Two numbers decide whether a projector can do the job in a particular room, and neither is the one on the box. The first is throw ratio: the distance from the lens to the surface divided by the width of the image it makes there. The second is where the audience is sitting relative to the light path.
Throw ratio is the specification that matters
A lens with a ratio of two makes an image half as wide as its distance from the screen. A short throw lens fills the same width from much closer. The room's geometry is fixed, the required image width is usually fixed by the design, and the lens is the variable that has to absorb both.
Getting this wrong is expensive in a specific way: the projector is already hung, the image is the wrong size, and the only remedies are a different lens, a different position or a different image. All three are load-in problems rather than adjustments.
Brightness is spread over area, and area grows fast
A projector emits a fixed quantity of light. Push the image wider and that light is spread over a larger surface, and because area grows with the square of the linear dimension, doubling the image width quarters the illuminance on it. A machine that looked more than bright enough in a demonstration at four metres wide is a quarter as bright at eight.
This is why brightness specifications are close to meaningless without the image size and the ambient conditions they were quoted for, and why the honest question in a venue is not how many lumens are available but how many arrive per unit of surface with the house lights and the stage lighting doing what they will actually do.
Keystone, and the difference between optics and arithmetic
When the lens axis is not perpendicular to the surface, the image arrives as a trapezoid rather than a rectangle. There are two ways to address it and they are not equivalent.
Lens shift moves the optical path without tilting it, so the image stays rectangular and every pixel is still doing its original job. Its range is limited by the lens design. Digital keystone correction does something else entirely: it distorts the image in software before projecting it, which means resampling. The result is a rectangle made of pixels that no longer map one to one, softer than the source, and worst on exactly the fine detail that titles and diagrams consist of.
The seats you lose, and the shadow you buy
Front projection puts a beam across the room, and anything in the beam appears on the surface. A performer downstage of the projection casts a shadow on it, and the steeper the angle, the more of the stage becomes unusable for that reason. Productions frequently discover this after blocking is set.
Rear projection removes the shadow entirely and requires depth behind the surface, which many stages simply do not have, along with a surface designed to transmit rather than reflect. That trade has its own article in this section.
Seating is affected in three further ways. A projector hung in the auditorium blocks sightlines from somewhere. High gain screens are brighter on axis and noticeably dimmer to the sides, so the gain that bought brightness for the centre block took it from the wings. And a projector is a machine with fans, which in a quiet house is audible to the rows nearest it.
The surface is a scenic element, and it is lit
A projection surface does not exist in isolation from the lighting plot. Every lumen of stage light that lands on it raises its black level, and black level is what carries the impression of contrast. A screen with a good deal of spill on it produces an image whose darkest value is the colour of the spill, and no adjustment inside the projector recovers it, because the projector cannot subtract light.
This is a scheduling problem as much as a technical one. The projection is usually assessed in a dark room and the lighting is assessed with the projection off, and the first time anyone sees the two together is the technical rehearsal, at which point the fix is to move either a fixture or a surface. Looking at them together early is free; looking at them together late is not.
Stacking, blending and the reason to avoid both
Where one machine cannot deliver enough light, two can be aligned on the same image, and where one cannot cover the width, several can be blended edge to edge. Both work and both add a class of problems: alignment that drifts with temperature, colour that does not match between units as lamps or sources age at different rates, and a blend seam that is invisible at focus and obvious from the third row.
The cheaper answer is usually to reduce the image. An image sized to what the room and the machine can actually do, with the design adjusted to suit, outperforms an oversized one that is dim, soft and seamed.
What we cannot verify
Lumen figures, contrast ratios and lens specifications are published by projector manufacturers under measurement conditions that are frequently not stated and are not comparable between brands. We reproduce none of them. Screen gain figures come from screen makers and describe on-axis performance under laboratory conditions. The reliable assessment is a test in the venue at the intended image size, with the stage lighting on, viewed from the worst seat rather than from the production desk.
The short version
- Throw ratio is throw distance divided by image width, and it selects the lens.
- Doubling image width quarters the light landing on it.
- Lens shift keeps the pixel grid; digital keystone resamples and softens it.
- Front projection buys a shadow zone that grows with the angle.
- High gain screens take brightness from the side seats to give it to the centre.
- A smaller image usually beats stacking and blending a larger one.