13 threads on CORRUPT frames in less than 2 MONTHS

[Conrad Hunziker]

Ive got to second Greg's post. Ive been privy to a lot of behind the scenes discussions as real people on big and small shoots try and figure out what when wrong. In just about all cases it was a hardware error (reader, cable, etc) or a workflow error.

In every case where the DITs had a proper workflow that included visually verifying in real time before formatting, no footage was lost.

My software makes it easy and does 90% of the work for you, but you've got to visually verify that all the zeros and ones add up to an image.

 

[Gunleik Groven]

Must agree with Conrad here.

After switching from the other tool to r3d data manager, the number of faults almost went away. BUT you CAN get verified copies off r3d data manager that contains faults which are not on the original media.

Transcoding as you copy, picks up all the problems I've seen.

I have not myself ever VERIFIED that a fault was on the original media since build 12/13 - I think. User error has accounted for all the things I have verified. That is not to say that it doesn't happen. I just haven't seen it myself.

I've seen a few bad frames due to unconcious data-handling, though...

 

[Don Harris]

All I was saying is that Nikon would have a PR nightmare if they stated data integrity was a low priority- especially when trying to sell cameras to wedding shooters. I am however, quite excited about adding the Scarlet to my business. And from the posts I have read from the Red Team members, it seems unlikely they would not hold data integrity as a high priority. I just wish RED could make a camera that was as good at stills as my D3x (or at least my D700) while still offering wonderful ff35 video- and at less than 10k for the whole "kit." That is my dream. When this happens, heads wont just turn, they will roll! Keep going Red Team! A lot of still photographers like me are rooting for you behind the scenes!

 

[Greg M]

FYI- we have been using R3D Data manager since the first day it was released. Thanks Conrad for an indispensable tool.

OH at most recent Red prep... 2nd AC to "DATA MGR"-
"dont use that R3D Data Manager tool they are recommending...it takes to long. Just drag and drop"

Our response was immediate to the "DATA MGR"
"DO NOT LISTEN TO THE 2nd, HE IS CLUELESS"

The real problem here is the "DATA MGR", which is in quotes because in most instances he has zero training.
IMO- The management of data is one of the most important jobs on set. To hire an individual with limited experience here is just foolish.

 

[Conrad Hunziker]

BUT you CAN get verified copies off r3d data manager that contains faults which are not on the original media.

Yep, which is why the workflow is so important. Ive always maintained that you must include a visual verification of all the footage in your workflow. Its entirely possible to have a bad reader read the data wrong twice, which would produce a correctly checummed copy that doesnt add up to a valid image. Visually verifing catches these, and usually letting the card/reader cool, or switching readers solves the issue.

In these 13 threads mentioned here, I dont see one that says they visually verified the COPY before deleting the original media.

Since 2007 Ive put more than 4Pb of material through my software, using this workflow. Ive had 9 copy errors, none of which resulted in lost footage

 

[Kim Frank]

As long as idiots operate these cameras, there will be corrupt footage.

If you take the time to learn your craft and follow basic rules you will not have issues.

Well Greg, I'm glad for you that you are obviously smarter than me and know your craft better.

But I don't know you, so I'll put your offensive statement aside.

I never had any issues when shooting with RED and suddenly I did.
I simply didn't know that the crashing of the Application on set was up to corrupt frames. We could scrub through the footage, so I intended it to be fine, but it wasn't.

In my Opinion RED would really help a lot if they would name that Issue, it's causes and soloutions in the manual.

 

[Greg M]

Well Greg, I'm glad for you that you are obviously smarter than me and know your craft better.

But I don't know you, so I'll put your offensive statement aside.

I never had any issues when shooting with RED and suddenly I did.
I simply didn't know that the crashing of the Application on set was up to corrupt frames. We could scrub through the footage, so I intended it to be fine, but it wasn't.

In my Opinion RED would really help a lot if they would name that Issue, it's causes and soloutions in the manual.

