The Evolution of Canon RAW Image Files
From CR2 to CR3 and Beyond: The Evolution of Canon RAW
Explore Canon RAW evolution from CR2 to CR3 and C-RAW, and consider what the future of RAW photography could look like.
A conceptual representation of Canon RAW evolution, showing CR2, CR3 and C-RAW progressing towards a possible next-generation RAW architecture incorporating greater efficiency, computational imaging, metadata and long-term photographic preservation.
Vernon Chalmers Photography examines the evolution of Canon's RAW ecosystem from CR2 to CR3 and C-RAW from the perspective of practical Canon photography and digital image workflow. The discussion distinguishes established Canon technology from informed projections about future RAW architectures, computational imaging and photographic archiving.
How Canon RAW Has Evolved — and Where It Could Go NextThe Evolution of Canon RAW
But behind that apparently simple workflow lies an important technological evolution.
Canon's RAW system has changed considerably over the history of EOS digital photography. The progression from CR2 to CR3, and from conventional RAW to C-RAW, reflects more than a change in filename extension. It reflects the changing demands placed on digital cameras as sensor resolutions increase, continuous shooting becomes faster, processors become more powerful and photographers generate increasingly large quantities of image data.
The interesting question, therefore, is not simply what Canon's next RAW file extension might be.
It is this:
What should a RAW file become as the digital camera itself becomes increasingly sophisticated?
That question provides an opportunity to look backwards at CR2, examine the transition to CR3, consider the significance of C-RAW, and then look beyond today's technology towards a possible next generation of Canon RAW.
Canon Early RAW Formats (.CRW and .CR2)
- .CRW: Used by very early Canon digital cameras (like the EOS 300D/Rebel). It was based on an older proprietary structure.
- .CR2: Introduced with the DIGIC II processor and used for over a decade across most EOS DSLRs. It stores 14-bit uncompressed or losslessly compressed sensor data.
- M-RAW / S-RAW: Mid-era reduced-resolution raw options that lowered pixel counts to save space instead of using advanced compression.
Canon Modern RAW Format (.CR3 and C-RAW)
- .CR3: Debuted with the DIGIC 8 processor (starting with the EOS M50) and standard across modern mirrorless bodies using DIGIC X processors. It uses the ISO base media file format structure.
- C-RAW (Compact RAW): A lossy compression option tied to .CR3 that keeps full sensor resolution while reducing file sizes by roughly 35% to 55%. Benefits of C-RAW
RAW: Preserving Information for Later Interpretation
A conventional JPEG is essentially a finished interpretation of the information captured by the camera.
The camera processes the sensor data, applies a range of image parameters, compresses the result and produces a relatively compact image suitable for immediate viewing and distribution.
RAW takes a different approach.
Instead of producing only a finished image, the camera records much more of the information generated by the sensor and preserves it for subsequent interpretation.
This is fundamental to the photographic RAW workflow.
Exposure, white balance, tonal relationships, colour information and other characteristics can subsequently be interpreted by software such as Adobe Lightroom Classic and other RAW processors.
The RAW file therefore occupies an unusual position in digital photography.
It is neither simply an image nor merely a collection of sensor measurements.
It is a photographic data resource from which an image can subsequently be developed.
That distinction becomes increasingly important as RAW technology evolves.
The CR2 Generation
Canon's CR2 format became one of the principal RAW formats of the EOS digital era.
For many years, CR2 was familiar to Canon photographers using DSLR cameras. Cameras ranging from enthusiast models to professional EOS bodies produced CR2 files, creating an extensive ecosystem of software, workflows and archival material around the format.
During the CR2 era, the fundamental RAW concept remained relatively stable.
The camera captured sensor information, encoded that information into the RAW file and transferred responsibility for much of the final image interpretation to the photographer's software.
But the photographic environment around the RAW file was changing rapidly.
Sensor resolutions increased.
Memory cards became larger.
Continuous shooting became faster.
