AVS Forum banner
1 - 20 of 24 Posts

·
Registered
Joined
·
8,103 Posts
Discussion Starter · #1 ·
This article explains that DVI, and HDMI was designed around the needs of content provider industries, not video quality. The short version is they both use twisted pair (cat5) which is more suitable for computer cables. Twisted pair was designed for 2-way comunication, and error correction. It has poor resistance tolerance, and lacks the bandwidth needed for high quality video. It says Broadcast engineers would have chosen coax instead. It's high bandwidth, one-way, etc. But it looks we are stuck with HDMI for a long time. So for now be happy with your content provider industry recommended $110 Monster brand twisted pair HDMI cable.

http://www.audioholics.com/education...tter-with-hdmi
 

·
Registered
Joined
·
9,578 Posts

Quote:
Originally Posted by icecow /forum/post/0


This article explains that DVI, and HDMI was designed around the needs of content provider industries, not video quality. The short version is they both use twisted pair (cat5) which is more suitable for computer cables. Twisted pair was designed for 2-way comunication, and error correction. It has poor resistance tolerance, and lacks the bandwidth needed for high quality video. It says Broadcast engineers would have chosen coax instead. It's high bandwidth, one-way, etc. But it looks we are stuck with HDMI for a long time. So for now be happy with your content provider industry recommended $110 Monster brand twisted pair HDMI cable.

DVI had nothing to do with the content providers. It never provided any of the protections they wanted. DVI was created for PCs to connect to displays sending digital signals that eliminated many analog artifacts.


HDMI simply extended the same concept to AV, added audio as well as video, different colorspaces, and least (or most) importantly incorporated HDCP as a virtual requirement (it isn't technically required but nobody puts out HDMI without HDCP) HDCP was also made available on DVI systems.


Personally I think HDMI/DVI look fine. It uses twisted pair, but that isn't a bad thing. You can get very affordable cables that transmit bit perfect data for very cheap. I call these digital cables the Monster-killer. Virtually all those claims that Monster likes to make with analog cables rarely apply with digital cables and it can be proven.


I think where HDMI really falls down is interop testing. It will eventually get better, but the early stuff had lots of compatibility issues.
 

·
Registered
Joined
·
8,103 Posts
Discussion Starter · #4 ·
"This article explains that DVI, and HDMI was designed around the needs of content provider industries..."


What I meant to say is DVI and HDMI use twisted pair and HDMI was designed around the needs of content providers. My writing has been sucking.


Other than that you don't agree with the author. I bought in to it, it was technical and well written.
 

·
Registered
Joined
·
5,928 Posts
I am using DVI and HDMI and HDMI>DVI. I am just not having any problems at all with any of it so the fact we are stuck with it isn't of much concern to me. My longest run is an HDMI cable of about 20'. If my installation was different, I might have some complaints but for mine and I bet most, it will work great.


Chris
 

·
Registered
Joined
·
290 Posts
"It has poor resistance tolerance, and lacks the bandwidth needed for high quality video."


It's a good thing they only do digital signals then, instead of analog. Otherwise people would probably give a rats.


It's digital. It's either a 1 or a 0. there is no "quality" margin.
 

·
Registered
Joined
·
9,578 Posts

Quote:
Originally Posted by kurhurdler /forum/post/0


"It has poor resistance tolerance, and lacks the bandwidth needed for high quality video."


It's a good thing they only do digital signals then, instead of analog. Otherwise people would probably give a rats.

Shh...


Don't tell those people buying $3k+ component matrix distribution systems based on CAT6 twisted pair wiring they lack the bandwidth for high quality video


https://www.audioauthority.com/indexh.php
 

·
Registered
Joined
·
186 Posts

Quote:
Originally Posted by sfhub /forum/post/0


... I think where HDMI really falls down is interop testing. It will eventually get better, but the early stuff had lots of compatibility issues.

