https://www.tomshardware.com/monito...opted-to-stick-with-dp14-on-the-rtx-40-series
TLDR;
DP 2.1 for the proper DP80 certification tops out at 1.2m ~4 feet.
TLDR;
DP 2.1 for the proper DP80 certification tops out at 1.2m ~4 feet.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
DP 2.1 has a cable length problem
- Thread starter Lakados
- Start date
The HUB video:
View: https://www.youtube.com/watch?v=WS5CAb6s5j8
I imagine there is good reason for it, but still strange for a fully new standard with 0 possible legacy issue, that decide to call Display port 2.1 a range of signal that literally has 40gbs up to 80gbs in it, that seem complete different tier of bandwidth.
I imagine there other benefit to 2.1 vs 1.4a to use the same term (eArc for example), but still....how to make it complicated with a bunch of people connecting 80 gbs with 40 gbs cable and not understanding what is going on (or just one of the device accepting that speed)
Calling it 2.0, 2.1, 2.2..... for the 3 different bandwidth seem like would have been easier... that show that being a VESA type free to use affair instead of something trying to sell Royalty like HDMI is not necessarily magically always more clear for consumer.
View: https://www.youtube.com/watch?v=WS5CAb6s5j8
I imagine there is good reason for it, but still strange for a fully new standard with 0 possible legacy issue, that decide to call Display port 2.1 a range of signal that literally has 40gbs up to 80gbs in it, that seem complete different tier of bandwidth.
I imagine there other benefit to 2.1 vs 1.4a to use the same term (eArc for example), but still....how to make it complicated with a bunch of people connecting 80 gbs with 40 gbs cable and not understanding what is going on (or just one of the device accepting that speed)
Calling it 2.0, 2.1, 2.2..... for the 3 different bandwidth seem like would have been easier... that show that being a VESA type free to use affair instead of something trying to sell Royalty like HDMI is not necessarily magically always more clear for consumer.
Last edited:
M76
[H]F Junkie
- Joined
- Jun 12, 2012
- Messages
- 14,090
All DP has a cable length problem. So far no matter what brand of DP cable I bought none worked at high res / refresh rate. The only cables that consistently work are the ones supplied with the monitors. That are coincidentally 1.2m.
Once we actually need DP80 levels of bandwidth there will be solutions to this problem. DSC is a wonderful thing, even though people are coming out the woodwork now all of a sudden thinking its bad just because its technically compression and therefor bad.
I'd love to see these people have to tell the difference in a blind test.
I'd love to see these people have to tell the difference in a blind test.
Armenius
Extremely [H]
- Joined
- Jan 28, 2014
- Messages
- 42,662
It's an issue with data transmission over copper. The higher the bandwidth, the greater the loss in signal integrity over distance. HDMI 2.1 has the same issue. I don't know why VESA and HDMI just don't switch over to optical for the masses, already. They're going to need to, eventually.
With all the features in DP 2.1 being optional already, separating cables like that would make confusion worse. Look how hard it is for people to grasp the simple concept in HDMI with High Speed vs. Ultra High Speed.
DisplayPort is royalty-free. That doesn't mean it's free to use.
I wonder if, like with PCIe4 and beyond, some kind of redriver would work.
I saw another article about this issue yesterday. OP mentions DP80, which refers to the full 80Gbps bandwidth. There's also DP54 and DP40, which are slower but allow longer cables. As long as the cable packaging lists the speed--and the manufacturers don't lie, of course--then teaching people to look for that should mostly obviate the problem. But yeah, it'd be nice if we didn't have to deal with this crap.
YOu send money to who if your device has a DP in or out port ?
Not sure to understand, if there would be optional feature if you call your cable an exact name, to be called 2.1 it need to have x-y-z, 2.2 all of 2.1 + something else, would it not be much clearer ?
Or some feature or so orthogonal it does not work ?
Darunion
Supreme [H]ardness
- Joined
- Oct 6, 2010
- Messages
- 5,411
Component video cables come to mind. Always having to fix people who plugged them into audio ports
They exist now, those cables are directional and the require an active power connector on the receiving side.
It’s a minor annoyance but I can easily see too many consumers messing it up.
Are they trying to sell you the Monster cables???
MavericK
Zero Cool
- Joined
- Sep 2, 2004
- Messages
- 31,942
I have an HDMI optical cable and it doesn't require seperate power on either side. It does, however, seem to have current/voltage running through the connectors, which is a bit weird...
EDIT: I suspect it's due to grounding issues. A decent chunk of my older house doesn't have ground wires, so I've had to use GFCI outlets, UPSes, etc. to get any semblence of surge protection.
MavericK
Zero Cool
- Joined
- Sep 2, 2004
- Messages
- 31,942
No, just people talking about how the stock monitor cable wouldn't do X resolution/refresh/whatever with any stability, but then they bought a "decent" cable and it worked.
I know for HDMI it can make a difference, that's why they have certifications and stuff. And of course, the longer, the more iffy as to if it will do full bandwidth.
