After the 3D aerial rendering is sent out, the client asks in the group chat, “Which buildings are in this launch?” Is it still in time to go back and regroup them?

Conclusion first: in most cases, it is worth separating them during modeling. By the time the client asks in the group chat, “Which buildings are in this launch?”, going back to regroup them is usually already late. When the subject buildings and surrounding existing buildings blur together in an aerial view, the common cause is usually not insufficient rendering color grading; more often, the modeling stage did not separate the subject, surrounding context, and existing context into groups that can be adjusted independently. Based on typical 2026 delivery experience, adding one more grouping layer during modeling has an experience range of half a day to one day; if the client says they cannot tell them apart after the image is produced and rework is needed, materials often have to be reassigned and the primary view re-rendered, with a typical cost range of 1 to 3 days. For images that will be passed around by multiple people and will need repeated range annotation, such as marketing displays, phased development, and planning approval presentations, it is generally worth grouping first; for process drafts used only for internal massing discussion where the scheme may be overturned at any time, it can be left ungrouped for now.

Why “cannot tell which buildings are in this launch” mostly starts in the modeling stage

Distinguishing the subject in an aerial view relies on the overlay of hue, lightness, saturation, roof materials, and entourage density, not on pulling a few curves in post-production. What can mainly be changed at the rendering stage is hue and lightness; object assignment, material assignment, and layer grouping are already fixed during modeling. If the subject buildings and surrounding buildings share one material sphere and sit in the same layer, at the rendering stage you can only adjust them as a whole block; changing one building moves a whole area, and the more you adjust, the messier it gets.

A more hidden layer is entourage density. If trees, cars, and people around the subject and in the existing context area are equally dense, the human eye will assume by default that they belong to the same category. In real aerial photography, people distinguish new and old buildings by warm/cool color, lightness contrast, and differences in roof materials; these all need to be prepared separately during modeling, and expecting post-production color grading to save them is usually too late. The client’s attention time on an aerial view is typically only 3 to 5 seconds; buildings that cannot stand out in those seconds are effectively not drawn.

• Rendering can only change hue and lightness, not object assignment; material spheres and layers are already fixed during modeling.
• Sharing one material sphere between the subject and surrounding context is a common reason aerial views blur into one mass.
• Making entourage density the same across the three layers visually groups areas of different nature into one category.
• Judging whether the grouping works by “something feels off” often leads to repeated revisions.

How should the layers be divided: three-layer grouping plus four-step color

Grouping is not for looking good; it is for making the subject pop out within a few seconds. Based on typical 2026 project delivery practice, first divide the scene into three layers, then follow four color steps; revision counts are usually lower. The three layers are divided by land-use nature and development sequence, not by distance.

1. Layer 1, subject: the buildings and site to be sold or displayed this time. Use warm colors or high lightness, give roof materials their own set, and provide full entourage density as needed for the primary view.
2. Layer 2, surrounding planning: other phases of the same project and known planned land. Use neutral gray or low saturation, downplay them, and retain massing relationships without competing for attention.
3. Layer 3, existing background: existing urban buildings, mountains, and water systems. Use low-saturation cool gray or desaturate directly, and darken them overall.

After the three layers are set, follow the four color steps: first set the main hue, then widen the lightness difference, then unify entourage density, and finally do one layered color correction in post-production. When setting the main hue, do not choose a tone that clashes with the surrounding context; when widening the lightness difference, it is recommended that the subject and surrounding context differ by at least one stop, or they will be flattened if too close; entourage density should decrease by layer, dense for the subject and sparse for the existing context; post-production correction should be done on layered files, not with a one-size-fits-all adjustment to the whole image.

• Lightness difference holds up better than hue difference: when shrunk to a phone screen or converted to grayscale printing, hue difference fails first, while lightness difference remains.
• Decreasing entourage density by layer is a relatively low-cost way to separate layers.
• Whole-image color grading in post-production is where grouping easily fails; doing it by layer is recommended.
• Do not scramble the order of the four steps; changing density before color matching will waste a round of adjustment.

How modeling, rendering, and post-production hand off to each other

Choose modeling tools by project scale: for small-scale communities, SketchUp plus standard plugins is enough; for projects with large terrain undulation and many curved surfaces, Rhino or 3ds Max is commonly used; when the scheme needs to be changed while checking lighting, Blender can be used as a base first. On the rendering side, 3ds Max with V-Ray or Corona is a common combination; when multiple angles need to be produced quickly, real-time engines such as UE5 and D5 save more time. In post-production, do color correction and range annotation by layer in PS.

What is easily overlooked is naming and units. Unifying model units to meters and aligning coordinate pivots are basic requirements; layer names should carry prefixes, for example MZ_ for the subject, ZB_ for the surrounding context, and XZ_ for existing context, so the rendering side does not have to guess when taking over. Material spheres should likewise be divided into sets by layer, with similar objects sharing one sphere, so color changes can be completed in one pass. Writing these items into the handover sheet can save a large amount of back-and-forth confirmation.

