Topic: CF card write / read / copy problems? Multiple odd happenings...help?

[Brian Skalinder]

Any ideas appreciated on...

<a>  root cause / problem
<b>  recovery

Recorded a couple sets this weekend.  ACM HD-P2 / Kingston Elite Pro 8GB CF recorded just fine, though I had trouble with not enough gain for the C34 (as expected).  Had trouble with my V3 > stock HD-P2 / PQI 8GB CF.  What I've experienced / steps I've taken so far:

<01>  Original recording source

Encountered no noticeable problems during the initial recording session:  Schoeps MK5 / KC5 / CMC6 > ACM V3 (24/48) > stock HD-P2 / PQI Hi-Speed 100 8GB CF card

<02>  Transferred files

Transferred files from the PQI CF card to PC via Lexar Pro CF reader (firewire), not the HD-P2

<03>  Error message opening files
Upon attempting to open one of the files in CD-Wave, received an error like "Bad FMT descriptor" or some such.  Can't duplicate the error and forgot to note the specific text of the message.

<04> upon successful open:  noise / artifacts

Tried opening a second time in CD-Wave and...success.    So, I opened in Samplitude SE v8 to take a closer look.  However:  HUGE noise / artifacts starting at ~37min and 35min into each of the two ~45min sets.  See attached screenshots (01 = zoomed out, 02 = zoomed in a bit) and MP3 sample for a listen.  Caution:  the sound source is very quiet at the start;  the noise starts at the ~20 sec mark and it's LOUD (as you can see from the screenshots), so keep the volume down.

<05>  Re-transfer attempt

Tried inserting the CF card into the stock HD-P2 to re-transfer.  P2 indicates something like:  "Media needs formatting.  Format now?"  Of course, I did NOT format.

<06>  Ran CHKDSK

Put the card back in the CF reader and ran CHKDSK /F from the command line.  Message indicated it wasn't a Windows 2000 disk, asked if I wanted to continue, I said Yes.  Message asking whether I wanted to save the bad links / lost chains to a file and I said Yes.  The results:

Bad links in lost chain at cluster 28286 corrected.
Lost chain cross-linked at cluster 28287.  Orphan truncated.
Bad links in lost chain at cluster 28424 corrected.
... continue above line in sequence until ...
Bad links in lost chain at cluster 28542 corrected.
Lost chain cross-linked at cluster 28543.  Orphan truncated.
Bad links in lost chain at cluster 28680 corrected.
... continue above line in sequence until ...
Bad links in lost chain at cluster 28798 corrected.
Lost chain cross-linked at cluster 28799.  Orphan truncated.
Convert lost chains to files (Y/N)? y
   The root folder on this volume is full. To perform a disk check,
   Windows requires space in the root folder. Remove some files
   from this folder, then run disk checking again.
892239872 bytes in 14137 recoverable files.
Windows has made corrections to the file system.
    8,124,448 KB total disk space.
    7,252,768 KB are available.

       32,768 bytes in each allocation unit.
      253,889 total allocation units on disk.
      226,649 allocation units available on disk.
Cannot write boot sector.

<07>  Read card again

Attempted to read the card via the CF reader from the DOS prompt, and received the following message:

The volume does not contain a recognized file system.

Tried again in Windows and received a message that the card needs formatting. 

<08> Additional card read and copy attempt

Rebooted, tried to read the card again in Win Explorer via the CF reader and...success!  Card contents:

• Proj05 directory + XML files
• Proj05/Audio files directory + my recorded WAVs

However, tried to re-copy from CF card to PC and received the following error:

Cannot copy <filename>:  The file or directory is corrupted and unreadable

Tried copying the XML files for kicks, and they copied fine.  But the WAVs still do not.

<09>  Ran CHKDSK again

Ran this time not from the command line but from within Windows:  Properties | Tools | Check, ticked the Automatically fix file system errors checkbox, did not tick the Scan for and attempt recovery of bad sectors.  Result:  Windows was unable to complete the disk check.

