Slashdot Mirror
HDMI Brands Don't Matter
adeelarshad82 writes "I'm sure most of us looking for an HDMI cable have been in a situation where a store clerk sidles up, offers to help and points to some of the most expensive HDMI cables — because apparently these are 'superior cables' which we all absolutely need for the best possible home theater experience. Well, as it turns out the claims are, for the vast majority of home theater users, utter rubbish. According to tests ran on five different HDMI cables, ranging in price from less than $5 up to more than $100, HDMI brands really don't matter."
For short cable runs, any old HDMI cable will do. When you get into the 50-100 ft lengths, the cable quality absolutely matters.
HDMI signals may be digital, so there's none of the subjective analog concerns, but it's also a real-time signal, which makes it susceptible to even small delays in transmission across the cable. This isn't a concern in a sub-20 ft cable, but becomes noticeable in the cheap longer cables.
The only difference between cables that really matters is dispersion (frequency-dependent losses.) A difference of 1 dB/meter in loss between cables is going to make quite a difference at 30 meters. However, I wouldn't bet one way or the other on which brands have better or worse loss characteristics.
Lacking <sarcasm> tags,
This is my favorite cable ever. Denon gets it - idiots want to give their money away, why not make it easy for them?
I think the point is more: Either the signal on the cable has distinguishable 0s and 1s, or it doesn't. Which is absolutely how it works. Over long distances, you might have some interference - that EVENTUALLY will lead to a 0 and a 1 not being that different anymore, or at least severely corrupted - but frankly it's entirely different than it was with analog cables, because so long as 0 and 1 are different enough for a given situation - it doesn't matter HOW DIFFERENT they are. Making them MORE different does NOT improve signal quality. Therefore - with HDMI, or any other digital cable - you should buy a cable that is the cheapest that will do the job. A more expensive cable will not improve signal in the situation where the cheaper cable works. A more expensive cable might have better connectors - by which I do NOT mean "plated with gold" I mean "designed in such a way that the cable does not fall apart on repeated unplugging and plugging back in" - so if that's a common use case, by all means, factor that in. If you are in the 0.00001% case where you absolutely need more sheilding around your cable, because there is just SO MUCH damned inteference, or because your cable run just HAS to be 200 un-amped feet over copper... well then, buy the more expensive cable. But there are $300 6-foot HDMI cables out there, with "features" that don't matter one damn, and nobody should be buying them.
Actually, gold plating decreases signal quality (by a tiny bit). The thing is that when current flows over changes in conductor material, noise is added. With gold, you usually have other material below, as copper diffuses though gold layered directly on it. So copper-nickel-gold---gold-nickel-copper is actually worse than copper-nickel---nickel-copper. One of the dirty secrets of audio contacts. Not that you could hear the difference.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
The point of the article being that many people think that a better cable will give them higher quality video... It won't... It'll give them video at all if a lower quality cable will fail. Not only that, but given that all HDMI cable is required to meet a spec, unless the consumer is doing something out of spec (very very rare), all HDMI cables, including the $1 ones will give them a signal.
(i promise i'm not a shill)
monoprice.com
they do an awesome job of getting any type of cable i need, at an awesome price.
Here's the link to the specifications:
From the WIKI:
What really ruins Cat5 is how good the shielding wire is. You can get good, cheap cat 5 cable. But if the shielding wire is 28-34ga while the main conductor is 24ga you get what you pay for. That also really makes a difference when they 'stretch' out the twists for shielding. Both twist, and gauge count matter in Cat5.
Om, nomnomnom...
You do realize that the vast majority of ethernet cables are unshielded right? And that the shielding actually decreases performance measurably?
What are you talking about? The vast majority of Cat 5 is unshielded. On the off chance it does have shielding it's usually foil. The main noise rejection strategy with twisted pair is running a balanced signal, thus the requirement for signal lines in pairs.
I used to think the same thing about Ethernet cables. it's all digital right? And yet I've seen speeds increase 10 fold when replacing old one.
Define "old one". Are you referring to a 10/100 cable that does not support gigabit? I've seen quite a few such cables (heck, even being sold today) that only have two pairs and will not negotiate at gigabit speeds. I recently replaced a few at a client's office actually, and they were installed only 5 or so years ago.