• Unify units to meters and align pivots, otherwise camera height and shadow direction will both go wrong.
• Layer names with prefixes are a relatively hassle-free convention when modeling and rendering are handled by different parties.
• Unify shadow direction according to true north at the project location; do not cast shadows differently in each image.
• Set camera height and focal length before grouping, to avoid the view changing after grouping is done.
• Material sphere names should also carry prefixes, which makes finding objects much faster in post-production.

At delivery: what if the surrounding context only has a satellite image and a few phone photos?

A common constraint is this: the project is developed in phases, the client provides only one satellite image plus a few site photos taken with a phone, the budget is tight and does not allow full existing-context modeling, but the aerial view still must show which buildings are in the current phase. In this case, investment is usually split into two levels—the surrounding context is modeled only as massing plus roof color, with no windows or facade details, and footprint and floor count are measured from the satellite image; the subject buildings are modeled to window-to-wall ratio, roof equipment, and entrance elements, with entourage placed only on the subject side. The result was that the client circled the current-phase range in the first round and did not enter a second round of revisions.

The opposite has also happened: to save time, the subject was also flattened into massing only, and the client’s first sentence after looking was, “This has nothing to do with my project.” Then the primary view had to be redone and materials reassigned, costing an experience range of 2 to 3 extra days; if it hits a presentation deadline, that means delay. The boundary for staged investment is usually not in the modeling software but in what the image will be used for—if it is for sales, the subject cannot be cut; if it is only for internal discussion, the subject can be cut. By typical range, making the modeling detail of the subject one to two levels higher than the surrounding context in a single aerial view is a common practice that the client accepts and that keeps costs under control.

Layered modeling and post-production color grading have different cost structures

Both paths can separate the subject, but the money spent and the cost of revisions are not in the same place:

• Layered modeling: about 0.5 to 1 extra day upfront; later changes are made by layer, with a typical range of 0.5 to 1 day per revision round; suitable for projects that need multiple views, multiple versions, and phased adjustments.
• Post-production outlining and color grading: no extra time upfront; revisions are fixed image by image, with a typical range of about 0.5 day per image; suitable for projects that need only one image and will basically not change afterward.
• Mixed approach: layer the subject in modeling and darken the existing context in post-production; this is a relatively common compromise in 2026, with mid-level upfront cost and mid-level revision risk.

To choose which path, working backward from the expected number of revision rounds is relatively reliable: if two or more revision rounds are expected, the upfront grouping will basically pay for itself; if only one image is needed and the client wants final approval the same day, it is reasonable not to invest upfront. There is no universal answer for this kind of judgment; the experience ranges above are for reference only, and the specific choice also depends on project milestones and the client’s review habits.

When layering is needed and when it is not

Scenarios where layering is worth it: aerial views used for marketing displays, sales office materials, or presentations to the client or competent authorities; projects developed in phases that need the current-phase range clearly marked on the image; and cases where the same base image must produce multiple views or versions. In these scenarios, the extra half day to one day is usually recovered within one revision round.

Cases where layering is not necessary should also be stated clearly: for images used only for internal massing studies while the scheme is still in the comparison stage, fast output matters more; for projects with only one building and basically blank surroundings, there is no comparison object for layering, so doing it is idle work; for cases where the scheme will be changed right after the image is produced and the image itself is a process draft, upfront investment usually cannot be recovered.

• Suitable: marketing displays, phased development, planning approval presentations, and multiple views and versions.
• Can be left undone: internal massing studies, a single standalone building, and clear process drafts.
• Boundary reminder: layering solves recognizability, not basis issues such as scale, north arrow, and building spacing measurement; those need to be checked separately in the output settings.

Frequently asked questions

How many days does a 3D aerial rendering usually take?

The experience range is 3 to 7 days per image, including layered modeling and two revision rounds; when the scale is large and surrounding existing context must be added, the typical range is 7 to 12 days, with rush delivery priced separately.

If the surrounding existing context only has a satellite image and phone photos, is that enough for grouping?

It is enough for massing-level grouping. Keep only massing and roof color for the existing context, without windows or details, and model the subject fully; by experience, the client can still point out which buildings are in the current phase in an image produced this way.

Does it cost more to hire separate people for modeling and rendering?

Not necessarily, but layer naming and material sphere conventions must be written into the handover sheet in advance; otherwise the party taking over often has to rename and reassign materials, adding an experience range of half a day to one day.

If the client later wants a night version, will the grouping still hold?

Yes, provided that material spheres are divided into sets by layer and lights are divided into collections by layer; with the same grouping structure, switching to a night scene only requires changing lights and exposure, and the typical range is 1 to 2 days to produce.

What files are usually delivered for an aerial rendering?

Typically a JPG or PNG bitmap plus a layered PSD file; when vector annotations are needed, a PDF is provided separately, and whether the source model is delivered depends on the contract terms.

If the project is developed in phases and the current-phase range must be clearly marked on the aerial view, it is recommended to write the three-layer grouping and naming prefixes into the handover sheet before modeling starts; this item is usually listed as first-round confirmation content. If it is only for internal massing studies and the scheme still needs changes, putting the budget into fast output first is more cost-effective—layering solves recognizability, while measurement and planning approval basis issues outside that boundary still need separate checks.