HDMI over Ethernet, Ethernet over HDMI?

example of a cable that supports ethernet and long distances

http://www.monoprice.com/products/subdepartment.asp?c_id=102&cp_id=10255#1025501

example of a HDMI extender for non redmere cables that uses ethernet cables

http://www.monoprice.com/products/product.asp?c_id=109&cp_id=10105&cs_id=1011012&p_id=8121&seq=1&format=2

example of non redmere HDMI cables that support ethernet over hdmi

http://www.monoprice.com/products/subdepartment.asp?c_id=102&cp_id=10250

I think folk are mixing up technologies and capabilities here!!!

HDMI Ethernet Channel (HEC) – must have seemed a good idea at the time but as no hardware manufacture I’ve come across has so far dared to try and implement it it’s a complete Dodo, http://www.hdmi.org/manufacturer/hdmi_1_4/hec.aspx

HDMI over Ethernet – utilises dynamic compression to allow you to pass HDMI signals over a single CAT cable via a managed switch (if you want multiple sources). Not sure it would be my first choice for my home AV network as no matter what some folk say the image has to be compressed to make this work.

HDMI over HDBaseT – utilise the Valens HDBaseT or HDBaseT-lite Transmitter and Receiver chip set to extend HDMI (uncompressed) out to 100m or 60m and by virtue of the Transmitter chip ‘packetizing’ the HDMI signal you can (note can) run 10/100 down the same wire, though many folk are shying clear of doing that and preferring to run a separate cable for an IP network. You wouldn’t want the HDBaseT 10/100 limitation becoming a choke point on your IP network!

RedMere – passive cables which steal power from the Source end of the HDMI cable run to power equalizer circuitry at the Sink end of the cable run! Seems to be pretty stable but ‘future proofed’ – not if the equalizer circuitry fails to talk to your Source or Sink or fails because of some new ‘feature’ in whatever comes next with HDMI!

Conduit – your best option if you want 100% future compatibility!

Joe

Cheers otto - don’t mind if you steal it

I’m sure some of the regulars will chip in too.

Joe

Joe,

You sure about that? From what I have been reading recently, Redmere steals power from the sink end because it requires more current than can safely be taken from the 5V line.

EDIT: the chip is on the display end and power is drawn from the source
Edited by PlasmaPZ80U - 1/10/13 at 11:29pm

? I didn't think there was power from the display.... I thought the powering was the same powering used to "read" display capability by the source

The 5V line is powered by the source. There is no equivalent powered by the sink. But apparently Redmere, in addition to using the 5V line for some power, is using power harvesting technques to steal power from signal lines, some of which originate on the sink.

so it's using power from both?

Apparently, in the current chip sets, anyway. The problem Redmere has to overcome is the same as any HDMI device that relies on the 5V line for its power. The 5V line is there for hot plug detect, not to power devices. The HDMI specification only requires the source to supply 55mA. Some can supply more, but there is no guarantee. Also, the fact that Redmere cables tend to use smaller wire than passive cables, sometimes as small as 36 AWG, doesn't help any because of the increased voltage drop. A patent licensed to Redmere mentions harvesting power from at least one differential line and one control line, IIRC CEC.
Edited by Colm - 1/10/13 at 11:57pm

Are these Redmere cables the only ones with these 'chips' embedded in the connector? Assuming this chip is an amp, why would it be placed at the far end not the source? Isn't that the same as placing a antenna preamp next to the TV vs next to the antenna?

Because it isn't an amp therefore when it restores the signal, it is done on the sink side. Amps are ineffective with long HDMI cable runs.

From Colm,
"The Redmere site only says it does equalization and skew correction on the TMDS lines..."

http://www.avsforum.com/t/1443594/do-monoprice-redmere-cables-boost-the-signal-for-the-audio-return-channel

This "chip" suppose to be just error correction?

There is more than one brand of cable with Redmere chips. The problem is not the signal level, but the shape and timing of the signals. The Redmere chip deals with the latter issues. There are a number of detrimental things that happen to the signal. The high frequency components of the square waves attenuate faster than the lower frequency components. The net effect is to round the shoulders of the square wave. And skew develops between the signals on the differential pairs, and between pairs. All of this can potentially affect how the signal is interpreted on the sink end.

It is not error correction, at least in the digital sense. If somehow a bit flips, it stays flipped.

FWIW one could deal with the attenuation of high frequency components of the signal by using pre-emphasis. In fact, some devices (e.g. some switches, matrixes, etc.) do this. It has its own problem. It is hard to get the value right, and it has the potential to increase cross talk. But even if pre-emphasis works properly that still leaves the problem of skew, which can only be handled at the sink end.

IMHO the real advantage of Redmere is that the current chips are programmable at the factory to match the characteristics of the specific cable.
Edited by Colm - 1/11/13 at 1:42pm

Ok, Redmere is the name of the chip, not the cable. I though ti was a name either Monoprice or the importer gave to those cables.

Excellent explanation, thanks. That all makes sense. The 'rounding' of the square waves is the biggest problem, or is it skew, or just both?

RedMere is an Irish company who supply chip modules or supply complete OEM ‘Active’ HDMI cable assemblies to many companies - http://www.redmere.com

The Products page gives you an overview of the various Chips they have on offer - http://en.redmere.com/business/redmere_products.php

Joe

Interesting. Thanks.

They have two related 'White papers' on HDMI including photos of these chips in the connectors.

The biggest problem facing HDMI is skew. A high speed cable is allowed a maximum of only 111 picoseconds (trillionths of a second) intrapair skew. The longer the cable, the greater the skew because absolute differences in wire length within, or between, pairs increases. That is the primary reason that you don't see high speed cables longer than about 25'. The rouning contributes to the problem because it essentially obscures where the edge is. There are other problems besides skew and the rounding of the square waves, like impedance mismatches that cause reflections, crosstalk, etc. They aIl take their toll. It is quite a credit to the chip manufacturers that they can make HDMI work at as high a bit rate and distance as it does.

When we tested 1394b cables we noticed reflections as our biggest problem. It was interesting because once you started to buildup bit errors, in some cases if you went longer in cable length, the bit errors went away.


HDMI is a different beast where you have parallel serial streams and so keeping them together is a challenge, as you said.

If gauge and/or shielding wasn't a limitation (size & space), would that make a difference?

It's kind-of like asking if friction wasn't involved would a car get better MPG. Well yes, but not with the physics we currently understand (and you would have other issues such as stopping).

If you didn't have loss with distance then the gauge wouldn't matter and theoretically since there would be no reduction in signal over distance you would have the same amount of skew at the end of the cable as at the beginning (again with no loss).

No interference would just mean that the signal didn't have noise added over any distance.

But, again, unless you go to superconductors I don't see how any of those could happen.