You can get some which can be powered through the HDMIs 5v out on Pin 18, but depending on cable length, refresh rate, and blah blah blah 5v might not be enough. The ones I have are easily 25’ and longer and they all require power on the receiving end.
And the 75’ one had to be done 3 times, they put it in backwards the first time, damaged it on the second, and got it right when they put in the new cable.
Yeah I can see that, I had some flickering issues with the stock cables that came with my GSync monitor, bought a generic Amazon Basics brand one and it went away.
I suspect at the facility that does the packaging they just grab a cable off the shelf and put it in the box. The worker doesn’t know nor care what it’s actually certified for or if the cable itself is any good.
Armenius
Extremely [H]
- Joined
- Jan 28, 2014
- Messages
- 42,662
You need to be a VESA member to access the full specification, which has annual dues.
2.1 has a minimum technical specification. As long as manufacturers meet it they can call their device 2.1 compliant. The aforementioned bandwidth differences are part of the optional spec. UHBR40 (10 Gbps per lane) is the minimum for 2.1, in this case. DSC, audio, and Adaptive-Sync (as far as I know) are all examples of optional features.
Cost. While fiber has gotten a lot cheaper, it still costs a lot more than copper. Not so much the fiber itself (though it does cost more) but the transceivers. For a real basic price check, we can look at a Fiberstore 100gig QSFP module. That is 100gbits over 4 lanes of data, send and receive, and runs about $100. So call it roughly $50/end/port to implement fiber. Now this isn't going to directly translate since you'd also need a return signal, and also it would probably be cheaper putting it in a device rather than pluggable but it would easily raise the price to consumer by $40-50/port on a device. That's a lot. Gonna be a hard sell to vendors that they want to go with your fiber optic standard that is going to drasticaly increase device costs. It also won't be backwards compatible, so people would need adapters if they were hooking it to a non-fiber system.
The other option is what we already have: Fiber optic cables, with transceivers on the end. Those are totally a thing, they work great, I use them at home and at work all the time. But, because of those transceivers, they cost more. Generally $100+/cable for higher bandwidth (HDMI 2.1) situations.
For DP80 Fibercommand makes such a thing. Goes from DP to MPO fiber. Can get up to probably about 150 meters over OM5. However, it's $200 for the adapters.
MavericK
Zero Cool
- Joined
- Sep 2, 2004
- Messages
- 31,942
This is what I got: FIBBR 8K Fiber Optic HDMI Cable 66ft, 48Gbps High-Speed HDMI 2.1 Cable - https://a.co/d/je9hpgI
Works fine getting 4K/120Hz to my OLED TV/AVR in the next room over from my gaming PC. 66ft was overkill for length, but whatever.
As an Amazon Associate, HardForum may earn from qualifying purchases.
I’ve got the rapid run optical cables, they have more termination options so if I need to change them from HDMI to Display Port down the road I can. Just need to change out the plate connectors, can also upgrade HDMI versions if needed that way. It seemed like a slightly more future proof cabling option at the time because walls need to be opened if we replace them otherwise.
As an Amazon Associate, HardForum may earn from qualifying purchases.
paradoxical
Limp Gawd
- Joined
- May 5, 2013
- Messages
- 301
I came into this thread to say exactly this only to see you had beaten me to the punch. The situation is frustrating to me though because I know the actual manufacturing cost of the transceivers is relatively low, and the serdes ASICS cost pennies to make even if they retail for $6-7 in QTY to companies like mine. If someone like Samsung decided to make the investment to implement a new standard and in-housed the development and production, the cost of fiber implementation would be drastically lower and perhaps in the $10-15 BOM area which is very feasible on premium displays. When it comes to high bandwidth over distance, copper is just limited compared to optical.
d3athf1sh
[H]ard|Gawd
- Joined
- Dec 16, 2015
- Messages
- 1,268
you would think they could make something similar to the optical digital cables we use for multi channel audio that wouldn't cost that much
The bandwidth for optical audio is tiny compared to video, and video needs two way signalling. Optical audio really cheaped out because they can, IIRC, the transmitters are usually led, not laser, and the cables are usually plastic, not glass fibers. Unidirectionality means no format negotiation.
For optical video, you probably need at least a laser for the source, eARC with 8-channel PCM might be enough that a laser for the sink is needed too. Then I guess there's the question of if it's cheaper to use two fibers or a single fiber and have source use a different frequency than sink.
Nah, those are garbage. As toast pointed out, they are way lower bandwidth, and they use LEDs. They are also POF (plastic) not actual glass fiber which means the cables are really lossy. You actually can get S/PDIF way farther over coax than TOSLINK because the POF cables are so lossy. You generally see only 96kHz supported over optical, and up to 192kHz over coax. The max bandwidth you get out of S/PDIF is about 8.7mbit/s (2 channels, 192kHz, 24-bits per sample). It doesn't actually natively support multichannel, that's all compressed, it is either Dolby Digital or DTS which then have to be uncompressed on the other end. In theory you could do 8 channels of 48kHz uncompressed, ADAT does that over the same fiber, though a bit different signaling, but that's as much as I've ever seen.
