An upscaler can only work with the information that is actually in the file. If you take a photo on an iPhone (which is stored as a HEIC or JPEG at modest quality), then convert it to PNG before upscaling, you are not feeding the upscaler more data. You are just wrapping the same JPEG-degraded pixels in a lossless container.
This matters because there is a popular myth that "PNG is better quality, so always use PNG". Better at preserving the pixels you give it, yes. It does not add any.
For these cases, exporting as PNG keeps every pixel intact for the upscaler to work with.
Different question entirely. Now you have a fresh, clean output. Format choice depends on what you will do with it.
| Use case | Pick | Why |
|---|---|---|
| Share to iMessage, Instagram, X | JPEG, high quality (90+) | Smaller file, indistinguishable on screens, faster upload. |
| Keep editing in Photoshop / Affinity | PNG | No further loss when you save again. |
| Print at A3 or larger | PNG or TIFF | Lossless detail matters at high DPI. |
| Email an old photo to a relative | JPEG quality 92 | Smaller attachment, visually identical. |
| Web hero image | WebP if you can, JPEG if not | WebP cuts file size roughly 30% at the same quality. |
If you are on iPhone, your photos are probably HEIC by default. HEIC is a much more efficient codec than JPEG (typically half the file size for the same visible quality), and most modern upscalers including ours accept HEIC directly. There is no need to convert to JPEG first; you would just be downgrading.
The exception is if you need to share the output with someone on an older Windows machine. In that case, export to JPEG after upscaling, not before.
On-device AI upscaling up to 4x. Accepts JPEG, PNG, HEIC, and exports any of them. · iPhone & iPad