R-44 SDHC problems

[Tommy-boy]

The other day I lost an hour of recording on my r-44. I recorded sucessfully for approx an hour then received the following error message: "SD Card Slow". So I stopped the recording and popped in another idential card, which worked fine.

There is an audio file for the recording on the first card, but I can't copy it onto my pc or import it into cubase as I get the following error message on my PC: "Cannot copy 091225181905: The file or directory is corrupted and unreadable."

Fortunately, this wasn't an overly important recording. It was an informal get together with my family. However, I'm concerned because I don't want this to happen to a recording that really counts.

I use 2 cards which are identical. Kingston 8 GB SCHD cards which are Class6 compliant. Class6 cards are the fast ones and should never have speed problems, especially when I'm only recording 2 channels of audio 24bit/44.1Khz.

I've recorded numerous takes with these cards and have never had a problem. Why the problem now? A perusal of r-44 FAQs on Edirol's site notes the following:

Sometimes I get a "SD Card Slow" error on the display. Also: sometimes I'll hit the STOP button but the project won't stop recording. (submitted 12/19/2008)
It's likely that your SD card is either not compatible with tthe R-44, or your SD card may simply need to be formatted. You should save any necessary projects to your computer, then format the SD card in the R-44. If the problem persists, your SD card is likely incompatible or corrupt. Try another SD card.

Any other r-44 users have this problem? Do you guys reformat the SDHC cards after each use to get them ready for the next recording?

Any thoughts for how to salvage my one seemingly unreadable wav file from the card?

Thanks,
Tom

[springer]

Hi Tom,
I have had the same problem on a critical recording and was able to fudge the file when it was connected to my PC. I "think" that I renamed the file and then was able to copy it???? I did this while it was in USB mode and was able to salvage the file up to it's failure point. Something simple along those lines.

I would LOVE more people to chime in on the class6 vs class4 cards on this unit. I think the majority of the Edirol "proven list" cards are class4?
I didn't run into problems until I did 24/96 with 4 tracks. I ALWAYS format now before every start. I still have had problems with 4 tracks at 24/96 using a class6 card that is not on the list. I do 24/44.1 now.

[NorseHorse]

I've used Class 4 and Class 6 cards at various sample rates. No problems yet, but maybe I should look at the list?

And yes, I always reformat before recording. This is also important if you use your SD cards for other things. For instance, my Canon Vixia uses SD cards. If I have an SD card with the video files on it, the R44 will not tell me that the card is already __% full, giving you the impression its an empty card when it isn't. Tabula rasa is the best way to record IMHO.

[Tommy-boy]

Originally Posted by NorseHorse

...
And yes, I always reformat before recording. This is also important if you use your SD cards for other things.

Do you always reformat the card in the r-44 using the r-44 format function (as opposed to reformatting when connected to a PC with a card reader)?

Just for reference, I've never used the two 8 gig cards for anything other than recording in the r-44. I would delete files after copying them to my pc, then record more. Maybe there's something about this process that causes problems.

Hmmm.....

-Tom

[NorseHorse]

Yes, I always reformat using the R-44's reformat function. Just because your computer "deleted" the info from the last show doesn't mean the card is formatted to accept hours of 24/96. And my example above demonstrates that even if it looks like there's plenty of space on the card, spurious files or fragmented drives could potentially get in the way if it isn't reformatted on the R44 before use.

In my mind, reformatting on the R-44 is akin to defragging. Which is a VERY good idea for any type of drive used for recordings.

[Tommy-boy]

That settles it then. I will reformatt my cards in the machine before each use.

Thanks for the heads-up.

-Tom

[Tommy-boy]

Some more info for anybody else who might have similar problems.....

I was able to recover the files from the SD card. I did this by putting the card in a USB reader. In Acronis Disk Director I searched the card for errors with the selection to fix any found errors. It found the errors and fixed them.

I then tried to bring the file into Wavelab, but had no joy. Fortunately Cubase was able to import the audio file. So that's how I got the file. There were some skips in the audio file, but at least I had something, rather than nothing.

This whole experience led me to RTFM, again. Pg 77 notes "as you repeatedly write and delete project files, the files on the SD memory card become fragmented, and this slows down the processing speed. If you continue recording in this state, a message of 'SD Card Slow!' is shown, and skips may occur in the recorded project." Sure enough - that was my problem. Never to be repeated as I'll always format before recording anything from now on!

-Tom

[Mark A. Jay]

Originally Posted by NorseHorse

Yes, I always reformat using the R-44's reformat function. ...In my mind, reformatting on the R-44 is akin to defragging. Which is a VERY good idea for any type of drive used for recordings.

I agree with this, and generally do this as well (format) before even packing the cards and heading out to the gig.

However, what I don't understand (and I have seen the "SD card slow" message on the R-44 once, but once was enough - it was during a gig) is that this happened on a card that was purportedly 'class 6' and that had been used maybe 2-3 times; basically, the card was 'new'. I was minding other things, and had not keep a constant eye on the display, so when I finally did turn back and look...that's when I saw the message. Bummer.

Yet, I have used SD cards (on the R-44) that are marked below class 6 and never had an issue.

