Cookies for analytics and advertising
We use cookies for analytics and advertising, both sent to Google. Refusing changes nothing you can see.Read the privacy page
Converting CR2 to JPG here pulls out the finished JPEG your Canon already wrote inside the raw file at the moment of capture, at full resolution on almost every body. It is the camera’s own rendering rather than a development of the sensor data, the EXIF is stripped out by default and kept if you ask for it, and it all happens on your own machine.
Up to 100 files at once. Mixed formats are fine.
They convert one after another and download together as a ZIP.
CR2 to JPG
A CR2 is a TIFF-derived container that Canon has been writing since 2004, and the reason it is four to six times the size of a JPEG is that it holds the sensor readings at 14 bits per channel rather than a picture at 8. Nothing on a normal computer can draw that directly; it has to be demosaiced first, which means reconstructing full colour from a grid where each photosite recorded only one of red, green or blue.
But the sensor data is not all a CR2 holds. Alongside it Canon writes a complete JPEG, produced by the camera’s own processor at the instant you pressed the shutter, and on almost every EOS body that JPEG is the full pixel dimensions of the sensor. It is what the rear screen shows you, what Windows Explorer and macOS Finder use to draw a thumbnail, and what this converter extracts.
That distinction matters enough to state plainly. This is not a raw developer. The picture that comes out carries whatever picture style the camera was set to — Standard, Faithful, Landscape, Monochrome — along with the white balance the camera chose and its own sharpening and noise reduction. If you shot a frame in Monochrome and expected the colour back, it will not be there, because the colour is in the sensor data and the sensor data is not what was extracted.
For the great majority of people typing "CR2 to JPG" that is exactly right: they want to look at the photograph, attach it to something or upload it, and the camera’s rendering is the one they already approved on the back of the camera. For anyone who wants to recover a blown sky or push two stops of shadow, it is the wrong tool and the CR2 should go to Digital Photo Professional, Lightroom or darktable instead.
On this pair the extracted preview is already a JPEG, so it is never decoded and encoded again. That makes this the one conversion on the site where the metadata question has a real answer in both directions. By default the "Remove metadata" option is on and the EXIF block is edited out of the file before you receive it: shutter speed, aperture, ISO, focal length, lens model, the exact capture timestamp and any GPS coordinates the camera recorded. Turn the option off and every one of those arrives in the JPG exactly as the camera wrote it.
Everywhere else on this site that choice does not exist. A conversion to PNG or WebP decodes to raw pixels and encodes again, which discards metadata whatever the option says. Here it is a genuine switch: leave it on for a photograph going to a stranger or onto a public listing, turn it off when the capture data is part of what you are handing over. Setting a maximum width used to be the way to strip the EXIF and is no longer the right advice — it only worked by forcing a re-encode, which costs a generation of quality to do what the option now does directly. The ICC colour profile is kept in both cases, which is what stops an Adobe RGB frame arriving washed out.
CR2 spans roughly 2004 to 2018 and covers the bulk of the EOS DSLR line: the 5D series through the Mark IV, the 6D and 7D, the 70D and 80D, and every Rebel and xxxD of that era. If your camera has a mirror in it and is not brand new, the files on the card are almost certainly CR2.
Canon’s earliest digital SLRs wrote CRW instead, and from 2018 the EOS R bodies and the newer DSLRs write CR3, which is a different container entirely rather than a version bump. If the files on your card end in .CR3, use the CR3 page instead — the extraction works the same way but the reason people arrive there is different, and it is worth reading.
Expect a substantial drop. A 22-megapixel 5D Mark III CR2 runs around 25 MB and a 30-megapixel 5D Mark IV around 35 MB; the JPEG inside is a normal camera JPEG and typically lands somewhere in the single-digit megabytes. The saving is not compression cleverness, it is the difference between storing sensor readings and storing a picture.
That is the whole appeal for anything with an upload form on the other end. Job application portals, insurance claim uploads, print shop web forms and marketplace listings routinely cap attachments in the low tens of megabytes and refuse .CR2 outright because they have never heard of it.
Drop the whole selection at once. Each file is handled separately with its own progress, and once several are done you can take them together as a ZIP rather than clicking through downloads one by one. Nothing is queued behind other people’s jobs, because the work is happening on your device.
The practical ceiling is 100 MB per file on the free tier, which no Canon CR2 approaches, and after that your own machine. Extraction is cheap — there is no decoder to run, only a scan through the file for the embedded JPEG — so a card’s worth of frames goes considerably faster than it would through a raw developer.
The converter looks for every JPEG inside the file and keeps the largest by pixel count, ignoring anything below 640 by 480 on the grounds that a 160-by-120 thumbnail is a worse outcome than a clear failure. If it finds nothing above that threshold it says so rather than handing you a postage stamp.
When that happens the file has usually not come straight from a camera. CR2 files written out by third-party software, recovered by an undelete tool, or truncated by a card pulled mid-write can be missing the preview or have it damaged. The fix is to go back to the original card copy if one exists; if it does not, the file needs a real raw decoder, which will read the sensor data even where the preview is gone.
