[jason]

First time for everything, I suppose. Yesterday, I recorded
a chamber music concert. Marantz PMD-671 & Rode NT4. I've
used this combo many times with excellent results. The
recording yesterday seemed just fine as it was rolling...it
wasn't until I got home that I discovered that three of the
seven files (one/track, where a track was an item on the
program) were reported as corrupted/unreadable on Windows
XP.

The 671 has a "3rd head" mode that decodes what's recorded
as it's recorded, just like a tape deck. Except... it
doesn't work at 88.2kHz/24 bits/sample.

Google revealed a ton of programs that purport to rescue
damaged data. They all seem to let you try for free, but
prohibit saving the result unless you buy the program.

Several of them gave up on the damaged files. One didn't. In
fact, it rescued two of three and they seem fine. $29... I
bought it.

It reclaimed the third one, but the result is nothing but
shrieking noise, except for a 2-second burst of applause at
the very end. The other two are fine.

I am betting that the data for that one are ok, but that the
..wav file structure is slightly damaged. (Else, why would
the applause at the end be ok?) A buddy suggested that I
fire up Linux and use low-level commands to poke around.
He'd recovered some "priceless" photos this way. Windows
..wav file format is well-documented. I think that surgical
use the linux dd command, and some others, just might let me
extract the data and rebuild it as a .wav file, so I'm
pursuing that. The other approach is to invite the two
players to my home to re-record the track. That might be
simpler in the long run.

So, my questions are these: does NV memory incorporate ECC?
And if not, why not? And, what's the majority opinion on the
reliability of these memory cards. I'd used this particular
Lexar card many times, so I don't know if the error was a
problem with the card or with the 671. It was running on
batteries, with fairly long mic cable runs, so I suspect
that static electricity may have played a part; no grounding
to bleed static. It was a very cold, dry day.

Before the 671, I used a Sony HD Minidisc for recording, and
it worked well. I think I could hook it to the Line Out or
Optical out on the 671 (I think they're active during
recording) for a backup. I haven't tried that yet, but I
sure will!


Thanks,

Jason

[soundhaspriority]

"jason" wrote in message
.. .
[snip]

So, my questions are these: does NV memory incorporate ECC?
And if not, why not? And, what's the majority opinion on the
reliability of these memory cards. I'd used this particular
Lexar card many times, so I don't know if the error was a
problem with the card or with the 671. It was running on
batteries, with fairly long mic cable runs, so I suspect
that static electricity may have played a part; no grounding
to bleed static. It was a very cold, dry day.

There are actually two types of flash:

1. NOR flash, which is made in very small capacities, and has individually
addressable cells. The cells are considered reliable, and there is no ECC.

2. NAND flash, used by all flash drives, cards, etc. The cells are
unreliable; it's part of the manufacturing process that they do not write
reliably, and it's all compensated for by writing the data in ECC,
error-correcting-code. NAND is only block-addressable, which gives the ECC a
nice long word length. The combination is very reliable, on paper.

Yet we all know flash drives fail. SD cards are particularly vulnerable,
because they rely on the external equivalent of a disk controller. But your
machine uses CF cards, which do contain integral controllers. If a CF card
fails, it's entirely the fault of the CF card.

There are two types of NAND flash cells: single layer, and multi-layer. The
multi-layer cells store three bits per cell, but have a short lifetime. If
you have a multilayer CF card, you could have worn it out. This would be
unlikely with a single-layer card.


Before the 671, I used a Sony HD Minidisc for recording, and
it worked well. I think I could hook it to the Line Out or
Optical out on the 671 (I think they're active during
recording) for a backup. I haven't tried that yet, but I
sure will!


Thanks,

Jason

It's always good to have a backup.

