Difference in HDMI cables besides the price?

[Constant Gardener]

Gee. My Oppo 971H included an inexpensive DVI-to-HDMI cable that works flawlessly. Has anybody had a bad experience with a digital connection?

[BJ_M]

no -- 2000$ for a power cable is what is even (more) stupid -- and there are several of those out there

[gshelley61]

There's a sucker born every minute.

[oldscubaguy]

FYI -
The physical interface of a type A HDMI is actally 3 separate differntial pairs for the Video and Audio data (TMDS Data 0:2), 1 diffential pair for the clock, 1 I2C channel (DDC) for communication between source and host, 1 CEC line, 1 Hot plug detect line, and 1 line with +5VDC.

For any of the data channels or the clock channel:

The nominal high-level voltage of the signal is AVcc and the nominal low-level voltage of the signal is (AVcc - Vswing). Since the swing is differential on the pair, the net signal on the pair has a swing twice that of the single-ended signal, or 2Vswing.
The differential signal swings between positive Vswing and negative Vswing.

This is interpreted as a 1 or a 0 by the reciever. Since there are 3 data channels there are 3 bits of video data for each clock cycle.

So if any of the data channels or the clock signal is corrupted you can have problems.

With that being said - A differential pair is very robust, has built in noise immunity since any noise that occures usually is canceled out. The problem comes on long runs when the voltage difference between the + and the - of the pair is not large enough for the reciever to detect. That is why they reccomend larger guage cables for long runs.

Something else to keep in mind:

All cable used for HDMI is made in China, in fact only the manufacturers that make custom cable lengths actually assemble them outside of China. Stock length cables are all made in china utilizing the pretty much the same cable - Including Monster.

For short cable runs (15 feet or less) there is no difference. Just make sure that the HDMI cable you purchase states the following:

A) Supports data rates to 5Gbps

B) HDMI™ and HDCP Compliant.
- this means that the design was tested by a certified tesing facility.

C) That it is at least 28AWG on the data and clock lines.

The HDMI cables I use are:

http://www.beachaudio.com/product_info.php?products_id=45295&prodid=516-806bk

The Poop sheet for this cable can be found:

http://www.sterenusa.com/bssd/news/NPR_PDF/516-806%20HDMI%20Cables.pdf

Have had no problems, and the PQ on my Sony KDL-40XBR2 looks perfect

[edDV]

FYI - ...

So if any of the data channels or the clock signal is corrupted you can have problems.

...The problem comes on long runs when the voltage difference between the + and the - of the pair is not large enough for the reciever to detect. That is why they reccomend larger guage cables for long runs.

Something else to keep in mind:

All cable used for HDMI is made in China, in fact only the manufacturers that make custom cable lengths actually assemble them outside of China. Stock length cables are all made in china utilizing the pretty much the same cable - Including Monster.

For short cable runs (15 feet or less) there is no difference. Just make sure that the HDMI cable you purchase states the following:

A) Supports data rates to 5Gbps

B) HDMI™ and HDCP Compliant.
- this means that the design was tested by a certified tesing facility.

C) That it is at least 28AWG on the data and clock lines.

So have you seen a failure trend for anything under 15 ft.?

The technology is essentially the same as DVI-D but with additional audio and bi-directional control lines.

I'm secretly hoping there is a massive problem with these cables when they turn on HDCP.

Ooops, no longer a secret.

[zeusdog2006]

As others have stated Monoprice is the best supplier of cables that I have found.

Also consider this. Component cable produces a much better picture then the current HDMI version, 1.3. Don't be fooled by the hype.

[edDV]

As others have stated Monoprice is the best supplier of cables that I have found.

Also consider this. Component cable produces a much better picture then the current HDMI version, 1.3. Don't be fooled by the hype.

So long as the rest of the analog path has quality.

[oldscubaguy]

The only failures I've seen are connection problems.
Because HDMI does not have any fastening like DVI it can pull loose especially the heavy cables that have EMI Ferrite clamps.
But I have heard rumblings of them comming up with some fastening system for future HDMI.

As far as Component inputs giving better PQ, that has not been my experience. the component transmission is:
Pixel from the MPEG decoder - converted to analog via video DENC - to 75 ohm driver - to coax cable - to buffer in display device - converted back to pixel data using an A/D.

As apposed to DVI or HDMI where the pixel data is directly transmitted serially via 3 data channels.

Seems to me you have more chance of screwing up the Pixel data with component than with DVI and HDMI.

But who knows what problems the HDCP will introduce when it kicks in.

[edDV]

Much of the production infrastructure is using analog component in segments but with high quality devices and mucho performance testing. HDMI is an attempt to make it cheaper and more foolproof. Quality is good enough for consumer equipment.

