CR3 vs CR2: What Changed in Canon's RAW Format

The biggest difference is the file container. CR2 uses a TIFF-based container — the RAW data is wrapped in a structure derived from the TIFF 6.0 specification. This made CR2 files relatively simple to parse and widely compatible with image processing software from the start.

CR3 switched to the ISO Base Media File Format (ISOBMFF), which is the same container used by MP4 video, HEIF/HEIC images, and AVIF. It's a more modern and flexible container that supports features the old TIFF-based structure couldn't handle easily — like storing multiple image tracks (full resolution, C-RAW compressed, and a small proxy) in a single file.

The container change is why older software couldn't open CR3 files — the tools knew how to read TIFF-based containers but had no ISOBMFF parser. Adding support meant implementing an entirely new container reader, which is why it took some open-source projects years to add CR3 support.

CR3 introduces a compression option that didn't exist in CR2: C-RAW (Compact RAW). When shooting C-RAW, the sensor data is lossy-compressed using a wavelet-based algorithm that typically reduces file size by 30–40% compared to full RAW.

A Canon R6 Mark II full RAW file is about 24 MB. The same shot in C-RAW is about 14 MB. Over a 1,000-shot wedding day, that's the difference between 24 GB and 14 GB of card space. Over a year of shooting, it adds up to hundreds of gigabytes of storage difference.

The quality difference? In most real-world shooting — ISO under 6400, no extreme shadow recovery — C-RAW and full RAW are visually indistinguishable. You can open the same scene shot in both modes, zoom to 100%, and not tell which is which. Push a C-RAW file hard in editing (lifting shadows 5+ stops or recovering heavily blown highlights) and you might see slightly more noise or faint banding compared to full RAW. For 95% of photographers in 95% of shooting situations, C-RAW is fine.

CR2 had only one compression option: lossless compressed. There was no lossy compact mode. You got the full data every time, whether you needed the editing headroom or not.

CR2 has been around since 2004 (debuting with the EOS-1D Mark II) and is supported by essentially every photo editing tool ever made. CR3 took longer to gain support because the ISOBMFF container was new for still images.

Adobe Lightroom — CR3 support started in Lightroom Classic CC 7.3 (April 2018). If you're on an older version — say, Lightroom 6 (the last perpetual license version) — CR3 files won't open. Adobe Camera Raw 10.3+ also handles CR3.

Capture One — added CR3 support in version 12.1 (2019). Older versions can't read CR3 at all. Current versions handle CR3 well.

darktable — CR3 support arrived in version 3.6 (2021). The open-source community had to reverse-engineer the ISOBMFF container because Canon doesn't publish a public CR3 specification.

Canon DPP — Canon's free Digital Photo Professional always supports both CR2 and CR3 from all Canon cameras. It's the guaranteed fallback if your preferred editor hasn't caught up — Canon updates DPP for each new camera launch, usually before third-party support arrives.

If you can't open your CR3 files, the troubleshooting guide has the specific fixes for both Windows and macOS.

CR3 files store metadata differently than CR2. The ISOBMFF container uses "boxes" (or "atoms" in QuickTime terminology) to store metadata, versus CR2's TIFF-style IFD (Image File Directory) entries. In practice, this means:

EXIF data is stored in the same fields but in a different structural location within the file. Tools that parse CR2 EXIF by looking for TIFF IFDs won't find the data in a CR3 file without being updated. Most modern EXIF tools (ExifTool, for example) handle both containers transparently.

Makernote data (Canon's proprietary shooting info — lens data, AF point info, Picture Style settings, internal temperature readings) is stored in a CNCV/CCTP box in CR3, versus a standard Makernote IFD in CR2.

The embedded JPG preview in CR3 is typically higher quality than the CR2 preview. CR3 embeds a full-resolution JPG preview, while CR2 embedded a half-resolution preview on most camera bodies. This is why thumbnail browsing in CR3 folders looks sharper than in CR2 folders.

No. There's no practical reason to convert between the two. Each format is a container for the sensor data from the specific camera that captured it. Converting CR2 to CR3 doesn't give you the CR3 features (like C-RAW compression) because those are capture-time optimisations performed by the camera's image processor, not post-processing ones you can apply after the fact.

If you want a universal RAW format, Adobe DNG is the closest thing. Both CR2 and CR3 can be converted to DNG using Adobe's free DNG Converter. DNG strips out the proprietary container and stores the raw sensor data in an open format defined by the DNG specification (Adobe, currently at version 1.7). Some photographers convert their entire archive to DNG for long-term compatibility; others keep the original proprietary files for maximum data preservation.

If you just need to share the photos in a universal format, converting to JPG is the simplest approach. The image converter handles both CR2 and CR3.

For a 24-megapixel Canon sensor (like the R6 III or R6 II), typical file sizes by format are: CR2 full RAW ~26 MB, CR3 full RAW ~24 MB, CR3 C-RAW ~14 MB. The slight size reduction from CR2 to CR3 full RAW comes from the more efficient ISOBMFF container. The big drop to C-RAW is the lossy compression at work.

At 45 megapixels (like the R5 II), the numbers scale up: CR3 full RAW ~50 MB, C-RAW ~28 MB. That's nearly a 2:1 ratio, which adds up fast on a multi-day shoot.

Many Canon cameras let you write RAW to one card slot and JPG to the other. With the R5 and R6, you can write full RAW to the CFexpress slot and C-RAW to the SD slot (as a backup), or RAW to one and JPG to the other. This gives you the editing flexibility of RAW and the immediate shareability of JPG without any conversion step.

Event photographers often use this setup: shoot RAW + JPG, hand over JPGs same-day for social media and quick previews, then process the RAW files for the full edited gallery later. It's double the storage per shot, but with 128 GB and 256 GB cards now commonplace, it's manageable.

If you shoot RAW-only and need JPGs later, batch conversion in Lightroom or through the image converter takes seconds per file.

Is CR3 better than CR2?
CR3 offers a more modern container format and the C-RAW compression option, so yes — it's technically more capable. But the image quality difference is negligible. The sensor and lens matter far more than the file container. A great CR2 photo from a 5D Mark IV is still a great photo.

Can Lightroom open both CR2 and CR3?
Yes, as long as you're running Lightroom Classic 7.3 or newer (released April 2018). Older versions open CR2 but not CR3. The current Lightroom subscription version handles both without any issues.

Do I lose quality shooting C-RAW instead of full RAW?
In normal shooting (ISO under 6400, moderate editing), no visible difference. In extreme editing — lifting shadows 5+ stops — C-RAW shows slightly more noise than full RAW. For event and portrait photography, C-RAW is perfectly fine. Landscape photographers who routinely do heavy shadow recovery may prefer full RAW.

Can older Canon cameras be updated to shoot CR3?
No. The RAW format is tied to the camera's image processing hardware and firmware architecture. Cameras that shipped with CR2 will always shoot CR2. Only cameras released from 2018 onward use CR3. There's no firmware update that changes the RAW container format.

Does the RAW format affect autofocus or burst speed?
Not directly, but C-RAW's smaller file size means faster card writes, which means the camera's buffer clears faster during burst shooting. On cameras with fast sensors like the R6 III (40 fps mechanical), the difference between full RAW and C-RAW buffer depth can be significant — dozens of extra frames before the buffer fills.