Bob Morein
(310) 237-6511

[George's Pro Sound Company]

"jason" wrote in message
.. .
First time for everything, I suppose. Yesterday, I recorded
a chamber music concert. Marantz PMD-671

disappointing to hear that
I record confrences for pay using a 671
but I do not record above 16 bit 44.1
and i always back my recording up with a minidisc
I reformat the cards prior to every recording
but I do not like the monitor feature
when I move my headphone from the 671 to the mixer if I forget to turn the
volume down a distracting bit of delayed audio is heard in the "VERY" quiet
room
george

[Gareth Magennis]

"Soundhaspriority" wrote in message
...

"jason" wrote in message
.. .
[snip]

So, my questions are these: does NV memory incorporate ECC?
And if not, why not? And, what's the majority opinion on the
reliability of these memory cards. I'd used this particular
Lexar card many times, so I don't know if the error was a
problem with the card or with the 671. It was running on
batteries, with fairly long mic cable runs, so I suspect
that static electricity may have played a part; no grounding
to bleed static. It was a very cold, dry day.

There are actually two types of flash:

1. NOR flash, which is made in very small capacities, and has individually
addressable cells. The cells are considered reliable, and there is no ECC.

2. NAND flash, used by all flash drives, cards, etc. The cells are
unreliable; it's part of the manufacturing process that they do not write
reliably, and it's all compensated for by writing the data in ECC,
error-correcting-code. NAND is only block-addressable, which gives the ECC
a nice long word length. The combination is very reliable, on paper.

Yet we all know flash drives fail. SD cards are particularly vulnerable,
because they rely on the external equivalent of a disk controller. But
your machine uses CF cards, which do contain integral controllers. If a CF
card fails, it's entirely the fault of the CF card.

There are two types of NAND flash cells: single layer, and multi-layer.
The multi-layer cells store three bits per cell, but have a short
lifetime. If you have a multilayer CF card, you could have worn it out.
This would be unlikely with a single-layer card.


Before the 671, I used a Sony HD Minidisc for recording, and
it worked well. I think I could hook it to the Line Out or
Optical out on the 671 (I think they're active during
recording) for a backup. I haven't tried that yet, but I
sure will!


Thanks,

Jason

It's always good to have a backup.

Bob Morein
(310) 237-6511



I recently bought a M Audio Microtrack 2 and 2 x 8GB Kingston CF cards, as I
need to make 8 hour uninterrpted recordings. I have been sent 2 replacement
machines, the last one has just been sent back for a refund, I've had enough
now, as I couldn't get any of the three with either card to reliably
complete the recordings - they would randomly hang and the card would be
empty on re-boot. Also, the first machines recording time elapsed display
was about 10% slow (!!!???), the second machine refused to boot up with one
of the cards, the second machine refused to boot with either of the cards,
though both cards have made numerous transfers of these large WAV's via a
USB reader and my laptop and desktop with no reported errors at all so far.

I hadn't realised CF cards are not considered reliable or I would never have
gone down this route in the first place. I don't know where the fault lies
here, the machines, the cards, or some incompatability/bugs. I'm not
impressed.



Gareth.

[philicorda[_6_]]

On Tue, 03 Feb 2009 04:31:01 +0000, jason wrote:

First time for everything, I suppose. Yesterday, I recorded a chamber
music concert. Marantz PMD-671 & Rode NT4. I've used this combo many
times with excellent results. The recording yesterday seemed just fine
as it was rolling...it wasn't until I got home that I discovered that
three of the seven files (one/track, where a track was an item on the
program) were reported as corrupted/unreadable on Windows XP.

The 671 has a "3rd head" mode that decodes what's recorded as it's
recorded, just like a tape deck. Except... it doesn't work at 88.2kHz/24
bits/sample.

Google revealed a ton of programs that purport to rescue damaged data.
They all seem to let you try for free, but prohibit saving the result
unless you buy the program.

Several of them gave up on the damaged files. One didn't. In fact, it
rescued two of three and they seem fine. $29... I bought it.

It reclaimed the third one, but the result is nothing but shrieking
noise, except for a 2-second burst of applause at the very end. The
other two are fine.

Sometimes this can be an 'off by one' problem.
It might be reading the 16 lower bits of one sample, and the 8 higher
bits of the next one as a single 24bit sample. Sounds like horrible
noise. Another missing byte near the end may have put it back in sync.


