Automated lights have very definitely come of age in the last couple of years, having gone through the same stages of growth all organisms experience. The first is the infant stage, when everyone adores them, but they need constant attention to keep from falling over and hurting themselves. Then come the “terrible twos,” throwing tantrums and breaking everything. Next, they learn to walk on their own and gain independence from the babysitters. They even got through puberty without too many psychedelic experiences. Now, you could say they've reached a steady middle age — solid and reliable, working a 9 to 5 job without complaining.
But does that solid reliability also mean they've become slow and staid? Does having a product that works the first time, every time, inevitably mean it's also boring? Have we lost our excitement with automated lighting, and could it be that it's the very fact that they are mature, commoditized products that's causing that ennui?
In those early, heady days of VL1s and other rental-only products, there was an air of mystery surrounding moving lights. Only the select few were admitted to the inner sanctum of the maintenance area to learn how they worked, and among the crew was always that slight ripple of fear in the operation of a show — will everything work tonight? It certainly kept the adrenaline flowing.
And you couldn't buy the products. They cost too much to manufacture, the technology was expensive to get into, and the learning curve was steep.
Nowadays, we get upset if they don't work. The maintenance areas have shrunk dramatically, and anyone can afford to buy an automated light. But have we lost something in the product itself to pay for this apparent progress?
Let's go through some of the issues surrounding manufacturing an automated light designed for mass production and mass use and see where this may be both good and bad. This process is inevitable in a consumer-driven market, unless there are connoisseurs prepared to pay a premium for a performance product.
WE'VE GOT CONTROL ISSUES
Let's start with control. Those early moving lights used dedicated proprietary control protocols designed specifically for automated lighting and its needs and foibles. Nowadays, everyone uses DMX512 and for good reason — DMX512 is ubiquitous, everyone speaks it, everyone knows how to deal with the cabling and distribution, and the hardware is inexpensive. It also works reliably. So, what's wrong with that? Well, DMX512 is really just too darned slow for automated lighting. It was designed to control dimmers feeding incandescent lamps with slow thermal rise times where a maximum refresh rate of 44Hz was just fine.
That same 44Hz is much too slow for controlling motors smoothly. A typical moving light using stepper motors might want to update those motors every millisecond. So, it needs updated data at 1000Hz — that's nearly 25 times faster than the 44Hz offered by DMX512! The motor driver must update this quickly so that the moves are smooth and crisp and acceleration ramps are natural looking. That way, you can do a five-minute slow move across the cyc or a snap across stage, and the light will respond in an organic, non-robotic manner. In order to get around this mismatch in update rates, the moving light has to interpolate values in between the DMX packets and try to predict what the next DMX command will be. (Figure 1)
Those prediction algorithms are getting very good, and you probably aren't aware they are even there most of the time, but they inevitably introduce a small lag in the response. Until the fixture has seen at least two DMX packets, it can't predict how fast it's supposed to move — that means at least one DMX packet delay. The alternative to that response delay is to use timing channels (variously called focus timing, Mspeed, or Vector by different manufacturers), where the fixture calculates all the intermediate positions internally and just uses DMX for the endpoints. That works fine but introduces problems of its own. Using timing channels means more programming and forces the designer to use linear moves. You can't do a profiled move using linear timing channels.
Arguably, using a control protocol designed specifically for moving lights could alleviate some of these issues, but that was the situation before we all chose to adopt DMX512. The pluses of universal control, perhaps, outweigh the minuses.
Although most noticeable in pan and tilt, the same argument applies to every parameter in the fixture. If you want to be able to smoothly rotate a gobo from one position to another, then 44Hz just doesn't cut it.
SERVOS VS. STEPPERS
These problems arise whatever kind of motors you use, and that brings us to the next question: servos or steppers? Each has its advantages and disadvantages. Some early moving lights used DC servo motors and with good reason — they are inherently smooth and, because they usually have an integral feedback mechanism, they always know where they are. This feedback mechanism can be either absolute or relative. Absolute mechanisms report the precise position of the system at all times, while relative systems, as their name suggests, report the position relative to some point, usually the “Home” position. If they are the absolute type (such as a potentiometer), the fixture will not need a homing or initialization process, whereas the relative systems (such as optical encoders) need some kind of physical homing to establish their fixed start point. With either type of feedback, the fixture will automatically correct any errors if the unit is knocked out of position. Servo motors are designed for exactly this kind of application and, you would think, are the obvious best choice.
Stepper motors, however, also have advantages. They are much simpler and more reliable, they are easier to make tolerant of high temperatures, and, probably most significantly, they are dramatically less expensive than servo motors. The difference in cost can't be ignored; it can be a factor of 10 to 20 times or more. There's a penalty to be paid for this, though. Stepper motors, as their name implies, step (surprise!), and we don't want to see steps. We want to see smooth, seamless movement. The most common uses of stepper motors outside of the entertainment industry are in printers, photocopiers, and fax machines, and the users of those devices don't care about smooth movement. They just want cheap and reliable. That means there's a lot of work going on behind the scenes by automated light manufacturers trying to smooth out those steps so you can't see them.
HEAVY METAL
These problems get worse as the fixtures get heavier. We started out with 250W and 400W moving lights, but now, the most common workhorse products are 1200W units with a corresponding increase in size and weight. In addition, they have more gadgets and gizmos inside — more gobo wheels, more color wheels, more everything, giving another increase in overall weight. All these extra features are positive and useful additions to the LD's arsenal, but now, we have a 60lb. head that we would like to move as quickly and as accurately as the old 30lb. one without breaking anything. The manufacturer has to err on the side of reliability and repeatability, and so, we see an inevitable reduction in movement speed.
The need for higher and higher output has other side effects, too. As the light source gets larger, the size of the optical systems has to increase proportionally. Thus, gobos and irises get larger, and lenses have to increase in complexity to try to obtain the same quality of focus. Bigger gobos can mean slower gobo changes and slower gobo spins. Add to this the increase in gobo weight caused by going from static to rotating, from metal to glass, from single color to full color, and you have an assembly that just cannot move as quickly as it used to. But would we be happy with a return to small, static, metal patterns just to get those snappy changes? Some would say yes, but many more would disagree.
Another area worth mentioning is color. Mats Karlsson (LD, August 2004, p. 40) described the development of color mixing mechanisms. You can see from the photographs in his article how complex some of these mechanisms are, especially when compared to a simple color wheel. Again, are we willing to sacrifice the ability to mix any color we want for instantaneous color changes? Pretty much nothing can bump colors as quickly as the early moving lights with their simple, fixed color wheels, but would we buy something that limited today?
LARGER = SLOWER
So now, we have reached a point in the evolution of automated lighting fixtures where there is an expectation of a certain performance and quality — no matter who the manufacturer is. A few years ago, one valid reason for choosing Manufacturer A over Manufacturer B was that Manufacturer A produced a product that you knew would work and would survive the tour, while Manufacturer B's product would fail the first week. Nowadays, you can reasonably assume that every manufacturer has reached a certain quality level. They all work, they survive pretty well, and they do a reasonable job lighting your production.
As in any emerging industry, demonstrably bad ideas have been discarded, while the good ones have perpetuated and spread. It's very much a Darwinian evolution — only the fittest survive. This means an inevitable leveling out of fixtures. They all work pretty much the same way and offer similar basic functionality, honed down by this natural selection to suit 95% of the user population 95% of the time. This is a symptom, good or bad, of a maturing commodity industry. Of course, each manufacturer and product has something to distinguish it, but those distinguishing features are evened out quickly.
0 TO 60…OR NOT
To pose an analogy with consumer level cars: we expect all cars to work. They all have the same basic functions, and they all perform those functions well. We have a level of expectation with reliability that just about all cars meet. They all look pretty much the same, and they are, in many ways, interchangeable. Frankly, most cars are boring!
Where this analogy breaks down is that you can buy an exciting car. The market is large enough that it can support niche products aimed at a tiny fraction of the total market. For example, even a tiny portion as small as 0.1% of the annual $500 billion automobile market is still $500 million, and that's worth having! However, that same 0.1% of the annual moving light market is less than $1 million a year and not a very attractive business proposition. It's difficult to justify a sports car level, high performance moving light product on those figures.
In the same way, it is certainly true that you could produce a high performance unit with enhanced functionality, where the fixture mechanics are pushed to the limit, but I'm not convinced we are willing to pay the price — in cash or maintenance — to justify this.
Are the manufacturers of automated lights holding back on performance in favor of reliability and a low price? Yes, of course they are, but they are only doing it because that's what the market has dictated. Once you consider inflation, the real selling price of moving lights today is something like 50% of what it was in the early 1990s. Again, this is a typical result in a maturing industry moving toward commoditization.
Maybe a prestigious touring production would like a customized high performance product, and maybe they'd accept a higher rate of breakdowns in the same way that sports car drivers do. Frankly, I'm skeptical. I suspect that LDs would scream like mad if the fixtures broke down and with good reason. This is where the analogy with cars breaks down further. People love cars for their own sakes; they become an object of desire disconnected from any job or task they perform. However, a lighting fixture is always a means to an end. It's a way for the lighting designer to express creativity and paint a vision on the stage. A light is a tool, pure and simple, and you want your tools to work, always, without exceptions.
After an early career as a lighting designer, Mike Wood has been responsible for the technological strategy, R&D, standards, and IP in many different companies within the entertainment industry. He is the current president of the Entertainment Services & Technology Association (ESTA). Mike can be contacted through his company, Mike Wood Consulting LLC, at [email protected].