So, in my case, despite using the format commmand in the R-44, that card produced errors. Fortunately, I had brought other cards with me and was able to swap cards and 'cut my losses', but that part of the gig was turned into discontinuous files.

So, is it possible that one could use something like Spinrite (with the SDHD card in a usb adapter) and give the card a once-over? The idea here being that bad 'sectors' (addresses) on the card could be marked as such by spinrite and thus, rendered unavailable when the system tries to access them.

Granted, you would lose some of the memory on the card, but if it solves the 'card slow' issue (by making the questionable addresses unavailable) and makes the card 'reliable' that would probably be worth the time spent doing it - at least the system would know those addresses are not relaible and marked as 'bad', and thus, no write attempts would be made there.

Has anyone else tried this?

Granted, it could simply be that the card in question was (is) bad, and even after an attempt with spinrite (assuming it would even work on an SDHC card) would still be problematic...and that SD cards are cheap, but I have used spinrite on hard drives before, and it's always done the trick (unless something is really wrong with the hard drive).

I can say this...the SD cards that I have purchased made by Sandisk have functioned perfectly, regardless of class.

Also, from what I have read, the classification of the cards (class 2, 4, 6 etc) can be somewhat 'fast and loose' and often times, a card that is labeled as a class 6 does not seem to perform as such.

Anyone?

[Ozpeter]

Fake cards are rife, I've heard. If possible buy from a respected supplier even if that means paying full whack.

[Mark A. Jay]

Originally Posted by Ozpeter

Fake cards are rife, I've heard. If possible buy from a respected supplier even if that means paying full whack.

Paying full whack is WELL worth it when they ae the real deal...but...how would one know with certainty?

[mathieujm]

Originally Posted by NorseHorse

And my example above demonstrates that even if it looks like there's plenty of space on the card, spurious files or fragmented drives could potentially get in the way if it isn't reformatted on the R44 before use.

In my mind, reformatting on the R-44 is akin to defragging. Which is a VERY good idea for any type of drive used for recordings.

Fragmented file is a problem on drives because of the heads shifting. But it's not a pb on cards. The real problem on cards is that the number of write cycles for each octet is limited, so the card controler manage (and fragment !) the files to minimize the number of write cycle for each memory cell. It's called Wear leveling. The important thing here is that performances change with wear.
Look at this document
http://www.flashmemorysummit.com/Eng..._T1B_Smith.pdf

JMM

[Mark A. Jay]

Originally Posted by mathieujm

Fragmented file is a problem on drives because of the heads shifting. But it's not a pb on cards. The real problem on cards is that the number of write cycles for each octet is limited, so the card controler manage (and fragment !) the files to minimize the number of write cycle for each memory cell. It's called Wear leveling. The important thing here is that performances change with wear.
Look at this document
http://www.flashmemorysummit.com/Eng..._T1B_Smith.pdf

JMM

I didn't know the formal name, but now that you mention it, I was told this same thing about the solid state hard drive in my laptop - that is, I should not de-frag it for the reasons that you have cited.

Anyway, my principal complaint is that if there is indeed a 'standard' that the cards must follow (and thus indicate which cards are best suited), then why are manufacturers not bound to demonstrate conformance with the standard?

After all, a standard without enforcement is merely a suggestion...is it not?

[Tommy-boy]

I've been reformatting on the R-44 just prior to recording and haven't had any more problems. I also bought a splitter and record into a little netbook as a backup (hopefully to be soon replaced with another R-44).

-Tom

[leddy]

Had this problem over the weekend on a pretty important recording. I can copy the file over, but parts of it are missing, as it is out of sync with another R44 it was slaved to.

I can make do with the tracks I have. The SD card was one I had used many times without issue. Do they just fail over time?

[rcrowley]

Yes, SD cards fail over time. They have limited number of read/write cycles. Furthermore, they are NOT "archival" media. I would NOT trust flash memory beyond 3-5 years at the very most.

Solid-state memory IS subject to fragmentation. Fragmentation is a function of the disk format, and not the storage medium. However fragmentation is a DIFFERENT kind of problem in solid-state memory as it is in rotating memory. There is a lot of discussion of this out on the interweb, especially now that SSD are becoming more common in computers.

Just as fragmentation causes problems in rotating memory because the head must physically move around hunting for all the sectors, solid-state memory also operates only in rather large blocks which must be completely written and read at once.

Formatting IN the recording device is always the best way of avoiding problems.

[leddy]

Originally Posted by rcrowley

Formatting IN the recording device is always the best way of avoiding problems.

I always format right before the gig. Formatting could not be the issue.

The card is probably a couple of years old though.

[Tommy-boy]

Since I started this thread, I've been reformatting the disk in the R44 just prior to recording. I also use the high speed San Disk brand disks. I have had zero problems since.

On a side note - I now run a 788T in tandem (using Radial JS3s to split signals). I was going to get another R44 (perhaps coupled with a monitoring mixer) but decided to stretch a bit and get a device with more inputs and monitoring functionality. I use the 788T as the main recorder and the R44 as the backup (with gain settings lower incase I set 'em too high on 788T). So far, so good!

-Tom