Definition. An output-format decision determines what the next application receives after document conversion. OfficeAPI lists PDF and image outputs for the Word, Excel and PowerPoint families. A PDF document, a bitmap image and a vector image have different roles even when they originate from the same office file.
Which output choices are documented?
The 3 office families list 2 documented PDF variants and 2 image groups. Grouped rows should be read carefully: they establish listed formats, rather than a guarantee about every possible setting of those formats.
| Output group | Documented formats | Decision to evaluate |
|---|---|---|
| Vector PDF, image PDF | A document container for a PDF reader | |
| Bitmap images | JPEG, PNG, BMP | Pixel-based imagery for a receiving surface |
| Vector image group | SVG, TIFF, EMF | Listed image outputs with differing format characteristics |
The product matrix labels SVG, TIFF and EMF together as a vector-image group. TIFF is a raster image container in general; the grouping should not be interpreted as a statement that TIFF is inherently vector. Evaluate the actual format your recipient accepts.
What should drive the choice?
Start with the consumer. A user downloading a document and a UI showing a thumbnail do not necessarily need the same result. If the task is only in-app viewing, consult the preview formats before adding a conversion step.
For a local model workflow, use only the formats accepted by that particular model and application. Conversion supplies a file; the model's interpretation and the application's data handling remain separate responsibilities. The on-device AI guide explains that boundary.
How do you evaluate an output format?
- Identify the source's office family and find the target output in the conversion matrix.
- List the receiving surface's requirements: document container, page imagery, readable labels or data fields.
- Produce candidate results through the real integration and open them in that receiving surface.
- Compare readability and output size using your own documents before committing to one result type.
Do not infer controllable dimensions, image quality settings, page splitting or a particular PDF standard from a supported-extension list. Those are integration details to confirm separately.
When should you keep more than one result?
A workflow may distribute a document while also presenting a visual review. Treat those as separate requirements: keep the source identity associated with both outputs and check each consumer independently. This is application orchestration over conversion and preview, not a documented multi-output SDK call.
Frequently asked questions
Is CSV another visual output?
CSV is documented for the Excel family and serves a tabular data workflow. It is not a substitute for an image of a spreadsheet.
Does output support imply input support?
No. Read the two matrix columns independently; several image formats are output-only.
Is one output always smaller or faster?
The product documents provide no universal benchmark. Measure with your documents and target environment.