I've even found some that have all four pairs - but only two actually crimped into the connector (the other two pairs simply terminate in the plug, uncrimped). Again, no gigabit speeds there.
I make my own ethernet cables from boxed wire bought at Home Depot or Lowes (Cat 5e - or occasionally Cat 6), and they all (even at various lengths, up to and including a few 100+ foot runs) perform just as well as any of the name brand, uber-expensive cables we've got lying around here. Oh - and I'm knowledgeable enough to actually check for things like retransmissions, "collisions" (ie: apparent ones due to echo/crosstalk, as switches shouldnt have such an issue), errors, etc. I can most definitely tell you, that unless you try very very hard to buy a crap cable, the results are generally within the norm regardless of price.
Now, if you MAKE cables, that's different. (1) I've found ones poorly crimped, (2) ones where the pairs have been unwound for feet, (3) ones where they used aluminum core wires for speeds such are not rated for, (4) ones where the wire gauge is not to spec (ie: smaller than it should be), (5) ones where the insulation is stripped off the wire before it's inserted into the plug, and so on.
But that's not a flaw in cheap cables - it's a flaw in having someone who doesn't know what they are doing making cables.
StarTrekPhase2 - The Five Year Mission Continues!
You can run HDMI over Cat-5 sometimes, depending on the resolution. The thing with HDMI is the bandwidth needed varies with the signal resolution. If you just want to do 1280x720@24Hz the bandwidth (in terms of digital bits) you need is very low and thus the cable bandwidth (in terms of analogue frequency) is also very low. If you want to do 1920x1080@120Hz it is much higher.
It also depends on how noisy your environment is. Your example with power cables is a bad one since that is too low frequency to matter to HDMI. However if you have noise in the 100s of MHz, that is the range of the signal over the cable and thus interference can happen if the run is too long, or if the shield is bad (or non existent as in your case).
So for consumers the easy guide to follow is just to check the cable's certification. Any cable worth buying will tell you if it is certified standard speed or high speed. Standard speed is a certification for 720p or 1080i, high speed is for 1080p. If you get a cable that is certified to the speed you need, you are good to go. All the cables from cheap places like Monoprice are.
Now the certifications are overkill, as is usually the case with this stuff. You'll find that you can usually get a longer "standard speed" cable and run 1080p over it no problem. However the reason for the overkill certifications is that it'll work in more or less any conditions. The farther you go out of the spec, the more likely a problem is.
Same deal with Ethernet. If you try it, you discover that you can indeed have cable runs over 100 meters, sometimes WAY over. Thing is, sometimes you'll have problems if you try. 100 meters is the "going to work almost no matter what" spec.
Thus "just follow the spec" is my advice for regular users. High speed HDMI cables are cheap as hell from Monoprice and you just won't have any trouble.
So upsampling is basically just using a technology that is better than nearest neighbour to increase resolution. DVDs are 720x480, and a full HD LCD is 1920x1080. Obviously you have to deal with that difference. If you just stretch the pixels that works fine, but doesn't look that good. What you can do is use more advanced math to try and make the upsampling look better.
A simple example would be bicubic interpolation. You can find that in most 2D graphics programs like Photoshop. Try taking something and playing it up with nearest neighbour, and then with bicubic. While it isn't magic, bicubic looks much better.
For more advanced examples look at 2xSal. hq4x and the like (http://en.wikipedia.org/wiki/Pixel_art_scaling_algorithms). These ones are designed for pixel art for things like old video games, but it shows you what I'm talking about. The result is much better than things scaled with straight pixel duplication.
In terms of the specifics for video upsampling, well it varies based on the chip used to do it, and it is all proprietary. They won't release the details. However the idea is the same. They use various algorithms to look at a frame (and sometimes data from surrounding frames) to do a more intelligent upsample.
The result is pretty good when done well. It is amazing how good an upsampled DVD can look. Not as clear as something actually shot in HD, of course, but not bad.
There's another possibility. Look at the history (click on the actual score). There's a good chance they already had 2 negative mods and 2 +1 informative (thus showing 0). When you add +1 funny, they get a final score of +1 and "informative" has more votes than "funny" and so takes the lead.