I am betting that the data for that one are ok, but that the .wav file
structure is slightly damaged. (Else, why would the applause at the end
be ok?) A buddy suggested that I fire up Linux and use low-level
commands to poke around. He'd recovered some "priceless" photos this
way. Windows .wav file format is well-documented. I think that surgical
use the linux dd command, and some others, just might let me extract the
data and rebuild it as a .wav file, so I'm pursuing that. The other
approach is to invite the two players to my home to re-record the track.
That might be simpler in the long run.

I've done similar things with dd for recovering bad cds. You have to use
the flags 'notrunc' (do not truncate after errors), 'noerror' (ignore
errors) and sync (so the blocks are padded with nulls to the right size
if data is missing, could fix the 'shrieking noise').

I can't remember the exact invocation I used.

Rerecording is probably more fun.


So, my questions are these: does NV memory incorporate ECC? And if not,
why not? And, what's the majority opinion on the reliability of these
memory cards. I'd used this particular Lexar card many times, so I don't
know if the error was a problem with the card or with the 671. It was
running on batteries, with fairly long mic cable runs, so I suspect that
static electricity may have played a part; no grounding to bleed static.
It was a very cold, dry day.

Before the 671, I used a Sony HD Minidisc for recording, and it worked
well. I think I could hook it to the Line Out or Optical out on the 671
(I think they're active during recording) for a backup. I haven't tried
that yet, but I sure will!


Thanks,

Jason

[jason]

In article ,
says...



There are actually two types of flash:

1. NOR flash, which is made in very small capacities, and has individually
addressable cells. The cells are considered reliable, and there is no ECC.

2. NAND flash, used by all flash drives, cards, etc. The cells are
unreliable; it's part of the manufacturing process that they do not write
reliably, and it's all compensated for by writing the data in ECC,
error-correcting-code. NAND is only block-addressable, which gives the ECC a
nice long word length. The combination is very reliable, on paper.

Thanks, Bob. I had tried to determine whether there was ECC and hadn't
been able to find a definitive answer with google.

Yet we all know flash drives fail. SD cards are particularly vulnerable,
because they rely on the external equivalent of a disk controller. But your
machine uses CF cards, which do contain integral controllers. If a CF card
fails, it's entirely the fault of the CF card.

There are two types of NAND flash cells: single layer, and multi-layer. The
multi-layer cells store three bits per cell, but have a short lifetime. If
you have a multilayer CF card, you could have worn it out. This would be
unlikely with a single-layer card.

The card is quite new. I had formatted it just before I made the
recording. The 671 has a Card Check feature and it passed. The check
also returns a performance number, and the pro Lexar cards "should" have
plenty of margin, even at 88.2/24.


It's always good to have a backup.

Amen, brother...


I buddy has used low-level Linux commands to rescue "priceless" digital
photos from CF cards. I have Linux installed, but it hadn't occured to
me to try it. .wav files have a well-documented format, so if I can suck
off all the data with the byte-level dd copy command, and assuming that
the data were recorded sequentially (a good bet, since the card had just
been formatted), I may be able to effect a rescue. Then again, it may be
simpler to invite the two performers to my home to re-record the
track...

[Scott Dorsey]

Gareth Magennis wrote:

I hadn't realised CF cards are not considered reliable or I would never have
gone down this route in the first place. I don't know where the fault lies
here, the machines, the cards, or some incompatability/bugs. I'm not
impressed.

So far I have not seen any recording system that I am willing to run without
a backup, save for the Nagra. I have seen some big CF issues but no worse than
what I have seen with DAT.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Gareth Magennis]

"Soundhaspriority" wrote in message
...

"Gareth Magennis" wrote in message
...

"Soundhaspriority" wrote in message
...

"jason" wrote in message
.. .
[snip]

So, my questions are these: does NV memory incorporate ECC?
And if not, why not? And, what's the majority opinion on the
reliability of these memory cards. I'd used this particular
Lexar card many times, so I don't know if the error was a
problem with the card or with the 671. It was running on
batteries, with fairly long mic cable runs, so I suspect
that static electricity may have played a part; no grounding
to bleed static. It was a very cold, dry day.