<10>  Read CF card again

Again tried to read the CF card via Win Explorer, only to receive message (again):  The disk in drive L: is not formatted.  Do you want to format it now?  Of course, I said NO.

<11>  Read CF card once more

Unmounted the CF reader, re-mounted, and succesfully read the card again.  BUT, contents have now changed.  Card now includes:

• Proj05 directory + XML files (same as before)
• Gone:  the Proj05/Audio Files directory + WAV files
• New dir:  FOUND.000
  
• New files:  FILE000.CHK sequentially through FILE999.CHK, total size of 312 MB
  
  

Attempted to copy the entire directory structure and files only to receive the following message midstream:

Cannot copy file FILE8574:  Cannot find the specified file.  Make sure you specify the correct path and filename.

Re-read the card to verify FILE8574.CHK exists, and...FILE8574.CHK through FILE9999.CHK are GONE.

However, there's another new directory named:  gibberish*

* Copied the dir name into a text file and tried to include in my post, but TS won't preview or post with the funky characters in the thread.

Attempting to read the oddly named directory returns the following message:  L"\FOUND.000\<gibberish> is nto accessible.  The filename, directory name, or volume label syntax is incorrect.

Tried re-naming the directory to WTF and could not.  WTF!!

<12>  Read CF card once more

Unmounted the CF reader, re-mounted, and succesfully read the card again.  Files FILE8574.CHK through FILE9999.CHK are now...BACK.  Tried to copy FILE8574.CHK through FILE9999.CHK once more.  Result:  Program error.  Explorer.exe has generated errors and will be closed by Windows.  You will need to restart the program.

Tried copying small groups of files with some success:

FILE9999.CHK - FILE9900.CHK | Success!
FILE9899.CHK - FILE9800.CHK | Success!
FILE9799.CHK - FILE9700.CHK | Success!
FILE9699.CHK - FILE9600.CHK | Success!
FILE9599.CHK - FILE9500.CHK | Success!
FILE9499.CHK - FILE9400.CHK | Success!
FILE9399.CHK - FILE9300.CHK | Success!
FILE9299.CHK - FILE9200.CHK | Success!
FILE9199.CHK - FILE9100.CHK | Success!
FILE9099.CHK - FILE9000.CHK | Success!
FILE8999.CHK - FILE8900.CHK | Success!
FILE8899.CHK - FILE8800.CHK | Success!
FILE8799.CHK - FILE8700.CHK | Success!
FILE8699.CHK - FILE8600.CHK | Success!
FILE8599.CHK - FILE8574.CHK | Success!

So...eventually successfully copied all 10,000 files.

<13>  Ran CHKDSK once more, copy fragments

Since I successfully copied the file fragments after the previous CHKDSK, I ran it again for kicks.  This time, 525 files:  FILE0000.CHK - FILE0525.CHK.  In attempting to copy them across, turns out there are two (yes, 2) files each of FILE0510.CHK through FILE0525.CHK.  Tried to copy a single set of FILE0510.CHK > FILE0525.CHK, plannig to rename to FILE0510a.CHK > FILE0525a.CHK, then copy the duplicate set and rename to b.  BUT, when I select and copy only a single set of FILE0510.CHK - FILE0525.CHK, Windows copies a pair of each file.  WTF


And now...I give up, and am at a complete loss.  I was initially encouraged by the 10,000 recovered files after running CHKDSK in Windows the first time.  But then I realized the files only account for 312 MB.  The 525 files after the 2nd CHKDSK account for only 16 MB. 

And if you've gotten this far...

I'm at a loss.  I'd like to re-transfer, but can't at the moment. 

[JackoRoses]

having just been frustrated with a bad xfer myself on a cf card all I can recommend at this point is to see about trying to use a file recovery tool at this point.

