What a Photo Composite Needs Before Production

Composites are unusual among visualization products in that most of the decisions that determine the result are made before any production work begins, and several of them cannot be revisited afterwards at any price.

A rendering can be re lit, re angled and re composed at the cost of render time. A composite is tied to a photograph that was taken at a particular moment from a particular place, and if that photograph was wrong, the only remedy is going back to the site.

This is the list of what has to exist first, ordered by how expensive it is to have skipped.

A decision about what the image is for

The same site, the same building and the same viewpoint produce different work depending on whether the image is marketing, submission material, or evidence in a contested setting, and no vendor can infer which from a brief that does not say.

Marketing composites choose the best available angle and the best light, and nobody is misled by a good day. Submission material has to survive being examined by somebody who knows the site. Evidence in a contested setting has to survive somebody actively looking for a reason to reject it.

The differences are not cosmetic. They change viewpoint selection, whether survey is required, what accuracy level is appropriate, and what documentation has to be retained. Deciding at the start costs nothing and deciding late means producing the work twice.

The viewpoints, chosen and justified

This is the decision with the longest reach, because photography follows it, survey follows it, and the modelling detail follows from how close the closest one is.

The positions that recur are the approach where the proposal first becomes visible, the point opposite where the height relationship reads most clearly, any distant vantage where it appears against a skyline, and any position made important by a review body, a neighbour group or a public space.

Writing down why each was chosen is a small discipline with a disproportionate benefit. If the images are ever questioned, a documented basis for selection is a very different position from a set of angles somebody liked.

It also prevents the most common late request, which is somebody asking for the view from the one place nobody photographed, months after the site visit.

The photography, or a decision to commission it

A project either has usable photographs or it needs a site visit, and the difference is frequently discovered late because everybody assumes the existing images will do.

Existing photographs are usable when they were taken from a relevant position, in appropriate conditions, at a known focal length, with the ground floor visible and the season correct. That combination is rarer than expected.

Marketing photography taken for a brochure is almost never usable for a composite, because it was shot for atmosphere: wide angles, dramatic light, deliberate cropping, and frequently from positions chosen because they look good rather than because they matter.

Commissioning photography specifically for composites is a small cost and it removes an entire category of compromise, and it has to be scheduled early enough that a return visit is possible if the day was wrong.

Survey control, if the images will be relied upon

The survey has to be arranged with a surveyor, coordinated with the photography, and ideally executed in the same visit or immediately after while nothing has changed.

What it produces is the camera position in three dimensions and control points visible in each frame that the render can be aligned to. That is the difference between a composite that is internally consistent and one that is tied to measured reality.

The sequencing matters more than the cost. Surveying positions after the photography means somebody has to identify exactly where the photographer stood, which is imprecise. Doing both together, or surveying first and photographing from marked positions, removes that problem entirely.

A confirmed drawing set

The building has to be modelled from something, and that something has to be current and confirmed rather than whatever version was circulating.

Floor plans, elevations for every face that will be visible from the chosen viewpoints, sections establishing floor to floor heights, and a roof plan. Materials at least to the level of what is reflective, what is matt and what colour each surface is under daylight.

Where a set is still developing, the workable approach is to produce at massing accuracy first, confirm the camera match and the composition, and add resolution once the design settles. That sequence wastes nothing, whereas modelling in detail from a superseded elevation wastes all of it.

Site levels, especially on any slope

A composite places a building on ground that appears in the photograph, and if the model sits on an assumed flat plane while the real site falls away, the building will be at the wrong height in the frame.

Ground levels from survey resolve it. On genuinely flat sites the error is small and tolerable. On any site with a fall across it, or with a retaining condition at the boundary, it becomes the dominant error and it lands directly on apparent height.

This is also the item most often assumed to be included and most often not, because it belongs to the survey scope rather than the visualization scope.

The weather window, which is a scheduling constraint rather than a preference

Composites have a dependency that no other visualization product shares: they require a specific kind of day, at a specific time of year, and no amount of budget compresses that.

Bright overcast is technically the easiest condition because the light is diffuse, there are no hard shadows to reconcile between render and photograph, and the facades of surrounding buildings read with even tone. Strong directional sun is more dramatic and considerably harder, because every new shadow has to agree exactly with every existing one.

Season matters just as much. A site screened by street trees in July is a site plainly visible in January, and the same viewpoint produces images that support opposite conclusions. Choosing which is honest depends on what the image is for, and having only one available means the choice was made by the calendar.

A project that plans composites three months ahead can photograph on the right day and return if the first attempt failed. A project that decides two weeks before a deadline photographs in whatever conditions exist, and every later stage inherits that compromise permanently.

What the client has to supply that is not a drawing

Beyond the documents, a composite package needs a small amount of information that lives in people rather than in files, and gathering it early prevents the most awkward late corrections.