There are actually two types of flash:

1. NOR flash, which is made in very small capacities, and has
individually addressable cells. The cells are considered reliable, and
there is no ECC.

2. NAND flash, used by all flash drives, cards, etc. The cells are
unreliable; it's part of the manufacturing process that they do not
write reliably, and it's all compensated for by writing the data in ECC,
error-correcting-code. NAND is only block-addressable, which gives the
ECC a nice long word length. The combination is very reliable, on paper.

Yet we all know flash drives fail. SD cards are particularly vulnerable,
because they rely on the external equivalent of a disk controller. But
your machine uses CF cards, which do contain integral controllers. If a
CF card fails, it's entirely the fault of the CF card.

There are two types of NAND flash cells: single layer, and multi-layer.
The multi-layer cells store three bits per cell, but have a short
lifetime. If you have a multilayer CF card, you could have worn it out.
This would be unlikely with a single-layer card.


Before the 671, I used a Sony HD Minidisc for recording, and
it worked well. I think I could hook it to the Line Out or
Optical out on the 671 (I think they're active during
recording) for a backup. I haven't tried that yet, but I
sure will!


Thanks,

Jason

It's always good to have a backup.

Bob Morein
(310) 237-6511



I recently bought a M Audio Microtrack 2 and 2 x 8GB Kingston CF cards,
as I need to make 8 hour uninterrpted recordings. I have been sent 2
replacement machines, the last one has just been sent back for a refund,
I've had enough now, as I couldn't get any of the three with either card
to reliably complete the recordings - they would randomly hang and the
card would be empty on re-boot. Also, the first machines recording time
elapsed display was about 10% slow (!!!???), the second machine refused
to boot up with one of the cards, the second machine refused to boot with
either of the cards, though both cards have made numerous transfers of
these large WAV's via a USB reader and my laptop and desktop with no
reported errors at all so far.

I hadn't realised CF cards are not considered reliable or I would never
have gone down this route in the first place. I don't know where the
fault lies here, the machines, the cards, or some incompatability/bugs.
I'm not impressed.



Gareth.

Gareth, I didn't mean to imply that CF cards are unreliable. The scheme of
unreliable cells combined with massive error correction combines to
produce a reliable media (on paper.) People in the r.a.m.p.s. forum use
these cards extensively. There are failures, and my personal impression is
that the user experience with flash is not as good as with hard disk.
However, in the case of low priced flash recorders, I would suspect a
design flaw in the recorders. A flash card requires steady power at a good
amperage to write.

It would be easy to be critical of the Microtrak after a look at the
plastic case, and how it's put together on the inside. But that isn't the
answer either. It is probably a case of the notorious Chinese quality
control. One production run will be good, the next, bad. Many well
reviewed products end up causing user misery that is not reflected in the
reviews, because the quality control slipped after the first run.

I wish I had an easy solution for you. It will probably take a lot of
persistence by you with M-Audio to get them to make good.

Bob Morein
(310) 237-6511




I've given up on the Microtrack, no point getting any more from the same
batch. Pity, because for the price and features it is very good, and one of
the very few that will record more that 2GB without stopping.

As it turns out, my Laptop (oldish Sony Vaio, Tascam USB interface, XP,
Audition) has been faultless - the Microtrack was bought because I didn't
consider a laptop to be reliable enough.




Cheers,


Gareth.

[philicorda[_6_]]

On Tue, 03 Feb 2009 16:31:19 +0000, jason wrote:

snip
I buddy has used low-level Linux commands to rescue "priceless" digital
photos from CF cards. I have Linux installed, but it hadn't occured to
me to try it. .wav files have a well-documented format, so if I can suck
off all the data with the byte-level dd copy command, and assuming that
the data were recorded sequentially (a good bet, since the card had just
been formatted), I may be able to effect a rescue. Then again, it may be
simpler to invite the two performers to my home to re-record the
track...