If it is really just up to the offloading Pipeline, why does the presents of this Problem increase. And please don't start that "crappy DIT" or "cheap sollutions" stuff again. I really dislike this attitude as long as you really KNOW it was the case and the threads don't state that.

Sorry, my comments were general in nature and not specifically to you (I was actually commenting on all your links). Truth is I didn't read the entire thread. I stand by what I said in the general sense, we see it every day where the person responsible for data has zero or very limited experience.

Regarding your specific issues, have you considered that you have some bad hardware? Bad cable, bad drive, etc.

From my experience, the problem is either a bad cable (or other hardware) or it is operator error. In either case the solution is usually quick to identify and easy to solve.

The first step is to call Red Support and discuss your problem with them. Send them a sample of the corrupt footage. Red support is first class and always responds immediatly to any problems.

 

[Kim Frank]

Thanks Greg...
My problem is solved, as I don't have them anymore.
I use the same hardware and it works.
It's really difficult to verify what exactly caused the problem.
But I learned it the hard way.
I just want others to not have to take the same route.
I don't say your wrong neither RED is with saying it's an offloading Pipline (including the person in charge) problem, but again I think it would be a great help if this would be made a subject in the manual.

 

[Jay Gannon]

I don't say your wrong neither RED is with saying it's an offloading Pipline (including the person in charge) problem, but again I think it would be a great help if this would be made a subject in the manual.

What we are saying is 99% of the time it is an offloading pipeline problem. RED is not there to hold your hand through data acquisition, DIT's are. So employ their services and you shall have no corrupt data.

 

[Jeff Kilgroe]

In my Opinion RED would really help a lot if they would name that Issue, it's causes and soloutions in the manual.

I guess it could be beneficial to some if RED were to print a notice about watching footage for verification, especially after copying. Beyond that, I don't know what else they could really say. Data integrity is always a concern and not by any means unique to RED, or even to high-end cameras. RED (or anyone) could write volumes on potential causes of data corruption and how to trace the cause or deal with the results. ...and many have. An application crash, potentially causing data corruption, is commonly accepted as a known hazard and reason to verify or even re-copy any potentially affected data.

I have been working with the RED One for over 2.5 years at this point. Most every data issue I have encountered has been user error. But not all -- I have seen a couple bad CF modules, bad RED Drive cables, bad card readers,FireWire and SATA cables, etc... Most anything related to the camera kit is noticed right away, it's the user errors in post or by un-trained or careless DITs that always get you. Faulty hardware and/or software do usually play a role. That is where user error often creeps in, as the data is not properly verified or maintained and problem hardware goes unchecked.

I'm not sure what Don was getting at with his comments relating to Nikon and PR nightmares... RED has always stated that data integrity is critically important to them. They take our data very serious and have gone above and beyond when called on to recover failed RED Drives and CF media. Nikon and Canon don't provide any such services, nor do their cameras provide any onboard data verification or control. Anyone who shoots tens of thousands of stills every year, as many of us here do, will have data issues at some point. I had one of my 32GB SanDisk CFs go bad during a time-lapse shoot a couple months ago, for example. The Nikon D3 didn't notice or care, it just kept snapping pictures..

 

[Roberto Lequeux]

Hi,

As one checks the gate after a circled take the no of unnoticed hairs would be zero. Generally film cameras are fairly simple, if they worked correctly when tested yesterday they will be fine today unless something stupid happened such as being dropped in the sea.

So that one time that we found a hair on a great Martini take was a dream. Only one I ever been on set for. Everyone laughed in confusion.
Since the negative couldn't be checked on the spot another take was rolled. Then the Director asked for a second for safety, can't blame him.

But I'll clarify for you that I wasn't saying film cameras are unreliable. My comments here are made in context of 99.9% + reliability, and I assumed everyone would at least look at film cameras as being in the same category.

Yours sincerely,
Another FUD purveyor (?!)

If it is really just up to the offloading Pipeline, why does the presents of this Problem increase.

