Read a Set Plan and Elevation Before Describing the Commercial's Space

The PDF arrives with a note that says "see attached." You need a paragraph about the space in the treatment, and the paragraph has to describe a room that will actually be standing on the day. So you open the plan, find the doorway, and start writing.

That is where the description usually goes wrong — not because the plan is wrong, but because a plan is only a fraction of the drawing. It answers a narrow family of questions: where things sit along the floor, how far apart they are, what sits next to what. It is silent about height. It is silent about anything behind the camera. If the doorway in the plan is what a shot depends on, the plan has told you it exists and nothing about whether you can see over it.

An elevation is the other half. Read both before the paragraph, and the description gets smaller and truer: three or four supported facts, one or two named unknowns, and a short list of questions for the person who drew the sheet. That is a better set of sentences than the confident version, and it survives the production meeting.

Identify the sheet before reading the marks

Start with the block of text in the corner. Title, project, sheet name, sheet number, revision letter, date, scale, who drew it. Then the status word, if there is one: proposal, sketch design, issued for discussion, not for construction. Status is not a formality. It tells you what the drawing entitles you to say. On a proposal sheet you can describe what the design intends. You cannot describe what exists, and you cannot promise what will be built. A proposal is a position, not a commitment.

Revision letters matter for a dull reason and a dangerous one. The dull reason is that sheets get reissued and you want the current one. The dangerous reason is that it is entirely possible to be holding a current plan and a superseded elevation in the same folder, because someone exported them at different moments. Cross-referencing a B-revision plan against an A-revision elevation is a quiet way to invent a room. Check that the letters match before you match the walls.

Then find each view's label, and look for the arrows. A plan usually carries one or more of them, showing which way each elevation is looking. Without those arrows you are guessing which face you are looking at, and a guess here propagates through everything downstream.

Every sheet should also carry its own key: what the line weights mean, what the dashes mean, what is hatched, what has been left out. Use that key. Do not carry a symbol convention over from a previous job or from a textbook, because a dashed line on this sheet might mark something overhead, something concealed, or the edge of a detail the sheet simply does not cover — and which one it is belongs to this drawing, not to drawings in general.

One borrowed principle is worth keeping, with its limits attached. David Neat's illustrated glossary of technical drawing for theatre, film and television, published in 2014, covers ground plans, elevations, sections, scales and broken lines, and what it sets out is the habit of cross-referencing the views and going by the written dimensions rather than treating every printed line as a measured length. Two things sit alongside that. It is one practitioner's terminology, from one country and one period — not a universal standard, and no substitute for the key on your sheet. And a glossary describes drawing in general, while your sheet describes one room. Where they conflict, your sheet wins.

Scale lives on the sheet too, usually as a ratio printed under the title or as a ruled bar. That ratio tells you what a line is worth in the room. If you have the sheet at that scale — printed at 100 percent, or on a screen at a zoom you genuinely know — you can scale a distance off it, which is ordinary drafting practice for anything the designer did not write down. The moment the image has been resized to fit a slide, a chat window or a moodboard, that value is gone. A resized screenshot is an excellent thing to point at while you talk to the designer. It is not a measuring instrument, and no number should enter a treatment because it looked right on one.

Match the same feature across the views

Both drawings are orthographic. The view looks straight at the subject from a fixed direction, so parallel lines stay parallel and every part is drawn at the same scale. That is exactly why these views can be read as measurements, and why neither of them is a picture. A render or a sketch perspective in the same pack converges toward a vanishing point, which makes it useful for intent and useless for dimension.

The first move is dull and it saves the most time: choose one feature that appears in more than one view and follow it. A wall with an opening in it is the usual starting point, because most treatments end up talking about an entrance.

Find it on the plan. Note what it sits beside — a corner, another opening, a fixed fixture. Then find the elevation that shows that wall, using the label and the viewing arrow rather than your sense of which drawing looks like the plan. Then ask two questions separately, because their answers come from different sheets.

• Along the wall, where is it, and how wide is it? That is a plan question.
• How far up the wall does it go? That is an elevation question, and no amount of staring at the plan will answer it.

A plan is a horizontal cut through the space, which is why it has no height to give you. It can show you that a wall line is interrupted. It cannot show you the height at which the interruption stops. An elevation can show you that an opening reaches the floor or stops short of it, and it cannot show you a single thing about the depth behind that wall — everything in an elevation is flattened onto one plane.

When two views agree about a feature, you hold a fact you can write. When they seem to disagree, the problem is usually your identification rather than the building: you have misread a label, you are holding two revisions, or the discrepancy belongs to a detail sheet you have not been given.

