Shed some light on this

[Coxy]

Okay, I've got my Friday night brain on so I'm looking for some perspectives on a lighting question from some of the more physics-minded readers of the forum.

N scale is approximately half the size of HO. This means that an N scale boxcar has about 1/4 the area of the same HO scale car. This means the N scale car is only able to reflect 1/4 as much light back to our eyes as the HO car, all other things being equal. Even less light bounces back from the same Z scale car but for now let's just focus on lighting needs of N vs. HO.

Does it follow then, that we should either view N scale from half the distance of an equivalent HO object to see it equally well, or, if the viewing distance is about the same, the N scale object should be lit with 4 times as much light for it to appear the same as the equivalent HO object?

And so that's the question, should N scale models/layouts be lit the same as HO, or, will N scale models/layouts be viewed comparably to their HO scale counterparts if lit with four times more light. Empirically, N scale models look better with more light, does the physics bear that out?

Your thoughts?

- Coxy

[Zox]

My thinking on this would be that no, lighting doesn't scale in that way. Even though you're getting back only one-fourth the light from the N-scale car, it's also coming back from one-fourth the area, so the intensity doesn't change.

There might be some benefit from brighter lighting, in that it causes your iris to "stop down" and thus increases depth of field, but it's not directly tied to scale.

I've have heard people advocate painting scale objects in colors that are lighter and grayer than the prototype, to simulate the atmospheric effects of viewing at a distance. I'm not convinced of the benefit, though, except for very distant (and stationary) objects. It's the same problem as forced perspective; it only works if you can closely control the viewing angle.

[Rossford Yard]

Presuming that you are lighting the layout area about the same either way, wouldn't you naturally be providing that 4X extra light to your N scale cars?

I know my light fixtures were located to keep gaps in the layout down, shadows down, etc.  While there was an MR article suggesting flourescent tube fixtures have no gaps more than 1 foot between them, mine actually have a bit more gap than that without shadows.

BTW, in practical terms, I think its more important to make sure the fixtures are shining on the front of the cars.  If your track is 2" in from the layout edge, your fixtures better be right above the layout edge, or if can lights, above the aisle.  In other words, placement behind the moving trains really cuts the % of light reflecting off the cars and makes the other discussions moot.

[GaryHinshaw]

Hmm, Friday night brain eh?  Still have the same questions Saturday morning?   

Generally, as Lord Zox notes, you don't need 4 times the light intensity (lumens) to perceive N scale trains as well as you would HO, but you really have to define what you mean by "perceive" to say much more.

First, it is true that you have to be 160/87 ~ 2 times closer to an N scale model to have it subtend the same solid angle (angular width x angular height) as the equivalent HO model.  But, assuming the painted surface does not scale, the surface brightness will be the same for both models, given the same room lighting, so they will appear equivalent to your eyes.  (As you get close enough to perceive differences in surface roughness, this equivalence will start to break down.)

However, we generally observe HO and N models from the same distance (a few feet), so the equivalent models will subtend different solid angles.  Now it's important to think about how you detect the light coming from the models.  Think of your eyes as a digital camera with a bunch of pixels (not a perfect analogy, but not too bad). Then you have to ask what a given pixel (rod/cone) detects.  For a fixed lens, a single pixel see a very narrow cone (really a square cone) of directions, so what matters for that pixel is the surface brightness of the model in that cone's direction.  Next, you have to ask what a neighboring pixel detects: the surface brightness in the neighboring cone, and so forth.  In the limit that neighboring pixels are just seeing different portions of a ~uniform surface, they will detect both models equally.

But of course we care about non-uniform surfaces (i.e. detail) so then next questions is: does your ability to resolve detail depend on brightness?  For a digital camera, as long as you are not in the shot noise limit (where you start to see graininess in the image) the answer is no: increasing the light level will not help you resolve detail, it only depends on the angular size of the pixels.  To first order, the same is true of the eye/brain system, so increasing the brightness will not help resolve detail, unless you're in a noise-limited regime.  You can be in that regime in a dim basement, so bright light is better, until it makes you squint...  But it doesn't need to be 4 times brighter for N than HO.

