> For the complete documentation index, see [llms.txt](https://docs.m-xr.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.m-xr.com/marso-studio/scenes/working-with-scene-results.md). # Working with scene results Once you've confirmed a selection, each generation stage runs across the whole group. This page covers what to expect while it runs, and what to do when part of a scene doesn't come back. **Group runs** The three generation stages — **Generate Reference Images**, **Generate 3D Model** and **Generate PBR Textures** — each run as a group across your whole selection, and each is charged per object. See [Credits & billing](/marso-studio/reference/credits-and-billing.md). Objects are processed in parallel, so a scene takes considerably longer than a single asset but not the sum of its parts. Larger selections take proportionally longer — plan for a long run on a big scene rather than watching it. You can review objects as they arrive. You don't have to wait for the whole group to finish before opening the 3D viewer on the ones that are already done. ### Partial completion A group can finish with some objects succeeded and some failed. This is normal on larger scenes — one object failing does not stop the rest, and does not invalidate the ones that worked. When a stage completes, check the group for anything that didn't come through. Failures are usually specific to one object rather than to the scene, and the most common cause is the reference image: * **The object doesn't fill enough of its reference image.** 3D generation needs the object to take up the majority of the frame. Objects that were small, distant or heavily occluded in the original scene produce references that are too sparse to reconstruct from. This is the failure to expect. * **A geometry the reconstruction can't resolve** — very thin, wiry or transparent objects are the common cases. * **A mesh without usable base colour** — PBR prediction needs a base-colour texture to work from. ### Retrying an object You can re-run individual objects rather than the whole scene. **Retrying charges only for that object and that stage** — you don't pay for the whole pipeline again. If a retry fails the same way twice, the problem is upstream. Go back a stage: a bad mesh is usually a bad reference image, and a bad reference image is usually an object that was too small in the scene to begin with. Running detection again and re-selecting is often faster than retrying the same object repeatedly — see [Detecting & selecting objects](/marso-studio/scenes/detecting-and-selecting-objects.md). ### Running out of credits mid-scene **Objects that already completed are kept.** They're yours, and you won't be charged for them again. But the run doesn't pause and pick up where it left off — **the remaining objects in that run won't complete**. Top up, then re-run the objects that didn't finish. To avoid it, work out the cost before you start — every object you keep is charged for all three stages. See [Credits & billing](/marso-studio/reference/credits-and-billing.md). ### Exporting a scene You can download: * **The whole scene** — every completed object in one download. * **A single object** — any individual asset on its own. * **Textures only** — the individual PBR maps for the objects you select. Objects that didn't complete are skipped rather than exported empty. Full format details are in [Previewing & exporting](/marso-studio/reference/previewing-and-exporting.md). Each object exports as its own asset with its own material set, so you can place, swap and reuse them independently once they're in your DCC. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.m-xr.com/marso-studio/scenes/working-with-scene-results.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.