Test the relationship the treatment needs

Treatments rarely describe objects in isolation. They assert relationships. You come in here and the counter is between you and the window. The doorway sits at the back of frame. The room opens out at the far end. Every one of those is a claim, and every claim has to be checked against the views that could support it.

Do it sentence by sentence. Underline what your draft says about space and mark which view supports each part. "The counter sits a meter back from the window" needs a plan. "The window has wall below it" needs the elevation of that wall. "You can see the whole room from the doorway" needs a camera height, a lens, and the height of everything between the doorway and the far wall — the set may give you none of those, and a low head over an opening near the camera can kill the sentence outright.

Keep one distinction visible in your own language. The drawing describes a room with a counter in it. The spot wants the counter at the back so a figure can walk past it and be seen doing so. Those are two different sentences, and only one of them is a description of the set. The other is a request to the design team, and it should be written as one.

Anything a sentence needs that the views do not carry becomes a question rather than an assumption. A sightline from a particular lens, what a wall can carry, whether a piece can be moved, whether an opening is permitted where the drawing puts it — those belong to the designer, not the writer, and they will not be settled by looking at the PDF harder.

An invented sheet, read twice

What follows is invented for the purpose of the lesson, and it is set out in words rather than as linework, so it is not a drawing you can examine. No set was surveyed and no designer reviewed it. The sequence of moves is the part worth keeping; you will run the same sequence on a real sheet.

The sheet, as invented: a title block naming the project, the sheet and its revision letter, marked as a proposal, with no construction issue. A ground plan. Two elevations, each labeled for a wall, with viewing arrows drawn on the plan. Two written figures across the whole set — 6.00 m on the plan, across the inside of the long wall, and 2.70 m on the elevations, floor to ceiling. Everything else has to be read.

Names below are sheet-relative, since nothing on the drawing says which way is north or which wall faces a street. The long wall at the bottom of the plan is Wall A. The right-hand wall is Wall B. Left is Wall C, top is Wall D.

On the plan:

• Wall A is interrupted by a gap about 1.8 m wide, starting about 1.2 m in from the left-hand corner.
• Wall B is interrupted near its top end by a gap about 0.9 m wide, leaving about 0.6 m of wall between the opening and the top-right corner.
• In the room, a rectangle about 2.4 m by 0.6 m runs parallel to Wall A, its near edge about 1.0 m up from that wall and its left end about 0.8 m from the left corner. A counter.

The distances other than the two written figures — 6.00 m on the plan, 2.70 m on the elevations — are scaled readings off the sheet at its stated scale, not written dimensions. That distinction will matter in a moment.

First reading, from the plan alone. The plan shows a room a little over six meters wide, with a large gap in the long wall and a second gap in the right-hand wall. This reads very easily, and it produces a paragraph like this:

A six-metre unit with a wide opening running the full height of the front wall, with a counter just inside it. A second opening at the back right leads through to the rest of the building.

Three things in that paragraph came from nowhere. "Full height" is a claim about height, and the plan does not carry height. "Front" supplies an orientation the sheet never gave. And "leads through to the rest of the building" describes what is on the far side of a wall when the drawing stops at the wall line. The counter is the one detail the plan did support, and even it got rounded into a relationship the sheet does not state.

Second reading, with the elevations. The elevation of Wall A shows a wall face six meters across and 2.70 m tall — those are the two written figures — with a rectangle cut into it left of center. Wall below it, wall above it. Its width and its position match the interruption on the plan.

That settles what the gap is: a window. Not a way through, not a full-height void. The plan gave you where it sits and how wide it runs. Only the elevation told you it does not reach the floor.

The elevation of Wall B shows its wall face with an opening that runs from the floor line up to a horizontal line well below the ceiling. So it is a way through: it reaches the floor, it has a head, and the head is not at ceiling level.

The counter appears in neither elevation. It stands in the room rather than against either wall, so the plan is the only sheet that shows it. That is the division of labor running in both directions, which is worth noticing once, because it stops you from asking the elevation to do a plan's job later on.

What is still open. The head height of the opening in Wall B. The elevation shows where the head line sits, and you could scale a rough figure off it, but a rough figure is not a dimension, and this is exactly the kind of number a shot can depend on. Also open: whether there is a leaf in that opening, since no swing appears on the plan. And open: what lies beyond Wall B. The sheet ends at the wall. It does not say whether that wall is internal or external.

The paragraph, revised.