Image processors became considerably more powerful.
And photographers increasingly expected their cameras to operate at speeds that would have been difficult to imagine when earlier digital RAW systems were developed.
The RAW file therefore faced a new challenge.
It was no longer simply necessary to preserve a single photograph.
Modern cameras increasingly had to preserve large quantities of high-resolution photographs at very high speed.
Canon CR2 vs CR3 RAW Image Files
From CR2 to CR3
Canon's transition to CR3 represented an important development in its RAW architecture.
CR3 was introduced with Canon's newer generation of cameras and became associated with the company's mirrorless EOS R system.
One of the most significant developments was the introduction of C-RAW, Canon's Compact RAW option.
The importance of this development is easy to underestimate.
C-RAW was not simply a lower-resolution image mode comparable to the earlier concept of reducing image dimensions.
Instead, it provided a way of recording a full-resolution RAW image while reducing the amount of storage required compared with conventional RAW.
That represented a fundamental shift in the relationship between RAW information and file size.
The question was no longer simply:
How much information can we record?
It increasingly became:
How efficiently can we represent that information?
That distinction is likely to become even more important in the future.
The Benefits of Shooting in Canon C-RAWC-RAW: The Beginning of a Different RAW Philosophy
The significance of C-RAW extends beyond saving space on a memory card.
For photographers working with high-resolution cameras and rapid continuous shooting, storage efficiency has practical consequences.
A smaller RAW file can mean:
less storage consumed per photograph;
more photographs accommodated on a memory card;
less data written during a burst;
less data transferred during ingestion;
and potentially more manageable long-term image archives.
But there is always a fundamental trade-off.
The more aggressively image information is represented or compressed, the greater the possibility that some information may not be recoverable in exactly the same way as from a conventional RAW file.
This becomes particularly relevant when a photograph requires substantial post-processing.
Severely underexposed images, extreme shadow recovery, heavy tonal manipulation and other demanding adjustments can reveal differences between different RAW approaches.
That leads to a much more interesting interpretation of C-RAW.
C-RAW can be viewed as an experiment in photographic information efficiency.
It asks whether a RAW file needs to preserve every aspect of the original sensor information in the same way, or whether a more efficient representation can retain sufficient information for the vast majority of real-world photographic applications.
For many photographers, the practical answer is already clear.
If the final photograph is indistinguishable for their purposes, the smaller file has considerable value.
But the technological question does not end there.
It may actually be the beginning.
Canon RAW Evolution: Why CR3 MattersThe Future Problem: More Pixels, More Frames, More Data
Consider the direction in which digital cameras have been developing.A modern high-resolution camera can generate tens of millions of pixels per frame.
Professional cameras can capture those frames at very high frame rates.
Autofocus systems can analyse subjects continuously.
Electronic shutters can operate at remarkable speeds.
Image processors can perform increasingly sophisticated calculations between exposures.
And future sensors will almost certainly continue to increase the amount of information available to the camera.
This creates a paradox.
The better the camera becomes, the more data it potentially has to manage.
A future camera with substantially greater resolution and faster continuous shooting could generate enormous volumes of RAW information during a single photographic session.
Simply making memory cards larger is one solution.
But it is not the only solution.
A more fundamental solution is to make the RAW representation itself more efficient.
This is where the evolution beyond CR3 becomes particularly interesting.
Could C-RAW Evolve?
There is no reason to assume that the current implementation of C-RAW represents the final form of compact RAW.
Canon has not announced a future C-RAW architecture of the type discussed here, so any discussion of its future must necessarily remain speculative.
Nevertheless, the technological direction is logical.
As sensors become more sophisticated, future RAW compression could potentially become increasingly intelligent.
Rather than applying essentially the same approach to every image, future systems could potentially exploit the statistical characteristics of the captured data more effectively.
The objective would not necessarily be to produce the smallest possible RAW file.
It would be to produce the most efficient RAW representation while retaining the information that matters photographically.
