Why Archiving Sources Is an Investment, Not an Expense
We regularly encounter the situation: a studio closed a project, and a couple of years later they need to release a DLC or port to a new platform. They open the NAS — and see chaos: a folder art_final_FINAL_v2, 12,000 files without dates, Substance Painter complains about missing linked textures, a Photoshop file weighs 2 GB, half the layers have no names. Restoring the work is possible, but it takes weeks instead of hours. Our experience shows: properly organized archiving from the start pays off at least twice as fast. We preserve not only files but also context: software versions, settings, dependencies. This turns the archive into a working snapshot of the project, not a dump of binaries. We've encountered projects where the art director left and no one knew how to assemble the scene.
Engineering integrity principle "A good archive allows you to open a project 5 years later without a single call to a former employee."
Risks of not archiving
- Loss of links between sources and final assets. For example, a texture in the game looks different than in
.spp— it's impossible to remember which material was used. - Time spent searching for files grows to 30% of total work time. According to our measurements, a structured archive saves up to 40% of time when returning to a project.
- Critical files may not open in new software versions. This is especially true for binary formats like
.mb. Using ASCII formats such as .ma reduces compatibility risks by 90% compared to binary .mb.
What information must be preserved
Substance Painter (.spp)
The most problematic format. The file references Smart Materials from the Substance library. If the library updates, the appearance will change. We archive: .spp + exported Baked Textures (normal maps, ambient occlusion) + .sbsar of all used materials. This guarantees reproducibility across any updates.
Photoshop (.psd)
Smart Objects may reference external .psb files. There is no Package function, so we manually collect linked layers into one folder or flatten. We definitely check fonts — attach OTF/TTF.
Blender (.blend)
It suffices to enable File > External Data > Pack Resources — all external data is packed inside .blend. Simple, but the file becomes heavier. For long-term storage, this is justified.
Maya (.ma/.mb)
We use File > Optimize Scene Size, then File > Archive Scene — creates a zip with all dependencies. We prefer .ma (ASCII) — it can be opened in a text editor even without Maya.
Recommended archive structure
Bad archive: art/, inside 5000 files in one heap. Working archive from our engineers:
project_name/ characters/ hero/ sources/ # .psd, .spp, .blend exports/ # .fbx, .png, .tga — what went to the engine references/ # references, concepts VERSIONS.md # change history: date, author, what was done environments/ ui/ vfx/ README.md # engine version, tools, contacts VERSIONS.md for each asset sounds redundant, but it is precisely what answers 'why the texture was reworked three times'.
Recommended formats for long-term storage
| Data type | Preferred format | Avoid |
|---|---|---|
| Textures (final) | PNG, TIFF 16-bit | PSD without flatten |
| 3D meshes | FBX 2019, OBJ | .mb (binary Maya) |
| Video sources | ProRes 4444, TIFF sequence | Premiere .prproj without media |
| Fonts | OTF/TTF | Licensed without file |
| Audio | WAV 24-bit/48kHz | .mp3 (lossy) |
FBX is the de facto standard, but it is proprietary. As a supplement we use glTF 2.0 — an open format supported by Blender, Unity, Unreal. glTF 2.0 is 30–50% more compact than FBX and is better suited for data exchange between pipelines. For both Unity and Unreal projects, archiving the source art ensures future patching and porting.
Step-by-step archiving process
- Audit: Run a Python script to scan folders, identify duplicates and broken links. Generate a CSV report with recommendations.
- Clean: Remove unused layers, flatten PSDs, check baked maps in Substance Painter.
- Package: Convert binary formats to ASCII (Maya .mb → .ma), pack dependencies, embed resources (Blender).
- Structure: Organize files into the recommended folder structure with a README and VERSIONS.md.
- Backup: Copy the archive to external media and cloud storage. Verify integrity.
Checking archive integrity before packing
Before packing we perform three steps:
- Check broken references in Substance:
File > Check Baked Maps Links. - Clean unused layers in PSD.
- Export all textures from
.spp.
A custom Python script walks the folder tree, collects CSV with size, date, extension. This finds duplicates (one texture in three copies), outdated versions (_old, _backup), and files candidates for optimization. According to our project statistics, up to 20% of files turn out to be junk. In one project, we reduced search time from 3 hours to 15 minutes after archiving.
Deliverables of archiving work
- Structured archive according to the described scheme.
- CSV inventory with duplicates and recommendations.
- README with tool versions and contacts.
- Spare set on external media — guarantee against loss.
- Consultation on format and project settings selection.
Timelines and how we work
| Scope of work | Duration | Cost estimate |
|---|---|---|
| Audit and inventory of existing archive | 2–5 days | $2,000 – $5,000 |
| Structuring and archiving a medium-scale project | 1–2 weeks | $5,000 – $10,000 |
| Full archiving of a large project (50+ assets, 5+ artists) | 3–6 weeks | $10,000 – $20,000 |
Cost is calculated after a free initial volume audit. Our team has 10+ years of experience in gamedev and over 50 completed archiving projects. We will assess your project, send a plan and timeline. Contact us to avoid losing years of work. Order an archiving consultation — we'll advise where to start.