Converting does not consume the original, and the original is worth keeping. Everything that makes raw worth shooting — the highlight headroom, the shadow latitude, the ability to change white balance after the fact without a penalty — lives in the 14-bit sensor data and is not present in the extracted JPEG in any recoverable form.
A reasonable working pattern is to extract JPGs for everything so you can browse, share and send immediately, and keep the CR2 files for the handful of frames that will actually be worked on. That way the fast path and the careful path do not compete, and you have not thrown away the thing you cannot get back.
The awkward part of CR2 has always been that no browser and few operating systems will display it. Windows needs a codec pack or the Microsoft Raw Image Extension, macOS supports many Canon bodies but lags new ones, and the file is simply invisible to anything running on the web.
This page needs none of that installed. The extraction is JavaScript running in the tab you already have open, which means it works on a borrowed laptop, a locked-down work machine or a phone, and the file never leaves the device it is on.
| CR2 | JPG | |
|---|---|---|
| Full name | Canon Raw 2 | JPEG Image |
| File extension | .cr2 | .jpg, .jpeg, .jpe |
| Media type | image/x-canon-cr2 | image/jpeg |
| Compression | Lossless — nothing is discarded | Lossy — file size is bought with quality |
| First published | 2004 | 1992 |
| Published by | Canon | Joint Photographic Experts Group |
| Specification | — | ITU-T T.81 |
| Licensing | Proprietary | Open standard |
| Standing today | Legacy, still read everywhere | Current |
| Bit depth | 14 | 8 |
| Colour it can describe | RGB | RGB, greyscale, YCbCr |
| Largest image | — | 65,535 px per side |
| Opens in a browser | No browser | Every browser |
| Considered instead | DNG, TIFF | WebP, AVIF, HEIC |
CR2 carries up to 14 bits per channel and JPG stores 8. The extra precision is what survives heavy correction without banding, so the conversion is best made after the editing rather than before it.
EXIF, XMP and GPS coordinates can cross over — both CR2 and JPG have somewhere to store them.
JPG opens in every current browser. CR2 has narrower browser support than that. If the file is going onto a web page or into a form, that is usually the whole reason for the conversion.
CR2 dates from 2004 and is largely superseded. JPG is what current software writes, so the conversion is as much about staying readable as about the file itself.
CR2 is a vendor format; JPG is a published specification (ITU-T T.81). That matters for anything meant to still open in ten years, when the program that wrote the original may not be around.
The usual programs do not overlap: CR2 opens in Adobe Lightroom, Canon Digital Photo Professional and darktable, JPG in Adobe Photoshop, GIMP and Preview — so whoever receives the result needs something from the second list.
The file gets substantially smaller, and it gets smaller by discarding detail. CR2 keeps everything; JPG keeps what the eye is poor at noticing. On a photograph that trade is close to free; on text, a screenshot or a line drawing it is visible as haloing around the edges.
CR2 is Canon's format, published in 2004. It comes out of Canon EOS digital SLRs. Underneath it is TIFF, which is why a program that has never heard of CR2 can sometimes still open one.
JPG comes from Joint Photographic Experts Group and dates from 1992, specified as ITU-T T.81. Adobe Photoshop, GIMP and Preview all read it.
No. This conversion runs entirely inside your browser, so the file never leaves your device. You can confirm it yourself: open the network tab of your browser's developer tools and convert something. You will see the page load, plus the analytics and advertising the site is paid for with — and nothing carrying your file. The engine behind this particular pair is a raw preview extractor, which pulls the JPEG the camera already embedded rather than developing the sensor data; your browser fetches it once and caches it.
No, and that is the honest answer rather than a limitation we would rather not mention. Every Canon writes a finished JPEG inside the CR2 at the moment of capture — it is what the camera shows on its own screen. This pulls that JPEG out. It is the camera’s own rendering with the picture style you had set, at full resolution on almost every body. Developing the sensor data is a different job that needs Digital Photo Professional, Lightroom or darktable.
Not unless you ask for it. "Remove metadata" is on by default, so the EXIF block — shutter speed, aperture, lens, timestamp and any GPS tag the camera wrote — is taken out of the JPG before you get it. Turn that option off and the preview is handed back byte for byte with all of it intact, because nothing re-encodes the picture either way. The ICC colour profile is kept whichever way you set it.
Because at the default setting nothing is re-encoded. The preview is copied out as it is, so the quality slider has nothing to act on. It only takes effect if you also set a maximum width, which forces the picture through a decode and encode.
Not from this JPG. The extra latitude is in the 14-bit sensor data, and the embedded preview is an 8-bit rendering with the camera’s tone curve already applied. If a frame needs rescuing, keep the CR2 and open it in software that demosaics it.
Up to 100 MB per file on the free tier, which covers every Canon CR2 comfortably — a 30-megapixel 5D Mark IV frame is around 35 MB. The work runs on your own machine, so a batch is limited by your computer rather than by a queue.
No. This conversion runs in your browser. You can confirm it the way we do: open the developer tools, switch to the network tab and convert a file, or disconnect from the internet and watch the page carry on working.
The claims this page makes about CR2 and JPG are checkable, and these are the documents that settle them.