Need help - Will Pay $ - Someone Familiar with Repairing CR2 files?


Recommended Posts

Had a hard drive go south prior to backup of a wedding. Am in desperate need. Recovered from original media card the CR2 files I shot; was near 800 in all. All but 100 of them are corrupted as recovered from the CF card. They are all recognized as CR2 files in Explorer and even have the correct file sizes. But neither Lightroom, Canon Photo Utility, Photoshop or any other editors will open them. I've tried ExiftTool to extract JPGs, got file format errors. I have a Hex Editor but can't make heads or tails of what I'm looking at. I'm out of options and don't want to have to tell this couple I have lost their photos. Anyone out there have experience with repairing this type of file corruption that can do the work or have suggestions? PLEASE. Need help!!!!

CR2Repair prompts me to download Stellar Photo Repair, which is the program I used to recover the CR2 files from the CF Card. But although it says in the description that the software can repair corrupted headers, I can't for the life of me figure out how to do that. 

you can use PhotoInfo to recover the JPEG from the EXIF, if the header is still OK or has minor damage; if the header is very damaged than even if the data is still OK (doubt it), you can't access it. Please post the EXIFTool result here, so we can see if it's still possible to recover something. Also did you try PhotoInfo? There's also Quick JPEG from CR2, but i haven't tried it: http://www.gdzid.com/QJFC.htm

_MG_8767.CR2

Main Info

 

Filename

_MG_8767.CR2

Directory

H:

Image type

Canon Raw File

Width

0 pixels

Height

0 pixels

File size

23.1Mb

File created date

11/25/2013 12:17:38 PM

File modified date

9/7/2013 5:30:30 PM

Bits Per Plane

8

Planes

3

Dpi

72 pixels

tada

 

or8tV1x.png

 

edit:zhangm beat me to it.

 

anyways, just mangled headers. the header is in the middle of the file. to fix the image, start at the header, cut from there till the end of the file, this will include the header and some image data, and paste this back at the top of the file, so whatever you pasted is the header+some image data,then the rest of the data will follow. save and you have your image. its tedious if you have many images. i might be able to write you a program to automate it. i cant do it now though.

I'm using XVI32 to view the files, find the second instance of FF E0 in each one, and delete up to that point.

http://www.chmaas.handshake.de/delphi/freeware/xvi32/xvi32.htm

http://www.mikekunz.com/image_file_header.html

XVI32 is scriptable, so check out the documentation on their site.

This is greek to me. Any help in salvaging these images would be so appreciated. Let me look over the post, get the program and take a stab. Worst case scenario I get nowhere and hire you :-)

There are 800 images. Need to salvage as many as I can. 
 

turns out, this is just the lossless jpeg image that is stored as grayscale. there is still plenty of info in the file such as color that needs to be parsed. i'll definitely parse the file further hoping to demangle it so it displays correctly.

ok heres the deal. it looks like in each of the 2 files you provided, there is the complete grayscale image, and some color data, about 1/4 color of the image. there is other data in the file, but this looks like color data of a different file. i tried replacing the 3/4 color data in one file from the other one you gave me, but it isnt the right one. the problem is that parts of different files are mangled in between and they may not be in order.

 

you provided image files 8767,and 8768. if you could upload 5 images before 8767 and 5 images after 8768, i could then try using the color information from those files into the 2 images i currently have. if the correct color exists in these for the 2 files i have, then hopefully there is a pattern where we could see for example 3 filename numbers back has the correct color information for us to complete the current numbered image.

If you want to get a raw byte-by-byte image of the entire memory card to me, perhaps from that I could attempt to repair the filesystem and see if we can get the entire set of files restored.

Instructions for doing the above (so you can get on with some of it if I've gone to bed):

 

Capturing the device image:

 

We're going to use the Linux dd command with a 'Live' Linux CD. (I'm presuming you're not using Linux as your main operating system right now, else some of this is redundant).

 

Note, in regards to storage space for saving this device image; The image file is going to be slightly larger than the total volume size of the memory card you're copying. Whether though you actually need this amount of space to store it depends on whether the file system of the device you're storing it on uses compression and if so how much space can be saved. FAT does not use compression, but NTFS does. If you're going to save it to a hard drive, it's perhaps unecessary to point this out, but if you're saving to a pen drive or another memory card, this fact may be important!

 

Also note, I'm generally presuming you're using a Windows computer and therefore NTFS; Ubuntu may 'mount' your NTFS partitions automatically, if not, the following instructions may help: http://superuser.com/questions/352157/mount-ntfs-windows-partition-on-ubuntu-live-cd

 

1) Download a Linux live CD (E.g the standard desktop Ubuntu install disk)

2) Burn it to disk

3) Reboot your computer and boot this (I presume you need no instructions here, ask if you do). At some point during booting it, it'll give you an option to "try ubuntu without making changes to your computer", that's the one you want.