The set is a single room, six meters along the wall that carries the window and 2.70 m clear to the ceiling. That wall is pierced by a window about a third of its length, set in from one corner, with wall above and below it — the room is lit through an opening rather than thrown open to the outside. A counter stands about a meter back from the window, parallel to the wall. Near one end of the adjoining wall there is a way through: it reaches the floor and stops short of the ceiling, and its head height is not dimensioned. What lies beyond that wall is not shown.

That paragraph is shorter than the first one and it says more, because every clause in it can be pointed at on a sheet.

Return a question, not an invented solution

The output of this reading is not a paragraph. It is a paragraph plus an annotation.

Annotate the exact thing. "Some dimensions aren't clear" is not a question, and nobody can answer it. "The opening in Wall B — is the head at 2.1 m, and is there a leaf in it?" is a question, and it comes back answered. Point at the wall, name the feature, ask for the missing view, dimension or status.

Ask for the smallest thing that unblocks the sentence. Sometimes that is a number. Sometimes it is a section — a vertical cut through the space, which will show heights and relationships that a plan flattens and an elevation isolates. If the pack has one, read it before you write anything about height. If it does not and height matters to your description, ask for one. Sometimes what is missing is not a drawing but a status: existing or proposed, retained or new.

Then hand the judgments back. Whether an opening can be built where it is drawn, how a wall takes load, how a camera mounts to a flat, how a counter gets through a doorway — those are structural, access and build questions. They are not answered by the sheet and they are not yours to answer. Passing them along as questions costs you nothing and keeps you from promising something on someone else's behalf.

Resist the blanket hedge, too. "Subject to design confirmation" attached to every sentence does not make a treatment careful; it makes it unreadable, and it hides the one thing you actually need. Put the uncertainty in one place, in plain words, beside the sentence it affects. A single "head height to be confirmed" reads as competence. Six of them read as a writer who did not open the drawings.

What you have when you finish is a short annotated reading: the sheet and its revision, the written dimensions it carries, the three or four spatial facts it supports with the view each came from, the points it leaves open, and the questions that follow. That reading tells you exactly how large a paragraph you are entitled to write.

For the invented room above, one number stands between the careful description and a specific one: the head height of the opening in Wall B. With it, you know whether a figure crossing the back of frame is visible above the opening or vanishes behind it — which is probably the thing the shot was about. Everything beyond that wall is a separate question, for a separate sentence, and it can wait until someone draws the other side.

Frequently asked questions

Why isn't a plan enough to describe a commercial set?

A plan answers only a narrow family of questions: where things sit along the floor, how far apart they are, and what sits next to what. It is silent about height and about anything behind the camera. A plan is a horizontal cut through the space. If a doorway in the plan is what a shot depends on, the plan tells you it exists and nothing about whether you can see over it. The elevation is the other half.

What should you check before reading the marks on a set drawing?

Start with the title block: title, project, sheet name, sheet number, revision letter, date, scale, and who drew it. Then the status word, because a proposal, sketch design, issued for discussion, or not-for-construction sheet tells you what the drawing entitles you to say. Check that revision letters match between plan and elevation, since a current plan and a superseded elevation can sit in the same folder. Find each view's label and viewing arrows, and use the sheet's own key rather than carrying over a symbol convention.

How do a plan and an elevation divide the work when matching a feature?

Choose one feature that appears in more than one view and follow it, often a wall with an opening. The plan gives where it sits along the wall and how wide it is. The elevation gives how far up the wall it goes. The plan cannot give height, and the elevation flattens depth onto one plane. When two views agree about a feature, you hold a fact you can write. When they seem to disagree, the problem is usually misreading a label, holding two revisions, or missing a detail sheet.

What does the invented sheet example show about reading the plan alone?

Read from the plan alone, the invented sheet produced a paragraph claiming a wide opening ran the full height of the front wall, with a counter just inside it, and a second opening led through to the rest of the building. Three things came from nowhere: full height is not carried by the plan, front supplies an orientation the sheet never gave, and leads through describes what is beyond a wall where the drawing stops. With the elevations, the Wall A gap becomes a window with wall below and above, and the Wall B opening reaches the floor but has a head below the ceiling. The counter appears only in the plan. The revised paragraph is shorter and every clause can be pointed at on a sheet.

What should you do when a sentence needs information the drawings don't carry?

Make it a question rather than an assumption. Annotate the exact thing, such as asking whether the opening in Wall B has its head at 2.1 m and whether there is a leaf in it. Ask for the smallest thing that unblocks the sentence: sometimes a number, sometimes a section, sometimes a status such as existing or proposed. Hand structural, access, and build judgments back to the relevant people. Avoid a blanket hedge attached to every sentence; put uncertainty in one place, in plain words, beside the sentence it affects.