Summer. That is my main guess. CF readers overheating more than usual.

And there's allays the stars aligning just so for more public voicing of negative experiences, or what have you... lol, let me know if you find a solid pattern. Might be nothing there other than heat.

I really get the point about the percentage of Red shoots vs. Posts about Corrupt frame issues, but not everyone that shoots RED is a Reduser. We don't know how many DOPs, Post facilities and Rental Houses deal with this issue every day and do not post it on the board.

There is some truth to that. Good point. However, there are also thousands and thousands registered under fake names here. And daily we see how many associate cameras getting more affordable with food being taken out of their mouths, or the economy, and how they come to take it out on RED here. That has to offset people "dealing with this problem every day" and not reporting it. But numbers would be hard to derive for either, so on this front everyone will have a different guess.

Let's say it is not a Recording error but happens while offloading, I think it would be very importand for RED to take this issue serious and offer some sollution. Easiest would be taking it in the manual for example. They could state that corrupt frames may appear and that it is a neccesarity to eye ball all takes, or at least the importand ones, after offloading, before reformatting the media.

Is check sum performed in-camera not enough? Didn't someone say nobody does it, and that EPIC may be the first? I remember when they said they were adding it to their list for the ASIC. Hopefully we get confirmation soon. I am no DIT, but I think this is one of the only things they can check on every take that is shot. And the equivalent for film cameras is to develop film.

It would perfectly fit into the OFFLOADING MEDIA section:


CAUTION: REDFLASH 16GB Compact Flash cards transfer Data at extremely high rates. RED reccomends a high performance, professional UDMA reader( from areputable manufacturer) when w downloading from 16GB CF cards. Using inexpensive or low-quality readers may cause damage to the cards and/or corruptin/loss of media files stored on them.

Edit something like:
To avoid the loss or curruption of any Footage, RED reccomends to watch the offloaded R3D Data in RedcineX to verify the complete data transfer, before reformatting the media. If corruption appears RED reccomends to switch out the cables and/or Card reader.


Something like this... I'm not native speaking, so please be patient. The issue also appeard with RED Drives, so it would need a Caution Warning on that section too. And I think it's such an importand thing to know, that it would need to be a point in the

POSSIBLE ERRORS, ISSUES AND SOLLUTIONS section.


Something like:
R3Ds won't Playback after offloading.
Or
Green and/or destorted frames in the Footage.


RED?
Someone?
You could save some people some serious problems...

All of this, and excuse me if I assume a base understanding, but sounds like making sure to wash your hands before you cook. I had internet in High School, not for all of it he he he, but I partially developed my core personality in an electronic world. Same deal with mechanics: lubricating, checking for wear, performing routine replacements of disposable parts, doing tune-ups. With electronics: check your cables are plugged in right (funny, but oh so true), watch out for damaged cables, feed clean uninterrupted power, allow every process to finish before you run off to score hot crafty NOM NOM NOM, and watch for heat and static.

Good thing for all of us, the new RED CF station will be out once the FCC or whomever it is that is holding up the release can get the paperwork processed.

We can all safely assume the RED station will provide better heat handling, but details aside, it should provide higher reliability, over all. The dedicated power module should add a reliability, on its own, by making sure the stacked modules get clean juice. And by being able to show the producer: "Lookie, this black box thingie has a RED logo, do you really want to be the one that said we didn't need the black-box thingie...??? I think we should get it."

But bottom line, engines have gotten much more reliable over their history, and the same is happening with computers, only at a far more accelerated rate. And computers are not engines, they are a processors and silicone. Treat them as such and they will treat you with equal tenderness.

 

[Gunleik Groven]

Actually I guess that the whole reason for RED to put out a CF reader, are these issues...

Because these things are weird. I'd have said something like: If you have a problem with one CF reader, try another. I have repeatedly had better results with a no-brand El Cheapo USB CF reader, than a lot of FW 800 reders. But when that fails, the FW 800 usually comes through... The issue has been only with the RED 16GBs. The 8GBs have had no issues.