[oldscubaguy]

But an MPEG decoder puts out Pixel data not an Analog representation. And the HD motitors whether they be LCD, Plasma, DLP, or LCOS use display chips that take in Pixel Data.

If you are talking about Analog Video to a CRT, that is a different ball game.

But these days even the SD is comming from an STB or a DVD player that utilizes MPEG Decoders, so in my opinion (which is only that) best to keep the data in a physically digital format so there are no conversion errors - as there are with all Video DENCs and A/Ds.

[edDV]

But an MPEG decoder puts out Pixel data not an Analog representation. And the HD motitors whether they be LCD, Plasma, DLP, or LCOS use display chips that take in Pixel Data.

If you are talking about Analog Video to a CRT, that is a different ball game.

But these days even the SD is comming from an STB or a DVD player that utilizes MPEG Decoders, so in my opinion (which is only that) best to keep the data in a physically digital format so there are no conversion errors - as there are with all Video DENCs and A/Ds.

I'm not arguing the trend to digital replacement. Both can do the job but digital communication is ultimately cheaper to manufacture and maintain. But don't fool yourself, at the engineering level, HDMI design is an analog RF design problem. Once the analog transmission problems are solved, digital data errors become nominal.

[oldscubaguy]

I'm curious, How is a differential pair transmission an RF design problem?
If that is the case, Ethernet would also.
If you look at the data sheets for differntial drivers and recievers you will see that there is no modulation involved.

I agree that when you get into high clock frequencies and low voltages, you can become more suceptible to external noise, wire capacitance, inductance and resistance. But I personally have seen 3 Gigabit Ethernet over cat 6 cable.

[edDV]

I'm curious, How is a differential pair transmission an RF design problem?
If that is the case, Ethernet would also.
If you look at the data sheets for differntial drivers and recievers you will see that there is no modulation involved.

I agree that when you get into high clock frequencies and low voltages, you can become more suceptible to external noise, wire capacitance, inductance and resistance. But I personally have seen 3 Gigabit Ethernet over cat 6 cable.

At the base level any transmission at GigaHertz frequencies is an RF design issue. You are insulated from this reality by the component manufactures and standard setting organizations ... that is until data errors manifest.

Ethernet is an analog technology in most respects. Data communication is the application.

[Constant Gardener]

Now let me get this right. We're watching analog video distributed as a digital signal atop an analog medium which is in a macroscaled quantum digital world - which may be based on tiny wee strings in an eleven-dimensional analog.... thread drift alert...

[oldscubaguy]

I guess in the strictest of forms because the ethernet PHY is connected to the cable via a transfomer it is analog.

But component goes through a digital to analog conversion process on the transmission end and then an analog to digital conversion process on the recieve end. in this process the Pixel data values are never exactly the same as the original due to DENC and A/D due to the resolution, accuracy, and nonlinearity of the parts used.

On a short run of DVI or HDMI you have more accurate pixel data.

The best solution in my opinion would be to have the MPEG Decoder in the Monitor, but that would be restrictive.

[oldscubaguy]

Come to think of it My Sony does have an MPEG decoder for the off air 8VSB. Too bad the Satellite, Cable, and DVD boxes don't have ASI output that we could feed directly into my Sonys decoder. Then we could have ASI over multimode fiber and not worry about loss.

[edDV]

Complex isn't it.

The A/D D/A issue is real depending on performance but you need to look at that in the context of the rest of the processing and if it makes a difference at the final display.

Stepping back in the process to the cable system or up/downlink you are in analog territory and at the mercy of their process. Take my word for it, Comcast or DirecTV have no loyalty to pixel by pixel transmission.

In other words. You worry about purity at the last stage in the process when your source is far from a virgin raster.

<-- I'm just trying to stimulate some conversation.

...Then we could have ASI over multimode fiber and not worry about loss.

huh? you have a fiber from your house to the production studio?

[oldscubaguy]

No, but I wish!

I was talking about MPEG transport directly from the STB so the decode process of the MPEG is done in the TV.

But I agree - we are at the mercy of the broadcasters and their budgets! I have seen HD from the major broadcasters (NBC, ABC, et all) that looks terrible because they are compressing it way too much. Were talking MPEG 2 HD at encode rates of 12Mb/s, when it should be at least 15Mb/s. I guess they think we can't see motion artifacts!

FYI - Verizon has Fiber to the home (FIOS) but not in my area.
But even then you are at the mercy of the original encoding process.

[oldscubaguy]

Check it out:

http://www22.verizon.com/content/fiostv

[edDV]

Now let me get this right. We're watching analog video distributed as a digital signal atop an analog medium which is in a macroscaled quantum digital world - which may be based on tiny wee strings in an eleven-dimensional analog.... thread drift alert...

close but no cigar.

