[quote]Bergman wrote:
The question then becomes how often packets get lost in a corporate network. I’m willing to bet close to zero, but I have no idea just how close to zero.[/quote]
Having tested many network cables in my days, with equipment where you put an emitter at one end and a receiver at the other (you can even try it at home if you have two computers: http://www.pcausa.com/Utilities/pcattcp.htm … you’ll have to test a crossover cable though.), I have never personally seen a cable that dropped “some” packets.
I’ve seen no packets (bad cable or, more often, bad end plug) but I’ve never seen a cable where only 95% of the packets made it through. That includes many cables where the length was way out of spec (which led to other problems, but dropped packets on the tester wasn’t one of them).
As for HDMI not running of TCP/IP, you’re right. Although I’d suspect that the spec (I haven’t read it, but that’s the way I’d bet) calls for some type of error detection and correction, if only because it supports encryption of the signal and all the “good” encryption methods I know will instantly turn to shit if you lose even 1 bit.
I’d also bet that the encryption is done by blocks of data so that any error only loses 1 block and does not corrupt the rest of the stream.
Yup. I’d suggest to everyone who thinks they can hear (or see) the difference between a $10 cable and a $200 (or $43,000) one to set up an ABX test (ABX test - Wikipedia) and really find out.
Maybe an oscilloscope can tell the difference between the best Monster Cable and the same length of zip cord, but as far as I know, no human can (ABX Double Blind Tests: Interconnects and Wires).
Unless you enjoy watching your music on an oscilloscope, save your money.
I’ve seen packet loss on Cat 5e cable a number of times. It’s almost always a bad termination in the RJ45 connector.
For some funny reading, check out the $1,000,000 challenge by James Randi against those audiophiles who claim the $7,000 (Pear Anjou) audio cables are better.
I consider myself educated then, though I guess I was trying to say garbage packets, whereby only a single bit was messed up. Entirely dropped packets only really becomes an issue at the network layer.
I think it would be pretty silly for HDMI to run off TCP/IP =p There’s no need for it to do any sort of routing.
You’re right in that there is error detection/correction (BCH ECC) built into the spec. However, if that fails, there is no way to request retransmission of the packet. But it’s a real-time stream, so think of UDP - why would you even want to request retransmission?
[quote]Bergman wrote:
You’re right in that there is error detection/correction (BCH ECC) built into the spec. However, if that fails, there is no way to request retransmission of the packet. But it’s a real-time stream, so think of UDP - why would you even want to request retransmission?[/quote]
I wouldn’t. I think we’re getting sidetracked a bit. The issue - on which we agree - is that you don’t need high priced cables for most signals and especially not for a digital one. The cheapest cable that meets the design specs will do just fine.