For high bandwidth applications you need a glass fiber, which then has to be protected and reenforced so as not to be super breakable. You need to use lasers, VCSELs normally. Also, you usually need more than one data signal. We can actually do 100gbit over a single pair but its more expensive. In networking it takes 2x 50gbit electrical lanes and you combine them to drive a single 100gig optical transmitter. That works, but those optics are in the $300 each range as compared to $100 for the ones that do them over 4 lanes. Likewise, DP and HDMI themselves are 4-lane solutions. They use 4 different pairs for signaling, so it makes sense to keep that when you go to optical. Just convert each lane electrical to optical, rather than any additional processing.
There's also communication back to be considered. S/PDIF is one way only. No handshake, no data back. The video formats have channels going back from the TV to the source for things like EDID, copy protection, CEC, and eARC. So you need to implement that too. The all-in-one prefab cables you get from companies like FIBBR usually do that with copper cales. So they'll have the video data going over 4 optical fibers, and then the control data coming back over copper. The pure fiber optic ones like the Fibercommand have to put it over fiber as well, in their case they use two fibers: one for EDID/control/HDCP/etc and one for eARC.
There are longer cables that claim DP80. I wouldn't be surprised if they have a redriver/retimer built in. I was a little surprised the prices weren't higher.
They also may just be lying. Lot of that shit goes on, has been since the HDMI 2.0 days. Plenty of cables that claimed they'd support something but were not certified, and when you got them it was a crapshoot if it worked.
d3athf1sh
[H]ard|Gawd
- Joined
- Dec 16, 2015
- Messages
- 1,268
yeah that's why i was saying something similar to the optical digital audio. like maybe take 2 or 3 fibers or whatever they need, but actually i guess modern video would need a lot more bandwidth than audio. but that's what i'm saying, why not make a multi-fiber cable, if that's what needs to be done, and price it similar to what we can get audio cables for? yes i know they use plastic but that's the whole reason they can do that to begin with is because it's usually not going a long distance. the same cost savings that you see with the audio cables, should apply to video even if slightly higher price than audio. but not $200-300 or whatever you were saying.
and optical digital / toslink has carried multichannel audio for quite a long time now and it's not lossy as long as it doesn't have to down-sample the audio for some reason. that's the biggest benefit of using a digital interface is to keep things lossless.
With how little amount of things in the field able to push and receive 80gbs right now, you could get away with it for a while.
Because, again, the loss, the bandwidth. The higher the bandwidth, the more interference and loss matters. The details get pretty technical pretty fast but simplified time and frequency are reciprocals of each other and as you increase the speed at which you pulse a signal, it spreads out in the frequency domain and this has issues eith internal loss, reflection, and interference. To keep this from happening you need to have precisely controlled fiber sizes in relation to the frequency of your light source and that means glass fibers and laser sources.
As I said: The optical used in audio wouldn't be useful for length. Max length on TOSLINK is 10m, and that's assuming a great cable. Coax can easy do 100m or more with AES signaling (works like S/PDIF but higher voltages). That's for very low bandwidth signaling. As bandwidth goes up, length goes down, yes even with optical unless it is single mode (which is only possible with lasers and again, more expensive).
Can you provide me a link to show me how that works? I ask because this is a topic I know more than a bit about and I've, literally, never encountered it. In the pro world yes, ADAT is 8 channels 48kHz uncompressed (or 4 at 96kHz, 2 at 192kHz) but I've never seen that in the consumer world. All consumer devices that I've seen that do multi-channel over S/PDIF use compression, either Dolby Digital or DTS.
So if this is not the case, I'd be very interested in seeing the devices that do it.
Yep, was the same shit in the early days of HDMI 2.0b. 99% of the people who bought those cables did it to futureproof, they were doing 1080p60 (or 1080p24) and it worked great. But that was only using 3.2gbps, not the 18 the cables claimed to be able to do. Even the few that did do 4k it was often 4k24 4:2:0 because it was all Blu-rays or the like, and so was less than like 8gbps.
Then, later, they'd get something that did 4k60 4:4:4 like a game console... and the cable would shit itself. Turns out when that full bandwidth was called for, it just couldn't hack it. But this was often years later so what are you going to do at that point?
I suspect there'll be plenty of that shit here. People will buy a "DP80" cable, it'll work fine because they are using DP 1.4. May even keep working with future devices because while right now they don't test the cables, they may learn how to and just say "oh, cable can't do 80gbit, let's just slow down and turn on DSC" and they don't notice. But some time down the road they'll probably try and actually make use of it and it won't work or will be flakey.
Or even DP 2.1 almost none of them are 80gbs... you have some AMD pro line gpu that can do it I think, but is there monitor for the other side ?
Likely the case. Spent a little more time and still didn't find any actually certified DP80 longer than about 1M. I did find retimers that hopefully will make it possible...eventually