What Do HDMI Spec Versions (1.2, 1.3, 1.3a, etc) Mean For Cable Choice?

Many people are worried, when buying HDMI cable, about the spec version of the cable in question. Is it the latest spec version? Will it support all the features of the devices it's hooked to? Most of the fears and doubts associated with this question are unfounded, but there are some interesting issues with regard to HDMI spec version, and we'll sort those out here.

What Are The Spec Versions?

At this writing, March 2009, there have been five versions of the principal HDMI spec document, and a new update (probably to be called 1.4 or 1.3b) has been announced but not yet released. Additionally, however, some references to specs are not to the principal HDMI spec document, but to the "Compliance Testing Specification" document, and there have been six versions of that document. The versions of the two, to date, are:

• HDMI Specifications:
  • HDMI Specification 1.1
  • HDMI Specification 1.2
  • HDMI Specification 1.2a
  • HDMI Specification 1.3
  • HDMI Specification 1.3a
• HDMI Compliance Testing Specifications:
  • Compliance Test Specification 1.1
  • Compliance Test Specification 1.2a
  • Compliance Test Specification 1.3a
  • Compliance Test Specification 1.3b
  • Compliance Test Specification 1.3b1
  • Compliance Test Specification 1.3c

The HDMI Licensing authority allows anyone to download a copy of the current version of the main HDMI spec document from its site at hdmi.org. For some reason, however, HDMI, LLC does not authorize release of the Compliance Test Specification to the general public. That's unfortunate, because it's in the CTS that the details of compliance testing are contained, and there can be some significance to that.

There is some room for confusion here, and unfortunately, HDMI cable vendors are often just as confused as everyone else--it's common to see cable labeled "1.3b compliant," when there is no such thing as HDMI 1.3b, and we are frequently asked whether our cables are 1.3a or 1.3b compliant. But while there is no "1.3b" version of the HDMI spec, the Compliance Test Spec does have a 1.3b version and a 1.3b1 version, as well as a 1.3c version (testing for cables was unaffected by these minor CTS revisions, except that 1.3c adds new criteria for "active" cables). Our Belden HDMI cables have been tested under main spec version 1.3a, CTS version 1.3b (and, in the case of our Series-1 Category 1 approval, CTS version 1.3b1).

What Hasn't Changed

With all these spec versions around, and equipment and cables having been designed and tested in accordance with one version or another of the spec, it's natural that people worry about the compatibility of equipment and cables. Device compatibility is a complicated subject and is really outside our area of expertise, but when it comes to cables, we can give some reassurance. The basic characteristics of HDMI cables have not changed from one spec version to another. They have the same conductors, in the same configuration, wired to the connector in the same manner.

What this means is that, in terms of supporting features and protocols (e.g., HDCP), all HDMI cables have been designed to the same basic standard, whether they were designed under 1.1 or 1.3a or anything in between, and whether they were tested under CTS 1.1, 1.3c, or anything in between. A 1.1 certified cable may not be officially (or actually) 1.3 "compliant" in the sense of having passed testing under the 1.3 standards, but it will always be 1.3 compatible, in the sense that it is built to handle the same job, and has all the necessary parts, for 1.3.

That might seem obvious, but for the fact that certain cable vendors seem determined to confuse the issue. We see quite a few cables marketed on the basis that they support "deep color," or new colorspace protocols, or new audio formats, when in fact the old cables support these features just as well. If your vendor is promoting a cable as supporting new features and protocols, run, don't walk, and find yourself a different vendor.

Likewise, the connector design has not changed, at least for the vast majority of HDMI applications. All ordinarily rack-mount or shelf-sitting HDMI gear on the market uses, and will for the foreseeable future use, the HDMI "type A" 19-pin connector. However, two miniature versions of the connector may also be seen (one, type C, is already available on a few miniature devices, and another is expected in the new spec) occasionally. The spec also provides--and has provided, from its earliest versions--for a dual-link "type B" connector, but to date that connector has not appeared on anything that's actually come to market.

HDCP: a Cable Non-Issue:

You'll also see, sometimes, references to gear being "HDCP compliant," meaning that it supports copy-protection features. This can be meaningful where source, display and repeater devices are concerned, but it has nothing whatsoever to do with cables. All HDMI cables are HDCP compliant, because the signals which constitute HDCP run on the same conductors that carry the other parts of the HDMI signal. A cable incapable of carrying HDCP would be incapable of carrying any signal at all.

So, What Has Changed?