Take HBO though cable - HD path,

1. Movie is transferred to Digital Betacam 480i, HDCAM 1080i or (future) Digital Cinema 4kx2k (and up)
2. For HD*, HDCAM (3:1:1 1080i 144Mb/s native) is uplinked over analog modulation to the local head end at ~ 25 Mb/s.
3. Local cable remux up to 25 Mb/s into analog QAM (analog RF) modulation for local distribution to cable box.
4. Cable box creates from QAM a YPbPr (720p or 1080i) or DVI/HDMI (720p or 1080i)
In both cases, 720p is interpolated and is not 59.94 fps progressive except maybe for ESPN.


* I'm being nice to HBO here because much of what I see is upscaled 480i.

[edDV]

Check it out:

http://www22.verizon.com/content/fiostv

Verizon has very few customers connected for fiostv so far. IPTV has its own set of limitations. Just where is the server bandwidth to supply unique VOD? How many customized streams will each home get? Why would they supply significantly higher bandwidth than the other guys?

[oldscubaguy]

That doesn't even include the changes taking place in the MPEG encode and decode process.

The MPEG encode compresses the original content by interpreting blocks of picture data and then predicting others based on the iterpretation. Depending on the desired output data rate and how may passes the encoder does can give an idea of how much of the true picture is left.

Then after it gets to the cable head end - the cable company has been known to compress it even further with a transcode process so they can fit more streams in a single QAM channel.

And I don't even want to talk about how compressed the SD is over cable and satellite -- too depressing.

In other words, if you have the Best STB and the Best HDTV, and cables made by God - you will still have artifacts.

[edDV]

Right on bro.

[oldscubaguy]

Check it out:

http://www22.verizon.com/content/fiostv

Verizon has very few customers connected for fiostv so far. IPTV has its own set of limitations. Just where is the server bandwidth to supply unique VOD? How many customized streams will each home get? Why would they supply significantly higher bandwidth than the other guys?

I was mainly looking at FIOS for my internet (blazing download and upload speeds).
As far as I know about FIOS (which is little) I believe they will be working in a multicast format. In other words they will only be transmitting what is being watched at that time. If a IP STB requests a channel that is not being streamed at that time to any other IP STB, then it will transmit that channel. If an IP STB is requesting a channel that is already being transmitted to another IP STB then it will simply join that multicast.

This saves them bandwidth because they are not transmitting all the channels all the time Like Video Furnace.

As far as how they will handle their VOD movies, I imagine the same as the cable companies do.

This is only my speculation not fact --- More will be revealed when they really roll it out I imagine.

[oldscubaguy]

Sorry, ment to say:
" Like Video Furnace This saves them bandwidth because they are not transmitting all the channels all the time "

[edDV]

The trend has to be to higher compression for cable and satellite. New set top boxes will be required.

I think ATSC is going to h.264/VC-1 for all but the primary MPeg2 channel. FCC rules say only the primary channel needs to be MPeg2. That means those buying TV sets with ATSC tuners today will only get the primary channel. A new tuner would be required to get h.264 or VC-1 secondary channels.

[edDV]

If a IP STB requests a channel that is not being streamed at that time to any other IP STB, then it will transmit that channel. If an IP STB is requesting a channel that is already being transmitted to another IP STB then it will simply join that multicast.

This saves them bandwidth because they are not transmitting all the channels all the time Like Video Furnace.

"simply join that multicast" means joining in progress.

The trick for cutting bandwidth on the "top ten Movies" is to have a very large local cache either in the local node or in the set top box that can play these elections without swamping the larger data path.

[oldscubaguy]

Considering the insanity that is h.264 right now, encoders and decoder utilizing some and not all of the toolsets, and few of them matching, I think we have a while to see widespread h.264 over 8VSB. Especially since most of them in the LA DMA spent all their money on their MPEG 2 systems. I was talking to an engineer for one of the broadcasters out here and he said that they are still trying to get the consumer informed enough to utilize their free 8vsb content.

[oldscubaguy]

[/quote]

"simply join that multicast" means joining in progress.
[/quote]

yes, as in let's say the SCI-FI channel broadcast.

[/quote]
The trick for cutting bandwidth on the "top ten Movies" is to have a very large local cache either in the local node or in the box that can play these elections without swamping the larger data path.[/quote]

Or Servers at nodes

[edDV]

h.264 or VC-1 are not ready for primetime. DirecTV and Dish went with proprietary MPeg4 so that it can be controlled and fixed without international committees.

The marketing of 8VSB (ATSC) secondary channels has been very poor. Most are amazed that this stuff is available OTA.