Architecture Video Glossary: Walkthrough vs. Flythrough and More

A brief that says "flythrough" and a spec sheet that says "safe zone" and "9:16" are describing three different parts of the same project video, and most lists online only define one of those worlds at a time. This glossary puts the traditional animation vocabulary, the social-platform vocabulary, and the newer AI image-to-video vocabulary in one place, grouped by type, so a term like architectural walkthrough vs flythrough sits next to first frame and safe zone instead of in a separate article. Each entry is one short paragraph; where a term has more to say, it links to the deeper guide.

Key takeaways

• Walkthrough and flythrough are traditional animation terms describing camera height and scope, not AI video formats.
• Social-platform terms like safe zone, 9:16 and the five-hashtag limit come from Meta and YouTube's own current specs, cited below.
• AI production terms — image-to-video, first frame, reference image, scene plan — describe how an image-to-video tool builds a film from still images rather than a 3D model.
• A handful of terms (project video, free video, allowance, extra video) are specific to one product's plans and worth knowing before you compare pricing.
• Where a term connects to a bigger decision — cost, disclosure, hashtags — this glossary links to the article that covers it in full.

Traditional animation terms

• Walkthrough. An architectural walkthrough animation moves through a building's interior at roughly human eye height, following real-world constraints like doors, stairs and corridors. ArchiCGI's comparison lists a walkthrough's focus as "Interior spaces, room-by-room flow," shot from "Ground-level, human eye height," with a typical runtime of 1 to 3 minutes, as of September 2026. Use it when a design review needs to follow room-to-room flow — a clinic's waiting area into its corridor into an exam room.
• Flythrough. A flythrough moves the camera above and around a building, exterior and interior both, from high or aerial angles a person could never stand at. The same comparison describes a flythrough's focus as "Full property overview, exterior + interior," shot from aerial and dynamic angles, also typically 1 to 3 minutes. Use it to establish a site or master plan — a coastal retreat's full parcel — before cutting into interior detail.
• Turntable. A turntable animation holds the camera mostly still and rotates the building, a furniture piece, or a physical model through 360 degrees. It suits a single object or a compact site plan better than a full building tour, since there's no traveling camera to plan.
• Architectural animation. The umbrella term for both of the above. Wikipedia's definition puts it plainly: "Architectural animation is a short architectural film or video created using computer-generated imagery," and unlike a rendering — "a single image from one point of view" — an animation "is made from a sequence of still images or rendered frames played in order to create motion."

Camera and edit terms

• Establishing shot. The opening wide shot that orients a viewer to a site or building before the film moves closer, common as the first scene in both a flythrough and a short AI film.
• Dolly. A camera move that travels physically through or alongside a space, rather than staying fixed — the shot type that gives a walkthrough its sense of forward motion.
• Pan and tilt. A pan rotates the camera horizontally from a fixed point; a tilt rotates it vertically. Both reveal more of a scene without the camera actually moving through space, useful for a sweeping shot of a courtyard garden from one spot.
• Cutaway. A brief shot away from the main sequence — a material close-up, a detail insert — that breaks up a longer animation or film without disrupting its overall path.
• Shot list. The pre-production document naming every shot a project needs before filming or rendering starts, whether that shoot is headed for a traditional animation or a set of source photos for an AI video.

Social-video terms

• 9:16. The vertical aspect ratio — taller than it is wide — used by Instagram Reels, TikTok and YouTube Shorts. Meta's Instagram Reels ad specifications recommend 9:16 at 1440×2560 pixels, as of September 2026, though most studios export and post at 1080×1920.
• Safe zone. The margin near a vertical video's edges to keep clear of text and logos, since a platform's own interface — captions, profile icon, action buttons — sits over that space. The same Meta specifications recommend keeping "at least 14% of the top, 35% of the bottom, and 6% on each side" free of key content; Instagram Reels for architects walks through what that means for a real caption.
• Hook. The opening one to three seconds of a short video, built to stop a scrolling viewer before they swipe past, usually the single strongest frame or moment available rather than a slow build.
• AI info label. Meta's disclosure label for AI-generated or AI-edited content, renamed from "Made with AI" to "AI info" on July 1, 2024, applied "through industry standard indicators" or a creator's own disclosure (Meta newsroom). AI rendering disclosure for architects covers how that applies across platforms.
• Five-hashtag limit. Instagram's current guidance, announced December 18, 2025, capping a caption at five hashtags, quoting Instagram that "using fewer (up to 5) more targeted hashtags... can improve both your content's performance and people's experience" (Social Media Today). Architecture hashtags for Instagram has a five-slot formula for using them.
• Shorts length. YouTube's rule that any video with a square or vertical frame, up to three minutes long, uploaded on or after October 15, 2024, is categorized as a Short automatically (YouTube Help), whether or not that was the intent.
• Reel. Instagram and Facebook's short-form vertical video format and the dedicated tab where it appears, distinct from a regular feed video post.