Try making an image of the drive with dd (with padding/error ignore etc)
and mount the image as a loopback device. If enough of the filesystem
remains intact you should be able to copy the files as normal from the
image with valid headers, though they will have small missing sections.
This is easier than hunting through the bytes with a hex editor.

You can also try other filesystem recovery operations on the image
without risking the original data.

[Robert Morein[_2_]]

"Gareth Magennis" wrote in message
...

"Soundhaspriority" wrote in message
...

"jason" wrote in message
.. .
[snip]

So, my questions are these: does NV memory incorporate ECC?
And if not, why not? And, what's the majority opinion on the
reliability of these memory cards. I'd used this particular
Lexar card many times, so I don't know if the error was a
problem with the card or with the 671. It was running on
batteries, with fairly long mic cable runs, so I suspect
that static electricity may have played a part; no grounding
to bleed static. It was a very cold, dry day.

There are actually two types of flash:

1. NOR flash, which is made in very small capacities, and has
individually addressable cells. The cells are considered reliable, and
there is no ECC.

2. NAND flash, used by all flash drives, cards, etc. The cells are
unreliable; it's part of the manufacturing process that they do not write
reliably, and it's all compensated for by writing the data in ECC,
error-correcting-code. NAND is only block-addressable, which gives the
ECC a nice long word length. The combination is very reliable, on paper.

Yet we all know flash drives fail. SD cards are particularly vulnerable,
because they rely on the external equivalent of a disk controller. But
your machine uses CF cards, which do contain integral controllers. If a
CF card fails, it's entirely the fault of the CF card.

There are two types of NAND flash cells: single layer, and multi-layer.
The multi-layer cells store three bits per cell, but have a short
lifetime. If you have a multilayer CF card, you could have worn it out.
This would be unlikely with a single-layer card.


Before the 671, I used a Sony HD Minidisc for recording, and
it worked well. I think I could hook it to the Line Out or
Optical out on the 671 (I think they're active during
recording) for a backup. I haven't tried that yet, but I
sure will!


Thanks,

Jason

It's always good to have a backup.

Bob Morein
(310) 237-6511



I recently bought a M Audio Microtrack 2 and 2 x 8GB Kingston CF cards, as
I need to make 8 hour uninterrpted recordings. I have been sent 2
replacement machines, the last one has just been sent back for a refund,
I've had enough now, as I couldn't get any of the three with either card
to reliably complete the recordings - they would randomly hang and the
card would be empty on re-boot. Also, the first machines recording time
elapsed display was about 10% slow (!!!???), the second machine refused to
boot up with one of the cards, the second machine refused to boot with
either of the cards, though both cards have made numerous transfers of
these large WAV's via a USB reader and my laptop and desktop with no
reported errors at all so far.

I hadn't realised CF cards are not considered reliable or I would never
have gone down this route in the first place. I don't know where the
fault lies here, the machines, the cards, or some incompatability/bugs.
I'm not impressed.



Gareth.

Gareth, I didn't mean to imply that CF cards are unreliable. The scheme of
unreliable cells combined with massive error correction combines to produce
a reliable media (on paper.) People in the r.a.m.p.s. forum use these cards
extensively. There are failures, and my personal impression is that the user
experience with flash is not as good as with hard disk. However, in the case
of low priced flash recorders, I would suspect a design flaw in the
recorders. A flash card requires steady power at a good amperage to write.

It would be easy to be critical of the Microtrak after a look at the plastic
case, and how it's put together on the inside. But that isn't the answer
either. It is probably a case of the notorious Chinese quality control. One
production run will be good, the next, bad. Many well reviewed products end
up causing user misery that is not reflected in the reviews, because the
quality control slipped after the first run.

I wish I had an easy solution for you. It will probably take a lot of
persistence by you with M-Audio to get them to make good. If you need some
help harassing them, I am very experienced at intimidating weak people to do
my bidding.

