So I was just as shocked as Linus when he found out that PASSIVE DisplayPort to HDMI cables can now do 4K 60p per cable in DisplayPort Dual-mode…
I know right?!?
So now I want to know what element caused this to finally work… Was it the fact the DP (electrical) drivers in dual-mode are capable of a 600Mhz TMDS clock speed on a Volta card and the GeForce lineup doesn’t? We’re not talking software limitations here, we’re talking limitations of the DisplayPort Dual-mode standard…
…no passive adapters capable of the 600 MHz dual-mode speed have been produced as of 2018.
Which is BS if you watch the Linus video, as those cables he used at the end were CLEARLY passive cables. (I saw zero in built circuitry. PC Cable World is famous in Vancouver for PASSIVE cables.) and they got up to the 600Mhz rated TMDS just fine on that TV.
That begs the question… How many GPUs actually work with DP Dual-mode 4K 60hz passively, and is there HDCP issues with this setup where you can’t initiate HDCP 2.2 for instance?
What would happen if you tried that with a Gefen 4K 60hz splitter? Would the splitter freak out because the GPU doesn’t know if it’s DP or HDMI?
These are all questions I want answered, as it’s been long held belief that DP to HDMI 2.0 ALWAYS had to be an active conversion… Turns out, maybe not…
8k requires 4x HDMI 2.0, right?
So 4x 14.4Gbit/s = 57Gbit/s according to wiki
So why not make that one cable? QSFP28 for examble. Then you got 100Gbit/s wich then has room to spare, almost enough for 8k120
Rant
OH, I know! Because we don’t even have the power to consistently hit 4k60 because game engines are stuck in 4c days!
Remember that manufacturers need receiver chips and transmit chips for HDMI 2.1, and the IP for that was only just released to the chipset manufacturers. TV manufacturers have a latency of at least a couple of months before they implement the chips.
Japan’s NHK wanted 8K HDR before that, so Sharp did 4x HDMI.
If Linus wanted to experience real 8K, he should have been in Taipei (basically half a year early for Computex) on December 1st, 2018. There was a viewing party for Japan’s NHK going live with true 8K HDR content:
Because at some point, HDMI is physically bandwidth limited.
QSFP28 may not allow for many connection cycles, but then again, how often do you move your TV?
I’ve also been wondering when they finally say enough is enough and move to fiber as well.
Copper cables are a complete crapshoot when it comes to getting rated speeds. With OM4 you could have your monitor on the other end of the house and be perfectly fine.
The problem isn’t copper it’s the cable specifications a) not updating when new HDMI/display port versions are released and b) not specifying a maximum length so it’s hard for the end user to know if a cable will work. So the problem isn’t that copper is just too unreliable to work even in a certified cable, it’s that the certification doesn’t produce reliable cables.
But I generally agree that HDMI would have been a good opportunity for another consumer fibre optic connection, but I’m pretty sure that at the time HDMI was first released high bandwidth fibre optic transceivers were still very expensive, hence why toslink is cheap plastic fibre and LEDs.
I re-watched where they had trouble, and they had WAY too many varied brands of active converters at first. They needed consistency, and they ALMOST went with 4 of the exact same brand of active converter before they went with the passive cables.
So, they just mixed and matched too much, with different converter chips on each converter.
Consistency is what got them to success in the end.
Ladies and Gentlemen: I have what I believe is the smoking gun:
4 simultaneous active DP to HDMI adapters can cause enough load on the 3.3V rail over DP that certain chips in those 4 active converters are prone to failure due to undercurrent. If the GPU manufacturers only engineered for the 500mA TOTAL spec for all DP ports, 4 at the same time is NOT designed to carry that much power.
In the manual, the Club3D converters warn to not use self-powered DP splitters and MST hubs with active converters, precisely for this reason.
I wonder if they had a passthrough device (a DP AC adapter) that generates 1.5A or 2A at 3.3V for the 4 DP connections instead of connecting directly could have remedied some of the problem.
But this proves that using passive cables in dual mode, only sends the data and not the power. Requiring power for the active DP conversion and putting the 3.3V rail under gigantic strain is what caused the failures where some were only running 50Hz.
Laptops likely don’t play well with active converters for the same reason. Active converters were designed to only have 1 connected per card, and on a desktop GPU only.
@Wendell, this means there could be a market for a 4x DP to HDMI box powered by a 12V power supply rather than power from the ports. If GPUs only have DP and someone wanted to matrix 4 HDMI displays in high resolution from the DP exclusive GPU, or you had a Mac Pro, this box would solve so many problems with GPUs being unable to supply enough power for the active conversion. This would be interesting if you could tile 4 4K displays (or TVs) with this box. EDID management would be a bonus of this box.
Macs only have 1x HDMI port at most. And would have the same problem with those Mac Pros with 6x Thunderbolt 2 ports. (with 3 max, due to 1 display per Thunderbolt Bus)
You also discount the use for digital signage, where sometimes they want to use consumer displays rather than commercial displays.
Another thing that could make more sense is USB-C to DisplayPort, with 5V AC adapter support and reclocking, because I guarantee you not every USB-C device is going to supply enough power to a active converter, and that’s the only way to get HDMI 2.0. (Passive USB-C direct HDMI mode only goes up to 1.4b, and I have doubts 5V 500ma is enough for active conversion, especially if you want multiple outs, like on the new MacBook Pros)
But for the case of “it just has to work”, passive Dual mode is what is sensible. I made this thread wanting to confirm if Dual mode had gotten up to 4K 60p levels of stability, because that would make the quote from Wikipedia outdated.
I just re-watched again… and they used ACTIVE cables… not passive cables.
FUCK
I double confirmed this when a cheap passive cable could only do 1080p max to a 4K monitor.
But then that would amplify the issue if it was power related, since the TV doesn’t supply power from it’s inputs. Some voodoo magic in active converting CABLES made it work.
They were NOT passive cables, they were active cables.