I'm not even sure how one could go about merging the files found back into the original file to get to what you want at this point.
I would hope a file recovery tool would get past all that and be able to pull the complete files for you.
I have not ever encountered this issue yet, other than having the firewire xfer issue that I posted about in the team board.
Good luck in your recovery effort!

[flipp]

No idea on cause and the only suggestion I can offer on recovery is to try http://taperssection.com/index.php/topic,72936.msg976154.html#msg976154 and see if it can recover the contents of your card.

[Brian Skalinder]

Update:  Tried to fix with TestDisk.  No joy.  App said the partition was too badly damaged to repair.  Now trying PC Inspector Smart Recovery.

[flipp]

One more thing to try, stellar phoenix http://www.stellarinfo.com/
Not sure if it will work on a cf card but it found and recovered files on my hard-drive other recovery programs couldn't even see. Thanks to darkstarjedi for mentioning it here on ts a few months ago.

[Brian Skalinder]

One more thing to try, stellar phoenix http://www.stellarinfo.com/

Thanks for the tip.  Is the demo fully functional, or did you buy it?  Not sure recovering one song from each set is worth the $100, especially since we record again next weekend.

[flipp]

The demo will show if it sees the files. To recover them you will have to purchase. I did because the fee was much less than sending the hard-drive to a recovery service. Worth it for what needed to be recovered. For two audio files, probably not.

[Brian Skalinder]

Now trying PC Inspector Smart Recovery.

No joy.  Recovered some ~20 sec WAV pieces from previous recording sessions, that's all.  I'm thinking I won't succeed with any of these recovery tools - I suspect it was a write problem during recording, in which case no recovery attempt will yield useful results.  Trying to duplicate this week, and if I'm identify to identify the root cause, I'll let everyone know.

[pigiron]

The first part of this saga sounds like a classic case of FAT filesystem corruption... then it all went to hell from there.

The file name, length, etc are stored on a spot in the low addresses... and this "spot' also points to the first cluster containing the actual data... this cluster points to the next cluster, which points to the next, and so on.

http://en.wikipedia.org/wiki/File_Allocation_Table
http://www.pjrc.com/tech/8051/ide/fat32.html

So if something goes wrong anywhere in the chain (or somewhere else in the FATxx filesystem), a perfectly good audio pull get's totally trashed.

These "file repair" utilities attempt to find the spot that's in error and fix it... but it's very easy for them to get confused and make the situation much worse... and I think maybe you just proved it 

yes, i'm a linux bigot, but if (hope the gods don't let this happen) you do recreate it, i've read multiple forums where the guys that support thousands of boxes carry a linux LiveCD for a situation like this... they swear that they've been able to successfully pull the data off of drives that windoze says doesn't even exist anymore.

will it work? who knows... but it just might be worth a try if your back's against the wall... but to be a fair shake, try it before the "utilities" have gotten their grubby little paws on it 

here's hoping that none of us see this again

[Brian Skalinder]

yes, i'm a linux bigot, but if (hope the gods don't let this happen) you do recreate it, i've read multiple forums where the guys that support thousands of boxes carry a linux LiveCD for a situation like this... they swear that they've been able to successfully pull the data off of drives that windoze says doesn't even exist anymore.

will it work? who knows... but it just might be worth a try if your back's against the wall... but to be a fair shake, try it before the "utilities" have gotten their grubby little paws on it 

Good info, pigiron - thanks.  I'm not averse to Linux, so I may go that route next time I run into trouble (hoping it doesn't happen, of course!).  It seems as though I was able to transfer all the data off the card - proper file size, data type, proper audio throughout the problem section of the waveform, but just interspersed with gobs intermittent noise.  At any rate, I'll definitely consider a Linux toolset next time. 

[Brian Skalinder]

FWIW, I've not experienced this problem again in my field use or testing.  I suspect it may have had to do with my ejecting the CF card without unmounting the project / media first - which the HD-P2 manual warns against - but can't duplicate.