AI and production terms

These terms describe how an AI image-to-video tool plans and renders a film from still images rather than a 3D model. The specific pixel and megapixel figures below, and the disclosure under "Consistent cast," reflect Arch2Video's own product, as of September 2026.

• Image-to-video. An approach that generates motion directly from existing still images — renderings, drawings or finished-project photographs — instead of building a 3D model first. Turning architectural renderings into video covers how that process works in practice.
• First frame. The starting image an AI video model treats as the base for one generated scene, anchoring that clip's composition and lighting.
• Reference image. An image supplied to an AI model to keep style, a building, or a character consistent across scenes, distinct from a first frame, which anchors one clip's motion specifically.
• Consistent cast. A recurring set of fictional people an AI video tool generates across a film's scenes so it reads as one continuous story; here, any real people visible in your source photos aren't animated and are replaced by that fictional cast instead.
• Scene plan. The sequence and content an AI director decides for each clip in a film before generation starts, including which of the source images becomes which scene.
• Long side. The longer of an image's width and height in pixels, used to describe minimum resolution regardless of whether a photo is oriented horizontally or vertically; the recommended minimum here is 1,500 pixels, with a hard floor of 512.
• Megapixel. An image's total pixel count, in millions, calculated as width times height; a high megapixel count from a modern camera can exceed a tool's decode ceiling — 120 megapixels here — before the file is even resized. Choosing project images for architecture video covers what actually makes a source image usable beyond its pixel count.
• H.264. The video compression standard used to encode most web-delivered MP4 files, balancing visual quality against file size for playback on a phone or browser.
• AAC. Advanced Audio Coding, the compressed audio format typically paired with H.264 video inside an MP4 file, used for a film's soundtrack.
• 24 fps. Twenty-four frames per second, the traditional cinematic frame rate still used for most short-form architecture and marketing video today.
• Faststart. An MP4 encoding option that moves key file metadata to the front of the file so a browser can begin playback before the rest has finished downloading — relevant for any video embedded on your own site.
• VideoObject. The schema.org structured-data type search engines use to understand a page's video. Google's documentation requires a name, thumbnailUrl and uploadDate, and states plainly that "VideoObject structured data must be added to a page where users can watch the video."
• Watch page. The page on your own site where a visitor can actually press play on a video, as opposed to a page that only links out to YouTube or Instagram — the page VideoObject markup belongs on.

Plan and account terms

These six terms come from Arch2Video's own plan and account pages.

• Project video. The basic unit of output: one roughly 30-second film generated from six images of a single project, planned as six connected five-second scenes.
• Free video. The one project video every account can create before subscribing, made exactly like a paid one but carrying a small arch2video.com watermark; subscribing to any plan re-exports it without the watermark at no extra cost.
• Allowance. The number of project videos included in a subscription each billing period — four on the Studio plan (USD 49.90 a month) and eight on Practice (USD 99.90), as of October 2026.
• Extra video. An additional project video bought beyond the monthly allowance for USD 30, available to an active subscriber at any time; unused extra videos carry over to later billing periods while the subscription stays active.
• Studio profile. A saved studio or brand profile — up to two per account — storing identity details and a logo so they don't need re-entering on every project.
• No-logo version. The second copy of a film, without the studio logo overlay, generated automatically alongside the logo version whenever a film is exported with one.

Frequently asked questions

What is the difference between an architectural walkthrough and a flythrough?

A walkthrough moves through a building's interior at roughly human eye height, following real routes like doors and corridors; a flythrough moves above and around a building, exterior and interior both, from high or aerial angles. Both typically run 1 to 3 minutes, as of September 2026, and the choice usually comes down to whether the priority is interior sequencing or a full-site overview.

Is an AI-generated film a walkthrough or a flythrough?

Neither, strictly. A walkthrough or flythrough renders a continuous camera path through a 3D model; an image-to-video film like the ones Arch2Video produces plans six independent five-second scenes from your own images and edits them into one short film, without a model to move a camera through. It can borrow the visual grammar of either — an interior scene, a wide establishing shot — without measuring the building the way traditional animation does.

What video format should I use for social media?

For Instagram Reels, TikTok and YouTube Shorts, that's 9:16, commonly exported at 1080×1920, encoded as H.264 with an AAC audio track at 24 fps. Keep text and logos out of the safe zone near the edges, since the platform's own interface — captions, profile icon, buttons — covers that space on a phone screen.

Do I need to label an AI-generated architecture video?

Under Meta's current policy, yes for organic posts with photorealistic AI video or realistic AI audio; the "AI info" label applies through industry signals or your own disclosure when you upload. Our guide to AI rendering disclosure for architects covers the wording and where it applies beyond Meta's platforms.

How many hashtags should an architecture post use now?

Five, as of Instagram's December 2025 guidance limiting a caption to that many hashtags; a longer list isn't an option anymore, gradually rolled out across accounts. Our architecture hashtags guide has a five-slot formula for choosing them by discipline, project type, material, place and community.