-gfh

[David Leonard]

The eye (brain) can adjust for light levels between 1 (minimum) to 10,000 times that, so the intensity of lighting within a small section of that range makes little difference--certainly not 2 or 4 times the intensity. In point of fact it's the optic nerve / brain "electronics" that create the actual image we perceive. The optics of the eye itself are not all that great--the image on the retina is somewhat blurry. The optic nerve / brain system sharpens it up, much in the same way that digital photos can be enhanced. Looking at a telephone wire from a great distance is a good example: on the retina the image of the wire can be kind of jagged (because the rods and cones "pixels" don't always lie in a  line relative to the image of the wire), but the brain enhances the image so that we perceive it as continuous. More important IMO is the angle of the lighting (as said above) and the color temperature of the lighting.

Geez, this is Saturday morning. Why am I writing this?

[ednadolski]

Generally, as Lord Zox notes, you don't need 4 times the light intensity (lumens) to perceive N scale trains as well as you would HO, but you really have to define what you mean by "perceive" to say much more.

Stereoscopic vision is another factor, but that is a case where perception IS affected by scale, because inter-ocular distance is a constant regardless of the scale of the model (unless you can come up with a way to move your eyes closer together in proportion to the scale)

Ed

[sirenwerks]

I agree with David, angle of lighting is more important. Size doesn't matter (had to say that). Having recently bent to the will of age and donned bifocals, I had a devil of a time 'tricking' my eyes out of the color separation the new lenses were causing. Of course, it wasn't the lenses so much as my eyes, the curvature of the new lenses was greater than my old glasses and my eyes refused to adjust (they finally have now, I think). The rods and cones are going to react whether the object is big or small, and the adjoining hardware and software will finish the job.

Placement and angle of lighting will be more important. You don't want the viewer to be subjected to direct view of the lighting, which will wreak havoc on their ability to adjust to viewing within the room environment, or harsh light angles that create a blinding reflection back at the viewer off the model.

Model surfaces will play a big role too, whether objects are dulled or not. Car sides, building sides, and backdrops are flat and create a direct reflection which needs to be accomodated, but I've also noticed in photos and in person lately how unrealistic CMW vehicles look when they're placed on a layout straight out of the box. They all look like they just came out of the carwash - highly unlikely, especially around a railroad with steamers or Alcos. Obviously, scenery items are going to absorb reflection, and train models that have their sheen removed with weathering and dullcote will too. But those autos always glare back at you. A shot of dullcote would help to 'minaturize' the coating's finish, or maybe Tom needs to do a vehicle weathering tutorial.

[Coxy]

Thanks for all the great comments, you guys have clarified things a lot. Sleep helps too! What got me thinking about this was how poorly some dark rust details were showing up on a dark green CGW covered hopper I was weathering.  Here's what I mean.



There is very little contrast between the dark brown rust and the dark green of the car side. But I couldn't recall having the same difficulty perceiving similar details on HO scale cars.  The choice appeared to be to make the weathering more prominent (not really the goal of the exercise) or reconsider the amount of light the car would be viewed under.  I didn't want to go to the trouble of adding weathering details that ultimately would be hard to see.

Given the popularity of valance lighting, or multi-deck layouts where the lighting for the lower deck is mounted under the upper deck, tracks near the front edge of the layout will often be lit more vertically. With that particular lighting angle fixed, the question seemed to be limited to one of more light.  It's cool that you guys gravitated quickly to lighting angle as being the more important factor.

I have a very basic lighting set up where I work at my desk. Normally, I have an articulated magnifier light which has two small fluoros.  I position the magnifier at about 45 degrees to the cars. But the cheap cast mount failed, so I'm temporarily down to a 20W halogen desk lamp while I make a new mount for the magnifier light.

Here's the same shot of the CGW car as above, this time lit with the same 20W desk lamp at about 45 degrees to the car, same distance, same camera settings and position. (Without any assistance from the desk light, the car shows up as a dark green blob.)



The contrast between the rust and the car side is more pleasing which is not surprising as contrast depends on light intensity.

The take-home appears to be:

• More light won't make as much difference as optimizing the lighting angle
• More light won't compensate for poor lighting angle
• Avoid vertical lighting angles
• Position layout lights so that they adequately illuminate the car sides
• If the lighting can't be moved back from the front edge of the layout, try to position tracks back from the front edge

...at least if you want to see the details.

Cheers, Coxy