Bob Morein
(310) 237-6511

[Arny Krueger]

"jason" wrote in message


So, my questions are these: does NV memory incorporate
ECC?

I doubt it. As a rule, consumer RAM does not use ECC. Hard drives do.

And if not, why not?

The cost of the extra bits and circuitry to manage it does not seem to have
a reasonable return for almost all single-user applications and even when
there is workgroup-level sharing. ECC is primarly used in high end servers
where 100's or 1,000s of users are impacted by random failures.

How long does your PC run between *having* to be booted? I have some that
only get booted for major hardware or software changes, physical moves and
power failures.

And, what's the majority
opinion on the reliability of these memory cards.

There are a jillion brands of flash cards. They use a small number of
sources for flash chips and interface chips, but there is still probably
plenty of mixing and matching.

I've never had one do anything but either work or totally fail. That
includes trips through pretty heavily chlorinated water in our municipal
swimming pool. That one worked after a reasonable drying cycle in my food
dehydrator.

I'd used this particular Lexar card many times, so I don't
know if the error was a problem with the card or with the
671.

I find this story interesting. I have a Microtrack, and it has been
flawless. One of my associates never could be comfortable with his
Microtrack, and eventually got a Marantz flash-based recorder. AFAIK his has
also been flawless.

It was running on batteries, with fairly long mic
cable runs, so I suspect that static electricity may have
played a part; no grounding to bleed static. It was a
very cold, dry day.

Good equipment should tolerate all that, no sweat. But back in the real
world - live and learn!

[jason]

In article ,
says...

Sometimes this can be an 'off by one' problem.
It might be reading the 16 lower bits of one sample, and the 8 higher
bits of the next one as a single 24bit sample. Sounds like horrible
noise. Another missing byte near the end may have put it back in sync.

Phil, it was the thought this might be an off by one issue
that made me think there was some hope for rescuing the
data..especially considering the pristine final couple of
seconds. You give me encouragement to proceed. :-)


Rerecording is probably more fun.

It will become more and more fun as I beat my head against
the wall...

[jason]

In article ,
says...


Phil, it was the thought this might be an off by one issue
that made me think there was some hope for rescuing the
data..especially considering the pristine final couple of
seconds. You give me encouragement to proceed. :-)

Success! ...well, partial. I read up on the .wav RIFF format
and was able to rescue almost everything by inserting a
single byte (=0) right after the header. I'm still missing a
bit at the end, but everything else is perfect. The data are
not out of order, just truncated. I will now dig a little
deeper into the format and see if the file is short because
sample counts are messed up somewhere. I suspect that's the
case now.

I will fax a beverage of your choice, Phil!

Jason

[jason]

In article ,
says...


I doubt it. As a rule, consumer RAM does not use ECC. Hard drives do.

I learned here that CF cards, unlike some other flash memory
packages (SD cards) incorporate their own DMA controller and
do incorporate ECC. Still didn't help... dangit.

[Arny Krueger]

"jason" wrote in message
.. .
In article ,
says...


I doubt it. As a rule, consumer RAM does not use ECC. Hard drives do.

I learned here that CF cards, unlike some other flash memory
packages (SD cards) incorporate their own DMA controller and
do incorporate ECC.

I was wrong about ECC not being used:

http://en.wikipedia.org/wiki/Flash_memory

"NAND (flash memory) relies on ECC to compensate for bits that may
spontaneously fail during normal device operation. This ECC may correct as
little as one bit error in each 2048 bits, or up to 22 bits in each 2048
bits.[7] If ECC cannot correct the error during read, it may still detect
the error. When doing erase or program operations, the device can detect
blocks that fail to program or erase and mark them bad. The data is then
written to a different, good block, and the bad block map is updated.

"Most NAND devices are shipped from the factory with some bad blocks which
are typically identified and marked according to a specified bad block
marking strategy. By allowing some bad blocks, the manufacturers achieve far
higher yields than would be possible if all blocks had to be verified good.
This significantly reduces NAND flash costs and only slightly decreases the
storage capacity of the parts.

Apparently NOR Flash is more relaible.