A lot of it still has compatibility issues. I'm working on an HDMI product for a client (fiber-optic HDMI solution), and in testing we've found about a third of the DVD players we've tried so far don't follow the HDMI spec. (The DDC doesn't support clock-stretching.) The funny thing is that the DVD players and our product both pass the HDMI compliance tester. (Note that these problems won't occur when using a regular HDMI cable.)


I'm curious to see what happens when CEC (Consumer Electronics Control) becomes generally available.
 

·
Registered
Joined
·
8,103 Posts
Discussion Starter · #10 ·
My working assumption is the poor resistance tolerance translates to lower bandwidth and/or shorter cable runs.


btw, I just spotted a Phillips HDMI cable on clearance for $48.74 at walmart (really).


another thing the author said is HDMI twisted pair has poor switching abilities. I'm guessing that means when it runs through a receiver's switch that some kinda problem is possible. know anything about that sfhub (what the heck does 'sfhub' mean anyway?)?


If HDMI quality is good enough it's good enough, but that logic troubles me because it is always possible to lower video quality to fit any given bandwidth. If a higher bandwidth was available I'd think that would constitute to a better picture. I couldn't tell you because I don't know what are the figures involved.
 

·
Registered
Joined
·
9,578 Posts

Quote:
Originally Posted by icecow /forum/post/0


another thing the author said is HDMI twisted pair has poor switching abilities. I'm guessing that means when it runs through a receiver's switch that some kinda problem is possible. know anything about that sfhub (what the heck does 'sfhub' mean anyway?)?


If HDMI quality is good enough it's good enough, but that logic troubles me because it is always possible to lower video quality to fit any given bandwidth. If a higher bandwidth was available I'd think that would constitute to a better picture. I couldn't tell you because I don't know what are the figures involved.

It's not that HDMI inherently has poor switching compatibility. If implemented correctly it switches fine. There are 2 ways to switch HDMI, pass-through method and repeater-style method.


Pass-through does exactly what you think, it just essentially connects the inbound and outbound wires. If implemented correctly it should be very compatible. However in my experience it is more susceptible to errors in implementation due to electrical/timing issues. Again, if done right, it should be very compatible. I have a manual 2-way switcher that has worked with everything I've thrown at it and I've tried a lot of stuff. I have a 5-way powered switch that is probably compatible with 98% of the stuff out there, but does sometimes have problems.


Repeater-style should theoretically be just as compatible. The switcher decrypts HDCP then reencrypts and resends further down the line. One advantage of this method is the switcher (possibly your AV receiver) can then do video processing, like scaling or deinterlacing) because it has access to the unencrypted data. However the incompatibility found in this method is the HDCP key negotiation sequence for repeater-style is more involved and many cable set-top-boxes did not implement this correctly, which meant many of them didn't work with AV receivers.


HDMI delivers 1920x1080p with bit-perfect transfers for me. This is as high or higher resolution than almost all mass-market video sources today, so for me it works fine. I also have run 1920x1080p at 50ft lengths, so that is plenty for me. The bandwidth has already increased with HDMI 1.3 to something that should satisfy people for the foreseeable future.
 

·
Registered
Joined
·
164 Posts

Quote:
Originally Posted by kurhurdler /forum/post/0


"It has poor resistance tolerance, and lacks the bandwidth needed for high quality video."


It's a good thing they only do digital signals then, instead of analog. Otherwise people would probably give a rats.


It's digital. It's either a 1 or a 0. there is no "quality" margin.


It's either 1 or 0 at a particular time - digital is very sensitive to timing (as measured by jitter). Those 1's and 0's are transported by analog voltages/currents and are subject to degradation. It's the digital interface that receives the signal that has the responsibility to reconstruct the original - with loads of bandwidth available over what is needed to transport the signal, redundancy can be built in to make the interface much more likely to rebuild the original signal. The qualities of the wire affect the timing and shape of the signal. Coax is not necessarily superior to well-constructed twisted pair in a balanced config.

 

·
Registered
Joined
·
9,578 Posts

Quote:
Originally Posted by rf75 /forum/post/0



It's either 1 or 0 at a particular time - digital is very sensitive to timing (as measured by jitter). Those 1's and 0's are transported by analog voltages/currents and are subject to degradation.

This is all true. One point not mentioned though is the big difference for end-users with HDMI is if there is an error it is immediately obvious as sparkles, full static, no picture, blanking picture, etc. You don't get the gradual degradations like with analog. Because of the way the signal is transmitted, you simply won't get stuff like faded colors, soft edges, diagonal scrolling lines, etc. There are a whole new set of degradation artifacts, but they are immediately obvious so if your picture looks fine, your cable is plenty good and buying a more expensive won't get you a better picture.
 

·
Registered
Joined
·
339 Posts

Quote:
Originally Posted by sfhub /forum/post/0


This is all true. One point not mentioned though is the big difference for end-users with HDMI is if there is an error it is immediately obvious as sparkles, full static, no picture, blanking picture, etc. You don't get the gradual degradations like with analog. Because of the way the signal is transmitted, you simply won't get stuff like faded colors, soft edges, diagonal scrolling lines, etc. There are a whole new set of degradation artifacts, but they are immediately obvious so if your picture looks fine, your cable is plenty good and buying a more expensive won't get you a better picture.

Exactly.
 

·
Registered
Joined
·
265 Posts
Haven't visited AVS Forum in a couple of weeks, and so this is my first opportunity to comment.


I read the article. It certainly has some valid points, but then we have to take into consideration the author's point of view: It was written by someone from Bluejeans Cable who is working on a superior HDMI cable for difficult installations (read: LONG cables), and who points out how much better things would work from an electrical engineering perspective if the HDMI standard had used coaxial cable from the get-go rather than twisted pair.


And he's right: whether the transmission is analog (e.g., component) or digital (e.g. HDMI), when very high bandwidths are involved, to be able to cover the longest possible distances coaxial cable is going to give you better performance than twisted pair -- especially when the signal is not just a single stream of bits but rather a group of signals (4 in the case of HDMI) that must be timed very closely to one another. As much as I am a proponent of twisted pair as an inexpensive solution in a lot of situations, this really is an area in which coax shines.


But the article is addressing how to make long cables -- an issue the typical consumer isn't going to be faced with, though the pro AV installer will be. Adopting coax for HDMI would make the cable much bulkier and the connectors more massive (more like the bulk of early DVI cables, though the digital part of DVI is twisted pair, just with the potential for more pairs). And the cost would be a lot higher -- again, an issue in mass-market consumer electronics. The choices made in the HDMI spec hold down the cost of consumer gear (the price of Monster Cable notwithstanding).


AV pros sending signals over 75, 100, 150, 200 feet are in trouble with HDMI, compared to component. There WILL be good technical solutions -- HDMI repeaters, converters that use coax or fiber optics over a long distance, etc. -- but they will all involve active electronics and fancier cable (or fiber) and will be pricy indeed.


99 out of 100 consumer installations won't have these issues, though, and will benefit from the lower cost of the approach that was adopted. For the typical consumer getting into HD, the twisted-pair HDMI standard is just fine.


(There are some interesting discussions over on the "Home AV Distribution and Networking" forum here at AVS. The preferred solution for long runs cable runs seems to be to run component over 3 RG-6 coaxes. How's THAT for bulk?)
 

·
Registered
Joined
·
9,578 Posts

Quote:
Originally Posted by Kenlex /forum/post/0


and who points out how much better things would work from an electrical engineering perspective if the HDMI standard had used coaxial cable from the get-go rather than twisted pair.

You realize of course that the HDMI standard specifically makes no mention of the physical cable medium, so as not to lock in manufacturers to specific material. The specs purposely only mention the electrical and timing characteristics of the signal, the negotiation of the signal, and the formats supported.


If you want to do fiber, go ahead. If you want to wireless go ahead. Just make sure you can match the signal characteristics specified.


IMO this gives *more* flexibility than analog over coax. The reason the fiber HDMI is so expensive simply reflects supply and demand. The people who really need very long HDMI installations are really a very small percentage of the overall consumer base.