Plenty of things have indeed changed from one version to another of the spec documents, but only a handful of them have to do with cables; and of those which have to do with cables, quite a few of the changes are minor, insignificant things--nothing affecting performance in any substantial way. But there are some changes to the requirements expected of cables, and to the way that compliance with those requirements is tested, and these can be meaningful. These changes happened at the switch from main spec version 1.2a and 1.3, so there really are, for practical purposes, only two sorts of HDMI spec compliance for cables: 1.2a and before, and 1.3 and after.

"Specified Bandwidth"

In spec versions 1.1 to 1.2a, a manufacturer submitting a cable for testing was entitled to declare a stated bandwidth or "maximum frequency" for the cable. This stated maximum frequency would be given by the manufacturer in MHz, representing the clock speed, and the actual bitrate it represents would be ten times that--so, for example, 74.25 MHz, which represents 720p or 1080i, would equate to 742.5Mbps. This had a substantial impact upon the cable's ability to pass compliance testing, as we'll explain.

Under all versions of the spec, a cable submitted for testing is subjected to two separate testing regimens, and if it passes either, it is deemed compliant even if it fails the other. The first of these is what's called an "eye-pattern" test, where an actual signal representing a "worst-case" source device output is fed through the cable, and the output at the other end of the cable is measured graphically on an oscilloscope. The second is a series of "parametric" tests where various attributes of the cable are measured, such as skew, crosstalk and attenuation. A significant difference between these types of testing, under 1.1 through 1.2a of both the main spec and the Compliance Testing Spec was that the eye-pattern test was run at a frequency determined by the cable's stated bandwidth, while the parametric tests, to the extent they depend upon frequency, were run at frequencies which were fixed and which did not depend upon the cable's stated bandwidth.

Differences in bitrate make a tremendous difference to eye-pattern results. Consequently, a cable could be tested at 480i or 480p, and could be certified at a much longer length than would have passed spec at 720p or 1080i because the bitrate is just over a third of the 720p/1080i bitrate. But a vendor did not necessarily need to share with the consumer the information concerning the bandwidth under which testing was passed, and consequently, one could not know what a certification meant. What length was certified as HDMI spec compliant, and at what bandwidth?

Under spec version 1.3 and later, this ambiguity is much reduced. A manufacturer can still submit any length of cable he wants, and he can still specify the bandwidth, but he has only two choices in bandwidth, which are "Category 1" and "Category 2." Under Category 1, the cable must either pass the eye-pattern test at 74.25MHz (742.5Mbps) or the parametric tests; under Category 2, the cable must pass eye-pattern tests at 165MHz and 340MHz (the latter test being aided somewhat by application of some software equalization to the output signal) or pass a stricter set of parametric tests. If you know the length of the cable and the testing Category, you have a pretty good idea what the result means. The best way to be sure of these matters is to ask your vendor for a copy of his HDMI Compliance Testing Certificate--unfortunately, many vendors either are not completely honest, or are themselves misinformed, about what tests their cables have passed, and at what distances.

Eye Pattern Test Strictness

The other substantial change which came about in the 1.2 to 1.3 changeover had to do with the conditions under which the eye-pattern testing is run. This change is, oddly, not reflected anywhere in the main spec document, and consequently, seems to be known almost to no one except those who do the testing. Instead, the change is buried deep in the details of the description of the Compliance Test Specification procedure for running eye-pattern tests.

As we've mentioned above, an eye-pattern test begins with a test signal being fed into a cable, and the result is read with a high-frequency oscilloscope which samples the waveform as it comes out of the other end of the cable. Needless to say, the quality and amplitude of the signal coming out of the cable is dependent on the quality and amplitude of the signal that goes into the cable. Under CTS 1.2a and before, the "swing" voltage of the test signal was 500 millivolts--that is, a "1" was signalled by a voltage of +500mv, and a "0" was signalled by a voltage of -500mv. Under CTS 1.3a and after, the swing voltage is 400mv. In other words, the signal is about 20% (about 1 dB) weaker when it goes into the cable, and it will come out about 20% weaker. But while the signal going into the cable "shrank" by 20%, the required output signal spec did not change at all. A cable that marginally passed the eye-pattern test under 1.2 will fail under 1.3.

Is That Really It?