Which neighbouring buildings are protected or sensitive, because those are the relationships the image will be judged on. Which parts of the surrounding context are themselves proposed or approved and may exist by completion. Whether any adjacent site has a live application that changes what the view will look like.

Also, and this is frequently unknown to the visualization team, whether there is a history to the site. A previous refused scheme, a local campaign, a specific objection that recurs. Any of those tells the producer which relationship the image most needs to address honestly.

None of it appears in a drawing set and all of it changes what the composite should show, which is why a conversation with somebody who has been on the project since the beginning is worth more than another round of file transfers.

An accuracy level, stated

The discipline has established levels and naming one converts a subjective brief into a specification.

A wireline outline establishing location and size. Resolved visibility showing what obscures the scheme. Massing accurate architectural form. Full materials and rendering. All on the same camera match, with only the finish changing.

Stating the level in the brief means the vendor prices the right thing and the client receives the right thing, and it removes the revision cycle that occurs when a client expected photorealism and a vendor quoted for massing, or the reverse.

The existing conditions that are about to change

Sites change between photography and completion, and the image has to be honest about which changes it assumes.

A building scheduled for demolition, a hoarding that will come down, temporary structures, a road layout being altered. Each is a decision about whether the composite shows the world as photographed or the world as it will be.

Both are legitimate and only one is honest without a note. Removing something that will genuinely be gone is reasonable if it is stated, and removing a permanent neighbour because it is unattractive is not.

Deciding this before production avoids the awkward late conversation where somebody asks whether the ugly building next door can be taken out.

Who reviews it, and when

Composites benefit from a specific review sequence and suffer when it is compressed into one approval at the end.

Viewpoints should be approved before photography, because that is the only moment they can change for free. The camera match should be approved on a simple massing block before detailed modelling, because if the match is wrong everything built on it is wasted. Lighting agreement should be checked against the photograph before final rendering.

A single review at the end catches only the things that are cheapest to fix and misses the ones that would have been free to fix earlier.

Somebody who knows the street

Worth naming as a requirement rather than a nicety, because it catches a category of error that no technical process addresses.

A person who walks past the site regularly will notice within seconds that the light never falls that way in the afternoon, that the pavement is never empty, that the tree is larger, that the neighbouring facade is a different colour than it appears.

None of those are process failures and all of them damage credibility with exactly the audience that matters in a contested proposal. Arranging one such review before publication costs nothing.

The documentation that will be kept

Decide before production what will be recorded and retained, because reconstructing it afterwards is unreliable and nobody does it.

Camera position, height, direction, focal length, date and time. Whether the position was surveyed and by whom. The accuracy level. The original unmodified photograph.

That set turns an image into a record of what was represented and when, which is what settles a later question about fairness. Without it, a challenged image can only be defended by argument.

A realistic sequence

Decide purpose and viewpoints. Arrange survey and photography together, with enough calendar margin to return if conditions are wrong. Confirm the drawing set and site levels while that is being scheduled.

Then produce a massing block through the matched camera and approve the match before any detail is built. Then model to the detail the closest viewpoint requires. Then light to the photograph rather than to preference. Then integrate.

Legal or factual review runs against the images as they complete rather than as a block at the end, since a single objection late can hold an entire package.

What each missing item costs

No stated purpose: the wrong product is produced competently.

No agreed viewpoints: a request arrives for the one position nobody photographed.

No commissioned photography: the composite inherits whatever position, light and season an old marketing image happened to have.

No survey: the image is plausible and cannot be defended, which matters only when somebody challenges it, and that is exactly when it matters most.

No site levels: the building sits at the wrong height, which is the one dimension every objection is about.

No accuracy level: a revision cycle that feels like a taste dispute and is actually a specification failure.

No documentation: an image that can be argued about but not checked.

Frequently asked questions

Can we use photographs we already have?

Sometimes, but rarely. They need to be from a relevant position, in appropriate conditions, at a known focal length, with the ground floor visible and the right season. Brochure photography is almost never usable because it was shot for atmosphere from positions chosen to look good.

Should the survey happen before or after the photography?

Together, or survey first with the photography taken from marked positions. Surveying afterwards means somebody has to identify exactly where the photographer stood, which is imprecise and reintroduces the error the survey was meant to remove.

What if the drawing set is still developing?

Produce at massing accuracy first, confirm the camera match and the composition, then add resolution once the design settles. That wastes nothing, while modelling in detail from a superseded elevation wastes all of it.

When should the client review the work?

Viewpoints before photography, since that is the only free moment to change them. The camera match on a simple massing block before detailed modelling. Lighting against the photograph before final rendering. A single review at the end catches only the cheapest problems.

Can we remove things from the photograph?

A structure that will genuinely be gone by completion, yes, and it should be stated. A permanent neighbour because it is unattractive, no, because it destroys exactly the credibility that made the format worth choosing. Decide this before production rather than during.