4) Connect the memory card if you haven't already

5) Connect the device/card you're going to save the image file to, if necessary

6) Load a terminal (command prompt)

7) Enter the following, modifying the references to the devices as necessary:

sudo dd if=/dev/hda of=/dev/hdb/image.dd bs=512 conv=noerror,sync

Where /dev/hda is the device to copy from and /dev/hdb is the device to save the file to.

8) When it's completed, you can shutdown and boot your normal OS

 

Getting it to me:

 

We'll use torrents as an efficient mechanism if you're happy with that.

 

1) Download and install utorrent (or other client of choice)

2) File > create new torrent

3) Point it to the image file (image.dd) created above

4) Leave all other options as defaults, i.e. leave the public trackers, leave 'start seeding' checked, leave 'private torrent' UNchecked

5) Click 'create' and save the small torrent file somewhere

6) Email me that little torrent file (I can give you my email via PM)

7) I'll open the torrent file in my torrent client which will start copying the file over to my computer, we just need to leave our computers on until I've got it, then I can start to work on it

Oh, something I forgot when writing the above instructions, when sending it to me, stick the image file in a zip file to compress it, then make the torrent with the zip file. That may reduce the size of the file to be transfered. I presume you need no instructions for this, but just ask if you do.

How large is this file going to be? Just thinking that using a service like http://ge.tt might be easier, if you are not a torrent user already.

 

 

What is the maximum/largest file size?

We do not limit the file size you can upload, but your account may not have enough storage space for the file. If that's the case you can upgrade your account to get more space.

Some browsers may limit you to 2GB per file. If this is a problem, then you can use a newer browser. Chrome is always a good choice.

 

What is the storage limit without an account?
You have 250MB of storage space without creating an account. If you create a free account you will receive 2GB space.
 
  • 2 weeks later...
This topic is now closed to further replies.
  • Posts

    • Plans. Christ at least editorialise this tripe for what it is or put your own journalistic take on it.
    • If you have a TV in your living room, chances are you can probably just use the Steam Link app and play your huge PC in big picture mode, effectively giving you the Steam Machine experience to see if you'd actually like it. The good news is the Steam Machine can have it's drives upgraded. It has a USB-C 10Gbs port as well, so the 512GB drive could be quickly moved to an external enclosure and repurposed.
    • This machine could very well be a second gaming PC for their living room as a console experience. So we would have to assume their main PC exists as well; With that said, I have 10gb home network with a 2.5gigabit internet connection here so we tend to have more than enough speed to download games. However, we can't make use of the 10gb LAN using Steam's built in transfer tool because it always compresses transfers and that slows the transfer down to well below a standard gigabit port speeds, sometimes as slow as 200-300Mb/s transfers. While that's probably still faster than most internet connections anyway, if they'd fix the LAN transfer issue it'd be upto x5 faster even on a gigabit LAN, than simply dropping a 2.5gbe port on there with hopes of a few people having fast internet connections. There are solutions, work arounds, like using LANCache if you run a NAS... or simply copying the files over manually using a network share.
    • Samsung announces ultra-fast UFS 5.0 storage to supercharge mobile AI by Paul Hill Local AI models tend to run a lot more slowly than cloud services like Claude and Gemini; however, Samsung has just announced that it has developed its UFS 5.0 solution, which increases data transfer to speeds of 10.8GB/s, enabling faster storage and processing in mobile memory that has the potential to provide more optimal local AI experiences. Commenting on this development, Jangseok Choi, head of Memory Product Planning at Samsung Electronics, said: If you’ve tried local AI, you’ll know it can be quite slow, especially if using the larger parameter models. By developing this new solution, Samsung says that storage is evolving from just storing data to a core piece of infrastructure that supports AI computation, too. The Korean company said that UFS 5.0 integrates the latest embedded memory interface standard from JEDEC and achieves up to 10.8 gigabytes per second (GB/s) transfer speeds. Regarding write speeds, Samsung UFS 5.0 can reach 9.5 GB/s. Both the read and write speeds are twice as fast as those of the previous UFS 4.1 standard. Aside from being ideal for local AI, Samsung’s UFS 5.0 is more power efficient by 40% compared to UFS 4.1. Samsung achieved this by implementing innovations such as clock gating and multi-voltage technologies. UFS 5.0 is also ultra-compact at just 7.5mm x 13mm x 0.9mm; that is 16.7% smaller than UFS 4.1. The company said it will be bringing it to multiple devices in the future, including mobile, wearable, and extended reality.
    • A bit like the steamdeck, this probably isn't for you.
  • Recent Achievements

    • Week One Done
      Almohandis earned a badge
      Week One Done
    • Rookie
      dorf went up a rank
      Rookie
    • First Post
      mike_rumble earned a badge
      First Post
    • Dedicated
      tuben earned a badge
      Dedicated
    • Week One Done
      mnsgroup earned a badge
      Week One Done
  • Popular Contributors

    1. 1
      +primortal
      496
    2. 2
      +Edouard
      209
    3. 3
      PsYcHoKiLLa
      99
    4. 4
      Michael Scrip
      86
    5. 5
      neufuse
      69
  • Tell a friend

    Love Neowin? Tell a friend!