But the data has ALWAYS been good on the card.

So, my conclusion is simply: Check your takes....

 

[Roberto Lequeux]

So, my conclusion is simply: Check your takes....

Sure, stunts and other expensive or critical takes... but unless you can afford a team of DITs to visually check all takes on the fly you are not going to be doing that.

So always shoot a safety, and ship footage with a list of marked takes to editorial right away. That way you move fast and light, while still getting to look at how things are cutting which you should anyway. Then you will have a good idea not only if scene won't cut but also if there is something that can make it better, before you move company across the State.

But there are several ways of doing things. And what is important is that you make sure you understand that things like food, transpo, and footage handling are not areas to save money... they are areas where you can lose serious money. Same reason why people go to established film labs.

just my humble 0.02

 

[Greg M]

Sure, stunts and other expensive or critical takes... but unless you can afford a team of DITs to visually check all takes on the fly you are not going to be doing that.

We use a single DIT on every shoot and he visually checks every shot on set after backing up to two locations. If you have the right equipment, he doesnt even work OT.

 

[Roberto Lequeux]

Yes, and that is great, and I totally agree that top-notch DITs are a fantastic investment (a must!) which smart producers will take note of. But every project will have different needs and be able to do different things.

Even with enough budget for the best DIT, some days you can't visually check everything. Say you have a couple of days shooting two cameras in deep snow far below freezing point and moving company a couple of times. You might still be able to check it all, but maybe not before you already moved. Or what if you had a third camera filling up a RED RAM while on a heli, or a third getting a couple of shots in an underwater casing under 2' of ice.

All I am saying is that it is not something you can rely on 100% of the time for 100% of the footage, so shooting a safety and having editorial start along with production can be a great place to be.

 

[Zakaree Sandberg]

I was just recently on a job, and they called me a few days later about a problem with creating quicktimes. the export would just freeze up and stop.. anyways turns out there was a corrupt frame..
I used multiple ways of checking the footage. r3dmanager, visual scrubbing, multiple drives for backing up, 2 different computers.
this little guy was hidden in plain sight...
so blaming it on inexperienced data management in my case wouldnt work... been doing it for a long time, and I used the right tools.
Ive obviously had issues before, but I have caught them before it got to post.


btw.. this particular job was on 16 gig cards, 4k 16x9, 24fps
and it wasnt just one clip.. it was 3 clips on different cards.
Nothing HIT on my end as being corrupt or bad.. everything was good as far as my tools and eye could tell

 

[Jay Gannon]

Yep the 1% happens to us all, no matter how thorough you are one clip might get through in every few hundred I would estimate.

 

[Greg M]

I was just recently on a job, and they called me a few days later about a problem with creating quicktimes. the export would just freeze up and stop.. anyways turns out there was a corrupt frame..
I used multiple ways of checking the footage. r3dmanager, visual scrubbing, multiple drives for backing up, 2 different computers.
this little guy was hidden in plain sight...
so blaming it on inexperienced data management in my case wouldnt work... been doing it for a long time, and I used the right tools.
Ive obviously had issues before, but I have caught them before it got to post.


btw.. this particular job was on 16 gig cards, 4k 16x9, 24fps
and it wasnt just one clip.. it was 3 clips on different cards.
Nothing HIT on my end as being corrupt or bad.. everything was good as far as my tools and eye could tell

Did you try running Red Undead on the file? It seems to be a miracle worker.

 

[Gunleik Groven]

Zakaree... Basically you say that you didn't do a visual or transcoding check before format.

 

[Roberto Lequeux]

Zak, remind me never to hire you. :thumbsup: What Zak is saying that the human element is always there. We don't count frames as they speed by before our eyes. All it takes is one of our blinks to line up with the wrong frame, and it all goes out the window. And he is also reminding us that footage checkers are not flawless either, unfortunately.

All of which supports my point, which makes me a genius, and I will be expecting my money in the mail. ;P