lines with no data from past state or neighboring lines, just so a display decoder can get by with a buffer in the low kBs.
          This might be a poor priority given that things like panel overdrive and panel self-refresh/low-Hz variable sync require full screen buffers anyway.

          If DP started doing crazier things like 8×8 block-based read-out modes in addition to scanline-based or incremental read-out (send high order bits whenever pixels change, then lower order bits on subsequent frames for no change), they could get massively higher compression ratios at the same visual quality.

          There is a decent chance that a radically different DSC 2.0 will be designed before a single DSC 1.0-1.2 system ever gets built.

    • the
    • 4 years ago

    Yes! My ego demands more pixels!

    • Sargent Duck
    • 4 years ago

    I have to ponder, if a manufacture is going to release a new monitor with shiny DisplayPort 1.2, how hard is it REALLY to put DisplayPort 1.4 in?

    After all, the standard is free and the port is the same, so….I’m not seeing why the market isn’t flooded with the latest and greatest DisplayPort spec.

      • brucethemoose
      • 4 years ago

      Wikipedia claims the DP royalty rate is $0.20 per unit, but this contradicts other info I’ve seen.

        • jensend
        • 4 years ago

        DisplayPort in and of itself is a free specification, no royalties required.

        The MPEG-LA, however, added DisplayPort manufacturers to the long list of people they’re trying to shake down for money. They call it royalties but it’s actually a protection racket.

        As a manufacturer using DisplayPort, you can thumb your nose at the MPEG-LA, who probably don’t have a leg to stand on regarding their patent claims, but you’ll probably be badly hurt by the legal fees, and there’s a risk the court system will rule in favor of evil and you’ll have to pay up big time. Or you can pay the protection money like a good peon and not get beaten up.

        Or you can just skip including DP entirely and thus not have to worry about it. Slowing DP adoption is probably most of why MPEG-LA is doing this, since Big Content people prefer HDMI, which they have control over.

        Our present IP law basically enables the tech world to be ruled by gangsters and thugs.

          • brucethemoose
          • 4 years ago

          What does MPEG-LA have on VESA/Displayport? Seems like their patents cover video encoding/decoding, but (until recently), Displayport hasn’t done that.

          Does it have anything to do with their current compression scheme? If so, it seems like a very compelling motive for VESA to move away from compression.

            • Krogoth
            • 4 years ago

            It is for HDCP support.

            Displayport doesn’t require HDCP but it doesn’t stop monitor and video card manufacturers from adding it on which requires them to pay MPEG-LA royalties for it.

            • brucethemoose
            • 4 years ago

            Ah, that makes sense.

            It’s also kind of sickening… A free standard has royalties forced onto it to support HDCP, of all things.

            • Krogoth
            • 4 years ago

            Pretty much, it is a stupid loophole.

            • jensend
            • 4 years ago

            Actually, this is not true at all. MPEG-LA is claiming their patents apply to any use of displayport whatsoever. This has nothing to do with HDCP. Please don’t engage in misinformed speculation.

            [url=http://www.mpegla.com/main/programs/DisplayPort/Documents/displayport-att1.pdf<]Here is their list of the patents they're willing to disclose publicly[/url<]. This includes such gems as a patent on "sending a different signal depending on the display's capabilities" and another on "sending two things at once by using packets and multiplexing." Prior art abounds.

            • jensend
            • 4 years ago

            See my below reply to Krogoth, which includes a patent list. See also [url=http://www.mpegla.com/main/programs/DisplayPort/Documents/DisplayPortweb.pdf<]this briefing[/url<]. They claim that any implementation of any DP standard infringes these patents; this has absolutely nothing whatsoever to do with HDCP or with the kind of video encoding or compression patents you're thinking of.

          • Krogoth
          • 4 years ago

          HDMI is an A/V interface while Displayport is a computer interface. They are cater towards two completely different demographics.

          MPEG-LA doesn’t care about Displayport. The real reason is that video card and manufacturer need to pay royalites to MPEG-LA to get HDCP support.

            • jensend
            • 4 years ago

            The most relevant differences between A/V video and computer video have long since disappeared. The technical requirements overlap almost entirely.

            The one real difference is the question of political control. That is the only reason a “demographic” difference persists.

            MPEG-LA most definitely does care about DP.

            • Krogoth
            • 4 years ago

            HDMI is found on equipment catered towards A/V market that usually finds itself in living room.

            DP is only found on monitors that are meant to be attach to some kind of computer. You will never find DP on A/V equipment or anything marketed to be in the living room. Gaming consoles are considered to be A/V equipment.

            The only reason that HDMI ports are even is found on computer monitors is because monitor manufacturer often use the same PCB and LCD panels for their smaller HDTV units on larger computer monitors (16:9 only). It saves the bottom line on production costs.

            HDMI is the successor to S-Video, composite, component and BNC interfaces while DP is the successor to DVI and VGA.

      • Tristan
      • 4 years ago

      For now there is no need for DP 1.3 None 4K panell support refresh higher than 60Hz, and 5K monitors are very rare (just only three models), and nonody support HDR. All this change, when faster 4K panels with HDR will be available.

