There are physiological disadvantages to HID auto headlamps that do not exist with glowing-filament lamps.
Probably the biggest issue is HID headlamps' significantly worse color rendering index (CRI), which is in the high-60s to low-70s range. Halogen headlamps' CRI tends to be around 90 to 97 or so. In English, this means that the human eye's color perception and differentiation is much, much better under halogen light than under the light produced by automotive HID headlamps.
Now, how do we reconcile this with the ad copy that seems universally to talk about how HID headlamps' light is closer to daylight? Well, from a color temperature standpoint, that's generally true. However, that's not the whole story. Color temperature is only one factor that goes into describing the quality of light from such a thing as a headlamp. But here's the tricky part: There's no evidence that "closer to daylight" is the right stuff to drive with at night. Sure, an easy case can be made by just assuming that because daylight is what we have during the day, daylight is what we ought to have at night, too. But it isn't that simple at all. The eye has a very different set of jobs to do, using a different box of tools, at night compared to the job and tools during the day. The extent to which this influences your safety behind HID headlamps is not currently known. The full extent of the current knowledge on the topic as relates to current HID headlamps is "You can probably see colors well enough to be safe". But that's it! (Note that it's from the same researchers who say that you can see well enough to be safe when driving e.g. a '93-'97 Camaro with those miniature oblong sealed beams...) So all the hype about "closer to daylight" is really meaningless in the real world, and may in fact be misleading; there is research showing improved distance perception with headlight of *lower* color temperature, for any given intensity. Certainly any kind of shift towards the blue (as with HID headlamps) is a step in the wrong direction in inclement weather (fog, rain, snow, etc.).
As an illustration of the fact that color temperature does not automatically bear on headlamp quality, consider this: The selective-yellow headlamps required for so many decades in France were shown, in a couple of studies, to improve (or at least not to reduce) driver performance at night. There are some color issues with this light color (it's hard to tell a yellow road lane marking from a white one, for instance), these are easy to deal with by changing the lane marking color or using only one color. BUT, a sodium lamp, which is a kind of HID lamp, would make a *lousy* headlamp even though its color temperature is not far from that of a selective-yellow headlamp. That's because sodium lamps have EXTREMELY low color rendering indices. Many sodium lamps suck almost ALL of the color from whatever they illuminate, creating a black-and-yellow world. There are sodium safelights for use in photo darkrooms, and the monochromatic-world effect you perceive in such a darkroom is very eery until you get used to it. In a driving task, this would be a disaster! Color is a very important information carrier in our task to see what we're seeing.
Now, HID headlamps do have a big technical advantage over most current glowing-filament headlamps in that the efficiency of the light source (measured in a straightforward, easy-to-understand way: Lumens per Watt) is quite high. As a result, an arc capsule that consumes only 35 watts can produce double to triple the *amount* of light (around 3200 Lm) as can a typical halogen bulb (around 1500 Lm in the better designs, e.g. H7). So that's good, right? Well, it gives the designer of the headlamp a couple of options: He can keep the same performance, but have a much smaller headlamp, or he can keep the same size headlamp, and have the potential for higher performance. But now we hit another "it's not quite that simple".
A headlamp isn't a flood lamp. With a flood lamp, we put electricity in one end, and we get light out the other end, and we don't really care where the light is, as long as it's in a big ("flood") beam. A headlamp is a much more complex animal, 'cause it has a much harder job to do. It has to maximize your distance vision while minimizing glare to other drivers. It has to light up everything that you need to see, BUT not light up areas that would detract from your ability to see what you need to see. The first point makes it very difficult to improve low-beam headlamps, because increases in seeing distance almost always bring with them higher glare.
The second point is even more crucial. You remember from above my statement that the eye has a different job to do at night than during the day; here's where that comes into play.
During the day, pretty much everything is illuminated relatively evenly. If it's sunny out, everything's bright. If it's cloudy out, contrast is reduced and colors are muted. If it's foggy out, everything's fuzzy. But at night, your world consists not of "everything", but of that which is illuminated by your headlamps. Everything outside of that "world" is dark. Which is fine, except that your "world" moves with you! The extreme contrast between your "world" (that which is illuminated) and everything else (not illuminated) creates the difficulty. That's why we have headlamp glare at night, why we squint when we come out of a movie theatre after watching a perfectly bright screen for 3 hours, why we hold our hand in front of our eyes when looking in the direction of the sunset and trying to read a road sign. It happens on that big scale (inside/outside your "world") and it also happens *within* that "world".
If you illuminate the foreground very strongly, your eyes will adapt to that big area of strong illumination, your pupils will become smaller, and your distance vision will be reduced. (On the other hand, if you take the route prescribed for so long by US headlamp regulations and have very LITTLE foreground illumination, you'll have a "black hole" in front of the car, and you'll be straining to see what you're about to run over...)
So how does this relate to HID headlamps? Well, suddenly we have all this extra light to work with, because we're using an HID arc capsule instead of a bulb. Where are we going to put the extra light? We can put some of it into the high-intensity zone of the beam (the "hot spot") to improve distance vision, but we can't put too much of it there, 'cause we'll glare other drivers (and exceed regulated maximum intensities). We can spread some of it around in the rest of the beam, but there are often even stricter maxima outside of the hot spot or zone, and too much "generalized" light causes veiling glare and backdazzle in bad weather. So we put a lot of it in the foreground. To an extent, that's a good thing, because US headlamps have typically had too little foreground light (see above). But over a certain level, which is quite easy to exceed with an HID headlamp, undesireable things start to happen. We sacrifice distance vision, we get high levels of reflected-light glare on wet roads, that sort of thing.
But what about all the rave reviews you read in magazines and on Usenet about how marvelous HID headlamps are? Well, here's the double whammy of high levels of foreground light: When you can see every last pebble in the road, all the way across, it's very comforting. You don't have to strain to see what's immediately in front of the vehicle. And the most recent research seems to be suggesting that this, to an extent, improves driver performance at night, possibly because with the foreground adequately illuminated, we tend to trust our peripheral vision to handle the foreground, and keep focussed out in the distance where we should be looking. (See above comments regarding too-low levels of foreground light in many US headlamps...). BUT, foreground illumination is *only one* performance aspect of a headlamp, and it's very, very easy to judge a headlamp "good" because of very strong foreground light, when in fact the distance vision isn't so grand. In short (finally!) an HID headlamp has a strong tendency to create a false sense of security.
There's (even) more to it than all of this, but I think it's time to stop for now. HID headlamps can be an improvement over halogen ones, but it's not automatically the case, hype notwithstanding. Beam design is much more crucial than light source.
There are new, extremely efficient halogen bulbs coming on the market that offer increased intensity without CRI problems or excessive-light problems. The glowing filament is going to be around for a long time. It's also reasonable to expect that HID headlamps will also develop.
Source: