Stefano Pasini


(Click HERE to download the Power Point presentation of this paper) 


(1) Warning- this paper was presented to the 3rd ‘The Eye and the Auto’ World Congress in Rochester, MI, 23-25 June 2005. I am indebted to the organizer of this wonderful congress, Dr Philip C Hessburg MD, for the opportunity he gave me to discuss the problems of the ‘unreadable speedometers’ with several other professionals of this sector. Ophthalmologists and ergonomic specialists agreed with my views. (Instrument panel manufacturers and car designers were less pleased.) This paper is designed to be read in conjunction with the similarly named *.ppt presentation that you can download from this same page.  

Abstract: Designing an ergonomically efficient instrument panel is not easy. A ‘rich’ dashboard has always been attractive for the customer; the evolution of the technology has brought on modern cars lots of useful systems most of which need their own visual interface. Fitting the essential instruments, the desirable gauges, the ‘vanity’ displays and various LCD screens inside the driver’s usable field of view is therefore quite complicated. This considerations notwithstanding, there are many cars around who have been fitted with interfaces that are probably not so well designed, and where several instruments look to have forgotten the Gestalt rules of rational design. The ophthalmologist understands these limitations but knows also the intrinsic dangers of complicated, irrationally-designed, time-consuming interface and would not like to have, in his own car, this type of nuisance. Some brands seem, in this respect, to fare better than others. 


 During the past 30 years I had the privilege to test the best cars in the world, and some of the worst as well. Some of those cars were so good that I bought them; some others were a delusion and sometimes I gave back the car to the manufacturer thinking, why did they build this one? Of course, in a high-quality modern automobile you would need to understand its character before judging its performance, as it would be silly to pretend a Rolls-Royce level of comfort in a Lamborghini. Each car has its flavor, anyway, and I enjoyed driving most of them.

I am however quite disappointed by the fact that comparatively few of these excellent motor cars have a really good, ergonomically efficient man-machine interface. My late friend Ferruccio Lamborghini declared in 1963, ‘in the past I have owned some of the most magnificent motor cars in the world and I found most of them wanting in some respect….now I want to build a car without defects. Not a technology marvel: very normal, very conventional. But a perfect car.’

I think that building a car ‘without defects’, if not ‘perfect’, would have to be the ultimate goal of any automotive designer. It is therefore surprising that in so many automobiles I find interfaces that I find to be less than desirable.

I hope that nobody will feel offended by what I shall say here and show in this paper: it is not my intention to offend anybody in particular. Please understand that I am not criticizing the fact that designers propose new ideas that are sometimes quite different from what I, as an ophthalmologist, would think to be good for this task. I might be wrong. I fully understand the need to express everybody’s creativity; and I know that there is some  pressure coming from the commercial department to offer to the potential customer a constant stream of ‘innovations’. A famed automobile designer predicted, some years ago, that headlights were bound to become ‘the car’s jewelry’; the instrument panel is already something the customers see as a special type of ‘jewelry’ for the interior of their car. What I’m about to show is simply what I would not like to see on my personal car: ideas and interfaces that I do not think to be ergonomically efficient.

I want solutions that make my life, and my driving, safer and, if possible, faster. Real life is not what appears to be on our PC screens or on the racetracks or on Soviet military airports, where a superfast car can be safely operated to its maximum performance. The average drivers’ real life is made of congested traffic, poor light, glare, fog, dirty screens, imperfect spectacle prescription, senescence (me, too) and so on.  For this reason I think that we would have to try to make things easier, enabling all the drivers to spend more time looking at the road and less at the dashboard.

In the past, airplanes and aeronautic engineering had a strong influence on some aspects of the man/ machine (driver/automobile) interfaces. The instrument panel of a Lancaster bomber was a no-nonsense workplace designed for aviators who were flying these complex machines at night, in the heat of the battle, over the skies of the Third Reich. It still may be looked at as a benchmark for the layout of an instrument panel that one must read quickly and accurately in critical illuminance conditions while actually conducting the vehicle with his eyes. For the automobile designer, this is a far more rational inspiration panel than the one fitted to a modern F-18; in the latter the visual predominance of the instruments is easily explained thinking that this panel is actually all that the pilot needs to fly his plane, fight, take off and land.

Things are very different in an automobile, where the eyes of the driver must first of all look around to direct the steering, operate the brakes, follow the traffic and the road indications, avoid moving or fixed obstacles and so on. A person driving his car at 100 mph on a highway at night is in a condition that could be assimilated to that of the Lancaster pilots of yore. And, as Santayana said ‘those who ignore history are doomed to repeat it’, I think that we must look with interest at combat aircraft… of 60 years ago.

To begin with, I wouldn’t like to have a completely digital display for the instrument panel of my car, exactly as I wouldn’t buy a digital watch. The reason is always the same: it takes more time to read a digital indication than an analogue one. Though the digital instrument may communicate a more precise data, the rapidity of reading of an analogue one compensates for its approximation. The legibility problems of an all-digital LCD layout when ‘washed out’ by a strong ambient light are known, so I like to see them used only for secondary displays.

Speaking of analogue interfaces, I don’t like white dials, as I already mentioned in my other paper here. Please see in this slide the difference of visualization of the two watches at the same time of the day and with the same ambient light. Several years ago I bought the fashionable chronograph of a very famous Swiss brand only to find that it was often nearly impossible to see what time it was on that white dial with chrome hands. I use since many years another model of chronometer of that same Marque and its chrome hands on a black background are always perfectly readable. This is why I would never, ever buy a car with white/silver/parchment instruments.

Another nuisance of certain modern speedometers is that the central part of their dial hosts a display covering most of the needle of the instrument, a solution that certainly does not improve its readability. This electronic instrument cluster is designed to display an array of information in less space than conventional clusters. This automobile manufacturer has a significant share of the taxi market, so it was seen as desirable to have a central multipurpose display where it is possible to integrate a number of ‘taxi functions’, for example, communication between a taxi driver and dispatch. Yes, but I’m not a taxi driver so I don’t need this possibility, while I would frankly prefer a full-length needle that would be easier to read than a short one, especially when the latter is reduced to be just a short segment of a pointer turning along the outer circumference of another display. And placing more small dials inside the speedometer may seem a brilliant idea; it surely is for a chronograph. It isn’t, for my tastes, in a sports car. Too distracting. Is a voltmeter so necessary nowadays?

To maintain an acceptable level of safety, the eyes of the driver must remain on the primary task (looking at the road ahead) the longer possible time. Glances away from the forward view must be limited; in-car tasks must therefore be designed so that the driver’s eyes will have to spend the shortest possible time looking for them. A special attention must be given to the reduction of the ‘transition time’, i.e. the time spent changing the target of the VS from the primary to a secondary task. This time increases proportionally to the increase of the angle between the forward view and the location of the in-car task; an added complication is that subjects over 40 years of age and as such with a tendency to presbyopia might experience trouble in the constant change of focal plane between forward view and in-car tasks: after having spent an effort in focalizing an instrument panel they could experience some ‘defocusing’ returning to the forward view. Finding a radio station must be simple and the audio system controls must be extremely easy to identify and operate. Sure, they are controls of no importance whatsoever for the operation of a motor vehicle, but they are an in-car task that some people spends a lot of time on, so designing buttons and knobs like this, is frankly irritating and potentially dangerous. Don’t matter how much secondary a task may be, if it’s not properly designed a driver could find himself spending a lot of time trying to operate it.

As the visual efficiency decreases from the centre of the visual field (macula) to the peripheral areas of the retina it would be advisable to put all the in-car task as near as possible to the line of forward view to reduce the driver’s mental workload, so moving the instrument ‘binnacle’ in the central part of the facia, between the driver and the passenger, does not seem to improve its readability. This location forces the driver to deviate his gaze away from the road ahead more than a conventionally-placed panel would require. As an author rightly remarks, ‘…it is generally recognized that as the angle between the forward view and the in-car task increases, transition times for the eyes also increases…. finally, the further down into the car the task is, the less likely it is that peripheral vision could be used to detect a hazard in the forward scene.’

Placing the main instruments in the center of the dashboard means also that they will be more distant from the driver’s eyes. Reading them at a glance will therefore be even more difficult for senior drivers wearing bifocals, as this position is usually too far for their ‘near’ correction, or progressive lenses, as they will have to turn their heads a bit more to centrate the instruments in the middle of the ‘optic channel’ of their lenses.

We must use digital interfaces for several instruments, anyhow, and full video screens are the most widely accepted interface for navigation systems. A careful design of these screens is necessary: avoiding glare that might impair the readability of a display would have to be one of the first tasks of a display engineer. Glare on a screen is very bad for readability, and we must avoid it as much as possible especially if it incorporates touch-screen controls. I’d do happily without this kind of satnav screen.

I’m happy to see that some ‘creative’ designs have disappeared altogether, as is the case of ‘rotating drum’ speedometers that have been fashionable on some Gallic product for a luckily short time. It is not necessary to explain to this audience why it is absolutely unadvisable to use a rotating scale against a fixed pointer to show the speed you’re traveling, or any other data: it was a bad idea, though at the time it was thought to be ‘innovative’. Speedometers must be designed so that they can be read quickly, accurately and mustn’t force the driver to take his eyes off the road for more time than strictly necessary, especially if one fears that a State Trooper might be waiting around the next corner with a laser gun.

Installing a feeble red light to illuminate the interior of a car is an excellent example of the application in the automotive field of a concept widely accepted in airplanes, where a red night illumination has always been used to protect the all-important mesopic adaptation of the eyes in low ambient light.

The choice of the colors for the night illumination of the instrument panel is also very important; some colors, like blue, might cause trouble. Blue is the most visible color of the spectrum in mesopic/scotopic mode and is therefore the hue that could more easily disturb this delicate condition of adaptation.

The other factor that may advise against the use of the blue illumination is that one of the most important warning lights of the whole instrument panel, the main beam light, is conventionally blue. This color has been chosen also because this allows to the warning light to be easily spotted in an instrument panel conventionally illuminated in amber, green or orange. As a result the driver will see at a glance that he is using the main beam and will switch them off; keeping the main beam accidentally on would cause a real danger for the oncoming motorists. But this color contrast, that is at the basis of the easy identification of this warning light, is reduced to unacceptably low levels when the whole instrument panel is illuminated in blue: the driver has therefore more chances to turn involuntarily on the main beam and this is not good.

The ‘Warrick principle’ about rotary controls is another worthy stereotype: an operator moves a knob so that the part of it nearest to the display moves in the direction he is trying to move the display indicator (Warrick, 1947) Sadly, the controls of several automobiles ignore this principle and sometimes the arrangement of the control/display combination happens to be the opposite of this and other ergonomic principles. This may cause a loss in the efficiency of the control itself and the driver’s mental workload would increase, so I don’t like knob/pointers that work the opposite way.

The Warrick principle applications remind us that a good way to improve safety  is to design controls and displays according to ‘stereotypes’, because, as Bruno Munari said, ‘we see what we know’. A classic example of a widely accepted stereotype is the clockWhise or ‘rising’ movement of an indicator to signal an increase in the value of the monitored data, be it speed, rpm, temperature or pressure. I drove several years ago a sports car where the fuel gauge needle indicated a full tank when it was at its lowest position and an empty tank at the top. Why?

Another aspect of interior design that nowadays leaves something to be desired is the control of glare on the windscreen. Some dashboards and facia are now very ample, their styling is elaborate and sometimes multi-colored, the windscreen is steeply raked: in a bright, sunny day the light will cause some very disturbing glare on the windscreen, reducing again the luminance contrast of what the driver can see through it. If the upper lids were all black this would be much less of a problem than with light colors.

So, is there any good dashboard/instrument panel around? Sure. Most of the less-expensive sedans, many utilitarian cars, a lot of small inexpensive sports cars have a decent (clean, rational, uncluttered) man/machine interface. They have one thing in common: they don’t offer much in terms of gimmickry, of visual entertainment. Their interfaces are always simple, sometimes crude. So they are perfect for their job.

On the other hand, if we look at the upper crust of high-performance sports cars, graphic ‘sophistication’ seems to make things worse. On my car I would not really want a red speedometer with a red needle, for example.

Some modern high-performance cars have, however, some very good solutions. There are some automobiles, mostly German, I really like also for the design of their visual interfaces.

This is a HUD that I found to be extremely well done and useful; this is the ‘sports’ mode, whilst a normal mode shows a very simple but effective navigation aid.

A classic approach: several analogue dials of adequate dimension and perfect visual structure was fitted to the most classic model of another famous manufacturer. In my humble opinion, these cars offer the best visual interfaces I can desire. Black dials, white numbers, orange needles, neutral night illumination are the strong points of the visual interface of this classic sports car. The clock is analogue, too, and it is all well readable in any illuminance condition. My favourite.


Designing a new visual interface for the instruments and controls of an automobile is a real challenge. In fact, keeping in line with the traditional black instruments with white dials and needles is always safe, as this type of instrument is cheap to build, simple to install and to illuminate for night driving, easy to read and understand. Any variation on this theme, or a radical innovation, could worsen the readability of the data transmitted by the instrument and force the driver to spend more time trying to read it, thus diverting his eyes for longer times away from the road. Maintaining a good luminance contrast and conforming the displays according to existing stereotypes is a good way to avoid the increase of mental workload, and the consequent dangers, that might be associated with an unusual visual interface.


Dott. Stefano Pasini MD, Bologna/Rochester, 25 June 2005




Hit Counter