326ft fiber HDMI seems plenty long to me
http://www.digitalextender.com/Searc...FSAcYAodCC9kKg
 

·
Registered
Joined
·
265 Posts
I haven't read the HDMI spec, but I infer from the fact that (nearly) all implementations today use twisted pair that 1) the circuitry being used today (drivers/receivers etc.) is intended for balanced signals and for the approx. 100 ohm impedance of twisted pair. Driver and receiver circuitry would be different for 75 ohm unbalanced coax.


Nothing wrong with transmitting digital signals over coax, of course.


So to interface to today's existing equipment, a system designed to transmit HDMI over longer distances using coax or fiber would need active converters to interface to existing HDMI-interfaced gear. Extra $$ required.


Also, the article points out that, at 1080p data rates especially, the wavelength of the signals transmitted is so short that differences in cable lengths among the signals become significant. The "run several RG-6's in parallel" approach that AV pros take for installations transmitting analog "component" signals could result in timing problems with high-res HDMI.


I think sfhub and I are in agreement that HDMI is a good thing, for its intended use, whereas the article knocks it for what it is not intended to be.


Perhaps one day we'll see gear (such as sources and displays) built with a fiber HDMI, much as we have digital audio gear with both fiber and coax connections.
 

·
Registered
Joined
·
9,578 Posts

Quote:
Originally Posted by Kenlex /forum/post/0


So to interface to today's existing equipment, a system designed to transmit HDMI over longer distances using coax or fiber would need active converters to interface to existing HDMI-interfaced gear. Extra $$ required.

I think it would be accurate to say that HDMI is designed so it can work over low-cost passive twisted-pair copper effectively up to 15m, however nothing in its specs prevents other physical material to be used for longer distances.


The reason you see so many copper-based twisted pair passive implementation is simply as you stated because of cost. I would say over 95% of the consumers out there have no need for cables longer than 15m (roughly 50ft)


However that is different than assigning blame to the specs for the copper implementations. The specs never forced copper. On the contrary they specifically envisioned the digital signals over fiber as can be seen in the DVI spec references (from an 8-yr old spec) listed at the end.


The TMDS DC-balanced signal, and line supplied power allow for elegant transparent in-line active converters. You can see in this image of a fiber HDMI product that has an in-line converter. Except for the small in-line dongle, you can hardly distinguish it from a standard HDMI cable. This product can reach 326ft.

http://www.digitalextender.com/images/HDOCLgif.gif
http://www.digitalextender.com/detail.aspx?ID=48



If you look at the DVI specs from 8 years ago (which HDMI is based on) they clearly envisioned fiber as a potential physical cable medium:

http://www.ddwg.org/lib/dvi_10.pdf
Quote:
2. Architectural Requirements

2.1 T.M.D.S Overview

...

The transmitter incorporates an advanced coding algorithm to enable T.M.D.S signaling for reduced EMI across copper cables and DC-balancing for data transmission over fiber optic cables. In addition, the advanced coding algorithm enables robust clock recovery at the receiver to achieve high-skew tolerance for driving longer cable lengths as well as shorter low cost cables.
Quote:
2.3.2 Alternate Media

The T.M.D.S. transmission protocol is DC balanced and capable of being transmitted over fiber optic cable. Specific details of a fiber optic implementation are not covered in this specification, but left to the designer.


Fiber optic implementation can be DVI compliant as long as the plug and play ability of the interconnect is still supported. For example, the system must be ab le to read EDID data and detect a hot plug event.


For alternative media to be DVI compliant it is envisioned that the alternate media will serve as a connector to connector adapter.
In case it isn't clear I think we are agreeing on most things about HDMI except the portion where blame is laid on the HDMI *specs* for poor long distance performance of twisted-pair copper implementations. I firmly believe this is an artifact of target market forces dictating price points for solutions, which then naturally led to passive twisted-pair implementations.
 
1 - 20 of 24 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top