This distinction is important.
A future compact RAW format might be judged not by how much data it removes, but by how little useful photographic information is lost relative to the reduction in file size.
That could produce a different way of thinking about RAW compression.
Instead of:
Full RAW = information
and
C-RAW = less information
the future could potentially be:
RAW = maximum retained information
and
C-RAW = maximum photographic utility per unit of storage.
That would represent a natural continuation of the philosophy behind C-RAW.
Beyond Compression: The Computational RAW
There is another possibility that may ultimately be even more significant.
Modern cameras are no longer passive recording devices.
They are computational imaging systems.
Autofocus systems identify subjects.
Subject-detection algorithms classify objects.
Exposure systems continuously analyse the scene.
Lens and camera information can be combined.
Noise reduction and other processing can increasingly occur within the camera.
Future cameras may perform even more sophisticated computational operations before the image reaches the RAW file.
This raises an important question:
What exactly should a future RAW file preserve?
Traditionally, the answer has been largely associated with sensor data.
But a future RAW architecture could potentially contain a much richer combination of information:
sensor-derived image data;
camera metadata;
lens information;
exposure information;
focus-related information;
computational imaging metadata;
and information describing how the camera interpreted the capture.
This does not mean that the RAW file would become a JPEG.
Quite the opposite.
The objective could be to preserve more of the photographic capture process, while allowing the photographer or future software to reinterpret that information.
The RAW file could consequently evolve from being primarily a digital negative into something closer to a digital photographic capture record.
RAW and Artificial Intelligence
Artificial intelligence adds another dimension to this discussion.
Today's AI-based photographic tools can already perform sophisticated operations such as noise reduction, sharpening, object recognition, masking and image reconstruction.
As these technologies mature, the boundary between camera capture and computational processing is likely to become increasingly fluid.
A future camera could potentially capture information that is specifically intended to support later computational interpretation.
This creates an interesting possibility.
Rather than asking:
What image did the camera produce?
future RAW systems may increasingly allow us to ask:
What information did the camera capture that can subsequently be interpreted in different ways?
That is a subtle but important change.
It moves RAW further away from being a static file and towards becoming a reinterpretable photographic data structure.
The Photographer Still Matters
There is, however, an important reason why RAW should not become completely dependent on computational interpretation.
The photographer must retain control.
Photography is not simply the reconstruction of what an algorithm believes was present in front of the camera.
It is an act of observation, interpretation and creative decision-making.
A future RAW system should therefore ideally provide greater computational capability without reducing the photographer's ability to make independent decisions during post-processing.
This is particularly important for professional, scientific, documentary and archival photography.
A future RAW file should not simply tell the photographer what the camera thinks the photograph should look like.
It should preserve enough underlying information for the photographer to determine what the photograph can look like.
That distinction could become increasingly important as computational photography advances.
The Archival Question
There is another issue that becomes more important with every generation of proprietary RAW technology.
What happens to a RAW photograph decades from now?
CR2 files created years ago remain part of photographers' archives.
CR3 files are now becoming similarly important.
But RAW formats are proprietary technologies, and their long-term usability depends partly on continued support from camera manufacturers and software developers.
For photographers creating significant archives, this raises a fundamental question.
Is the RAW file merely a working format for the present, or is it intended to become a permanent photographic record?
Ideally, it should be both.
A RAW photograph may contain considerably more information than the JPEG eventually published on a website or social-media platform.
For long-term photographic archives, that original data can therefore have enormous value.
The future evolution of RAW should consequently consider not only compression efficiency and computational capability, but also interpretability and preservation.
A highly efficient RAW file that becomes unreadable in fifty years would be a poor archival solution.
Could There Eventually Be a New Canon RAW Architecture?
It would be premature to predict exactly what Canon might call its next RAW format.
It might eventually be CR4.
It might use an entirely different architecture and naming convention.
Or CR3 could remain Canon's underlying RAW architecture for considerably longer than expected.
The important issue is therefore not the name.
The important issue is the technological requirement.
A future Canon RAW system would need to accommodate an imaging environment in which cameras capture considerably more information, at considerably higher speeds, while photographers expect that information to remain accessible and editable.
That suggests several possible design priorities:
Greater storage efficiency.
Higher-resolution sensors make this increasingly important.
Improved compression efficiency.
More sophisticated encoding could potentially reduce file sizes without materially compromising photographic latitude.
Greater computational awareness.
The RAW file may increasingly need to coexist with computational imaging processes.
Richer metadata.
As cameras and lenses become more intelligent, metadata could become an increasingly valuable component of the photographic record.
Long-term accessibility.
Future RAW systems should ideally remain interpretable as photographic archives rather than becoming dependent upon obsolete software ecosystems.
From Digital Negative to Digital Capture Record
This may ultimately be the most significant evolution.
The early concept of RAW was relatively straightforward:
Capture the sensor information and let the photographer develop the image later.
CR2 represented an important era of that philosophy.
CR3 and C-RAW demonstrate the increasing importance of efficiency.
The next generation could potentially take the concept further.
The RAW file could become a comprehensive record of the photographic capture rather than simply a container for relatively unprocessed sensor data.
In such a system, the photographer might retain access to the underlying image information while also benefiting from increasingly sophisticated computational metadata and processing possibilities.
The RAW file would therefore become less like a static digital negative and more like a digital photographic master.
That could be one of the most important developments in digital photography over the next decade.
Where Does This Leave C-RAW?
C-RAW should therefore not necessarily be viewed as a compromise between RAW and JPEG.
It can instead be understood as part of a much broader evolution towards efficient photographic data representation.
Its future could follow the same trajectory as Canon's broader RAW architecture.
As camera-generated data increases, the value of efficient representation also increases.
If Canon eventually develops a substantially more sophisticated RAW architecture, it would be logical to expect some form of compact or efficient RAW representation to evolve alongside it.
The precise technology is impossible to predict.
The underlying requirement is not.
Future cameras will need increasingly efficient ways of preserving increasingly large quantities of photographic information.
From CR2 to CR3 — and Beyond
Looking back, the progression can be understood as a series of responses to changing photographic technology.
CR2 emerged during an era in which the principal challenge was capturing and preserving increasingly sophisticated digital sensor information.
CR3 arrived in an era of mirrorless cameras, higher resolutions, faster processing and much greater image volumes.
C-RAW addressed another part of that problem by asking how full-resolution RAW photography could be represented more efficiently.
The next stage may be less about simply creating another file extension and more about redefining what the RAW file represents.
The future RAW file may need to balance:
information, efficiency, computational capability, editability and archival longevity.
That is a considerably more complex problem than simply producing a larger memory card.
And it is one that will increasingly affect every photographer working with high-resolution digital cameras.
A Personal Perspective
As a photographer working with Canon cameras across different generations, I find the evolution from CR2 to CR3 particularly interesting because it demonstrates that RAW is not a fixed concept.
It is an evolving relationship between the camera, the sensor, the processor, the storage medium, the software and the photographer.
My own experience working with both RAW and C-RAW also reinforces an important practical principle: the best format is not necessarily the one that preserves the greatest theoretical quantity of data.
It is the format that preserves sufficient useful information for the photographic task while fitting the photographer's workflow.
That distinction is likely to become increasingly important.
As Canon cameras continue to evolve, RAW will have to evolve with them.
CR2 was not the final destination.
CR3 is unlikely to be the final destination either.
And C-RAW may ultimately prove to have been an important step towards a broader future in which the RAW file becomes not merely a digital negative, but a highly efficient, computationally aware and enduring digital photographic master.
The fascinating question is no longer simply:
“What will Canon call its next RAW format?”
It is:
“What should the next generation of Canon RAW preserve?”
That may ultimately determine what the future of digital photography looks like.