    • kcarlile
    • 4 years ago

    I love the term “visually lossless” And here I thought the devs at work made that up themselves.

      • ptsant
      • 4 years ago

      Well, considering that your monitor already introduces measurable overshoot, color deviation, gamma deviation and all that, I think that the contribution of a decent compression algorithm at 3:1 should be minimal. If you use an external source of images like a digital camera, you also have to live with sensor noise, limited dynamic range etc. If you use games, you already live with 16-bit floating point approximations and I can tell you that game shaders are not written to be correct at the 20th decimal place. If you see movies, consider too that x264 at 10:1 is usually “high quality”.

      So, the visual chain is very long and there is a lot of noise in there already that most people forget about. An additional minor loss of information has to be empirically evaluated. If people do see it, it’s bad. If people don’t, it’s OK.

        • Cannonaire
        • 4 years ago

        Additional lossy/noisy steps [i<]amplify[/i<] the problem. It's why transcoding an mp3 to another lossy format even once can make the music sound like it's coming from a tin can. [url<]https://en.wikipedia.org/wiki/Digital_generation_loss[/url<]

      • TheMonkeyKing
      • 4 years ago

      I think they dropped “virtually” from the quote because even they realized even that was too much.

    • chuckula
    • 4 years ago

    [quote<]Display Stream Compression 1.2. VESA says DSC 1.2 can allow for up to 3:1 compression of the video stream with end results that its members claim are "visually lossless."[/quote<] Well here's a whole new way to differentiate between crappy monitors that won't have the ASIC quality to really implement the decompression properly and higher quality monitors that try to get it right. Additionally, will there be a way to turn this off for professionals who actually want to see the real images without compression/decompression artifacts?

      • brucethemoose
      • 4 years ago

      On top of that, the compression/decompression could add some latency as well.

      If the lossy compression can’t be turned off/bypassed, I’ll be avoiding DP 1.4 like the plague.

      • the
      • 4 years ago

      The real question is how well DSC 1.2 compares to alternatives already on the market. I’ve seen what SVSI claims is ‘virtually uncompressed’ 1080p over Gigabit Ethernet and it looks [i<]horrible[/i<]. Resolution gets cut to 720p and then the 4:4:4 gamut gets shifted to 4:2:0 so everything looks like a bad 1970's soap opera. Not sure if the transmitted refresh rate got chopped to 30 Hz as the scenes I was watching had little motion (the receiver did link to the display at 60 Hz though).

        • brucethemoose
        • 4 years ago

        [s<]Fortunately, cutting the uncompressed bitrate by 2/3 isn't as impossible as squeezing 1080p through a gigabit line. I think some lossless video codecs can even achieve that rate with some videos, but an ASIC in a monitor probably can't do it fast enough.[/s<] EDIT: Actually, I looked at the numbers... Maybe that was a crumby adapter. Or maybe it's a bad omen for the cheap compression schemes we'll see with DP 1.4 🙁

      • Tristan
      • 4 years ago

      of course you can disable this compression, but framerate will drop

      • ptsant
      • 4 years ago

      Correct me if I’m wrong, but usually loss happens mostly at the encoding part, which will obviously be highly standardized. Decoding should be practically bit-wise identical. I am aware that some decoders are said to obtain better quality (rounding errors, maybe some post-processing?) but the x264 standard requires all decoders to obtain similarly. It’s the encoder part that has freedom, not the decoder part.

    • SlappedSilly
    • 4 years ago

    I can count the number of DP1.3 products I’ve seen on 1 finger. Do you suppose we will see 1.5 or 1.6 ratified before 1.4 products reach the market?

      • chuckula
      • 4 years ago

      [quote<]I can count the number of DP1.3 products I've seen on 1 finger. [/quote<] You've seen one?!?!

        • SlappedSilly
        • 4 years ago

        I… I thought… but… ;_; I dreamed it?

        • Tristan
        • 4 years ago

        AMD promises on slideware

      • mczak
      • 4 years ago

      Well, 1.4 is only a “minor” update, since at the electrical level it’s identical. I wouldn’t expect another “major” DP update soon. There is definitely a need for DP 1.3 products (the new AMD “polaris” based graphic cards arriving in a couple months will definitely have it, near certainly Nvidia “Pascal” based cards too), since right now you can get 5k monitors which require the kludge of needing 2 DP 1.2 connections (well for 60 Hz operation anyway). Hopefully 5k monitors will appear which support DP 1.3 rather sooner than later (traditionally, monitors lag behind adopting new standards compared to graphic cards, however).
      I suppose the new color space support of DP 1.4 will see pretty rapid adoption (really doesn’t require anything special at all neither for source nor sink), and I bet the display compression is very cheap to implement, so it shouldn’t take too long neither (that said, on the monitor side, I’d suspect only really 8k display would bother with that, albeit it would be useful for 5k displays with high refresh rate too).

Pin It on Pinterest

Share This