Initializing... drag & drop files here
Supports: X3F
X3F is Sigma's proprietary RAW container for cameras built around the Foveon X3 sensor — a three-layer silicon stack that records red, green, and blue at every photosite rather than interpolating them from a Bayer mosaic. The format first shipped with the Sigma SD9 in 2002 and stayed through the Merrill (46 MP equivalent) and Quattro (39 MP Bayer-equivalent, 1:1:4 layer ratio) generations. The trade-off is that almost nothing outside Sigma Photo Pro, dcraw, and a handful of third-party tools opens it natively — so once you've picked your keepers, exporting to PNG turns the file into something every browser, OS, and image editor can read losslessly.
Need a different output? See X3F to JPG for smaller share-ready files, X3F to TIFF for 16-bit print masters, or X3F to PDF for proof sheets. Already have PNGs and want them smaller? Run them through Compress PNG.
| Property | X3F (Sigma RAW) | PNG |
|---|---|---|
| Type | Proprietary camera RAW container | Open raster image format |
| Standardised by | Sigma Corporation (no public spec) | W3C / ISO/IEC 15948 (2003) |
| Compression | Lossless RAW with Huffman + embedded JPEG preview | Lossless DEFLATE with prediction filtering |
| Bit depth (typical) | 12-14 bits per channel sensor data | 1, 2, 4, 8, or 16 bits per channel |
| Colour model | Three stacked Foveon photodiodes per pixel | Greyscale, indexed, RGB, RGBA |
| Transparency | No alpha channel | Yes — 1-bit or full 8/16-bit alpha |
| Software support | Sigma Photo Pro, dcraw, LibRaw, Sigma X3F PS plug-in | Every major browser, OS, image editor since the late 1990s |
| Best for | Capture, full editing latitude | Distribution, web, print, archival masters |
| Typical file size (24 MP equivalent) | 45-75 MB | 15-60 MB depending on bit depth and content |
| Use case | Resolution | Colors | Compression level | Notes |
|---|---|---|---|---|
| Archive master | Original (full sensor) | True color | High (slow) | Largest file; locks in today's demosaic so a Sigma Photo Pro update can't shift the rendering |
| Print at 300 DPI, 8x10" | 2400x3000 px | True color | High (slow) | Matches commercial print pipeline without a JPEG step |
| Portfolio web image | 2048-2560 px long edge | True color | Medium | Sharp on a Retina display without the file becoming a download |
| Social-media post | 1080-1440 px long edge | True color | Medium | Fits Instagram/X without server-side re-encode artifacts |
| Email proof | 1600 px long edge | True color | Medium | Stays under typical 25 MB attachment caps even multi-up |
| Icon / UI asset | 256-512 px | 256 (indexed) | High | Smaller than JPEG for flat graphics and logo-style content |
X3F is only safe to keep if you (and everyone you share with) have a current copy of Sigma Photo Pro, dcraw, or LibRaw. PNG is readable by every browser and OS, including ones that won't ship with X3F support a decade from now. Most photographers keep the X3F as their archival negative and export a PNG (or TIFF) as the deliverable.
No — once the X3F has been demosaiced, the three-channel RGB result is preserved exactly by a lossless PNG. The catch is that demosaicing happens at conversion time, so a different RAW processor on a different day may render slightly different colour. Save a 16-bit PNG (rather than 8-bit) if you want the most tonal headroom for later edits.
No — the conversion uses an open-source X3F decoder pipeline (LibRaw-style) running on our servers, so you don't need Sigma Photo Pro installed. The trade-off is that very early X3F files from the SD9/SD10 era and some Quattro-specific metadata are less battle-tested than they would be in Sigma's own software. If you see colour shifts on those frames, run them through Sigma Photo Pro first and export to TIFF.
X3F stores compressed sensor-side data (one value per photodiode, three diodes per pixel). When you demosaic and write to PNG you produce full RGB (or RGBA) data for every pixel, which can be larger — especially at 16-bit. Drop to 8-bit, use a smaller resolution preset, or accept the larger file as the cost of a universally readable output.
No — that preview exists for in-camera review and quick thumbnails, and using it would throw away the whole point of shooting RAW. The converter demosaics the full Foveon data so your PNG reflects the actual sensor capture, not the camera's JPEG-engine guess.
Yes for the most common tags — camera model, lens, ISO, shutter speed, aperture, and capture date are read from the X3F and written into the PNG's tEXt/eXIf chunks where supported. Some Sigma-specific maker notes (white-balance fine-tuning, custom film modes) don't have standardised PNG equivalents and are dropped.
TIFF is the traditional choice when you want 16-bit-per-channel data plus layers and the widest editor support. PNG is a lighter, universally browser-readable alternative that still supports 16-bit and lossless compression. For a viewer-first archive, pick PNG. For a "this file will live in an editor stack later" archive, pick X3F to TIFF.
Any Sigma body using the Foveon X3 sensor: SD9 and SD10 (early 2000s), the SD14/SD15/SD1 line, the DP1/DP2/DP3 fixed-lens compacts, the Merrill generation (46 MP equivalent), the Quattro generation (39 MP Bayer-equivalent with the 1:1:4 layer ratio), and the SD Quattro / SD Quattro H. Files from third-party tools that have re-wrapped X3F may have lost metadata the decoder needs.
Modern Quattro and Merrill X3F files are typically 45-75 MB each; that's well inside the converter's per-file ceiling. If a frame is unusually large (HDR bracket re-saved as one file, for example), split or re-save it from Sigma Photo Pro first. For multi-gigabyte shoots, upload in smaller batches rather than one giant queue.