Yes, that's it. There have been some other minor edits and revisions, but those don't really go to the substance of the spec, and don't affect compliance testing. As we've indicated above, there is no reason to worry about your HDMI cable failing to support a new feature because it was designed under a prior spec version; the only thing that is a potential area of concern is that as the bandwidth being pushed through HDMI cables increases, the likelihood of failure increases as well, and a cable which was perfectly adequate to carry 480i under 1.1 may not work properly when fed 1080p deep-color video.

HDMI Licensing Position on Cable Lengths, and Noncompliant Cable Being Sold

The HDMI documentation has said since Compliance Testing Spec version 1.2a that cables incorporating the HDMI trademarks must have passed Authorized Testing Center ("ATC") compliance testing at the longest length placed on the market. The HDMI Licensing documentation is, however, something of a muddle; the Adopter Agreement doesn't impose such a requirement, and doesn't appear to leave room for such a requirement, in that it only requires Authorized Testing Center testing of an Adopter's first product in a "Category" ("Category," for this purpose, meaning type of product, e.g., source or cable) and then permits the Adopter to market further products in that Category without additional testing; and the Compliance Testing Spec isn't disclosed prior to entry into the Adopter Agreement, and so there is an interesting legal question whether anyone is actually bound by provisions it contains which are in conflict with the main Agreement. Consequently, the obligations of HDMI Spec Adopters are somewhat unclear.

What is not unclear, however, is that a lot of noncompliant HDMI cable is sold bearing HDMI trademarks. Blue Jeans Cable currently holds an HDMI compliance certificate under 1.3a (CTS 1.3b1) for a 45-foot long cable, and this is, to our knowledge, the longest distance at which any passive cable (that is, one not incorporating an amplifier) has ever passed ATC testing. We are regularly contacted by manufacturers and vendors of HDMI cable, and we routinely ask them for their compliance certificates (which, incidentally, seems to end about half of the conversations; we never get a compliance certificate from quite a few vendors). To date, we have never seen any compliance certificate for a cable, in any gauge, as long as our 45-footer, under any version of the spec. We have seen 40-foot certificates under 1.2 (without indication of the testing bandwidth), and of course since the test signal amplitude dropped about 20% under 1.3, it's likely that the cables bearing those 1.2 certificates would not pass the 1.3 testing at 40 feet.

It's important to bear this in mind. There's no such thing as a cable type being deemed, categorically, HDMI compliant. Cable assemblies are deemed compliant after they have passed testing at a particular length, and are deemed compliant for that length and shorter lengths. But just because a manufacturer has passed compliance testing, that doesn't mean that he's passed it for the length of cable he's selling you. How to know? Ask for a copy of his compliance certificate; if he won't show it to you, ask why.

There are a lot of 50-foot passive HDMI cables in the world with the HDMI trademarks stamped on them. But how many are compliant? Likely, based on what we have seen, none. If anyone has a 50-foot compliance certificate, he ought to be waving it in the air and bragging about it; the fact that there seems to be nobody doing so is perhaps as good an indicator as any that there are no compliant cables at that distance (we exclude from this remark, by the way, "active" cables incorporating amplifier circuits; it's possible to extend distance somewhat with a booster/EQ unit, but our current discussion is for pure "passive" cables, made out of nothing but wire and plastic). If anyone knows of one--and can produce an Authorized Testing Center certificate--we'd like to know about it.

Noncompliant Cable: Should I Worry?

If all of the noncompliant cables that are sold actually failed to work in-use, there would be an awful lot of complaining going on about cable quality. As it happens, however, there is ordinarily a good deal of "headroom" between what will pass the spec and what will actually, under real-world conditions, pass signal. The way the compliance testing is set up, a cable is tested to ensure that it will work when it is connected to the world's worst minimally compliant source device at one end, and the world's worst minimally compliant display at the other. With any luck, you do not personally own the worst source and the worst display in the world, and consequently, you have some amount of safety margin between noncompliance with the spec and loss of function. How much? That, of course, is hard to say and will depend upon the particular units in use.

If you're using what you suspect may be a noncompliant cable, but everything seems to be working just fine, then there's nothing to worry about, at least with your current setup. When HDMI cable fails, it fails conspicuously, generating pixel dropouts ("sparkles"), line dropouts, flashing or jumping picture, or no picture at all. If at all possible, it is of course nice to have the security of knowing that your cable is actually spec-compliant, and we hope that this article has been helpful in clearing up what that means and how to ascertain it--but if you're not seeing conspicuous picture problems of these sorts, then your cable is working fine, regardless of whether it's spec-compliant, and regardless of which version of the spec it was tested under.

Many thanks to Blue Jeans Cable for contributing this article.