Page 1 of 1

WARNING TO URSA OWNERS

PostPosted: Sat Jul 25, 2015 11:29 am
by Mattwright
Hi All.

I just wanted other folks to be aware of a serious problem with the URSA that has really screwed me over twice now.

I have had 2 shoots now, very fast run and gun (which isn't easy with an URSA) where after the shoot I have gone back to check the recordings and they are completely corrupted, It is just a pink recording, it is the correct file size, but the video is completely screwed and pink. On both occasions, these weren't scenarios I could re-set up, as they were just live realtime events.

Having spoken to BM, they said it is a bug which will get fixed, but so far this has not happened. The workout around is, every time you power on the camera, you must record a second of footage, and play it back, if that footage works, then you will be good until you cycle the power again. If the first recording is corrupt, everything you shoot will be corrupt until you cycle the power.

This is a MASSIVE deal breaker, when this is supposed to be a professional camera. Sony, Panasonic and others would not allow this out in the wild in this state.

Just wanted to warn others, so you don't have to tell your clients that their shoot was screwed, like I have had to tell mine.

and BLACKMAGIC, please sort this issue out, with the utmost urgency that this requires. !!

Re: WARNING TO URSA OWNERS

PostPosted: Tue Jul 28, 2015 4:00 am
by Benton Collins
That is a bummer to hear! I hope BM fixes this ASAP! The work around sounds simple enough, BUT, having the operator to remember and do this record checking every time is an accident waiting to happen.

Re: WARNING TO URSA OWNERS

PostPosted: Tue Jul 28, 2015 3:28 pm
by John Fishback
There have been other posts about this kind of file corruption, but most pointed to bad SanDisk CFast2 cards. Did Blackmagic say it was definitely the camera as opposed to the CFast 2 card?

Re: WARNING TO URSA OWNERS

PostPosted: Tue Jul 28, 2015 4:17 pm
by David White
Whether it's the card or camera, it shouldn't be hard to have the camera automatically verify recording integrity during recording. I sometimes see this kind of warning when burning discs when one of the discs is bad. I'm not talking about the separate verify pass. I get warnings during the writing phase.

David

Re: WARNING TO URSA OWNERS

PostPosted: Tue Jul 28, 2015 10:11 pm
by shanepeters@bellsouth.net
Good thing you know how to verify it yourself now.....just don't forget to check each time.

Re: WARNING TO URSA OWNERS

PostPosted: Tue Jul 28, 2015 11:47 pm
by JamesParker
Weird, we film with the URSA on every shoot. Haven't experienced it yet.

Using Lexar 256

Re: WARNING TO URSA OWNERS

PostPosted: Wed Jul 29, 2015 7:26 am
by Tom
David White wrote:Whether it's the card or camera, it shouldn't be hard to have the camera automatically verify recording integrity during recording. I sometimes see this kind of warning when burning discs when one of the discs is bad. I'm not talking about the separate verify pass. I get warnings during the writing phase.

David



Have you ever had to create a data integrity process for video content before? Or are you just guessing that it shouldnt be hard.

The data rate of a disc being burned is a tiny fraction of the data rates being written on these cameras, not to mention that a disc burner can use a buffer.

You are comparing apples to oranges.

Depending on where the corruption is happening will also substantially effect how easy it would be to check or even have a non-intelligent system identify what is corruption and what is normal signal. Let say it happened during the write stage to the card - it would have to read every packet of data each time one is written - doubling the data rate going from card to camera. Lets say it happened from the sensor to the DSP - how would corruption be identified?

Before pointing fingers and implying that this issue could have been avoided if the designers had implemented something apparently easy - let's first try to diagnose where the fault is.

Re: WARNING TO URSA OWNERS

PostPosted: Wed Jul 29, 2015 7:29 am
by Jamie Hammond
Hi Guys,
I raised this issue in a previous thread but on to mostly deaf ears:
viewtopic.php?f=2&t=36805

Watch out for the dual post haters....

This isn't a card issue this is a camera issue amongst a bevy of others....

Best

Jamie

Re: WARNING TO URSA OWNERS

PostPosted: Wed Jul 29, 2015 8:09 am
by Tom
Jamie Hammond wrote:This isn't a card issue this is a camera issue amongst a bevy of others....




I don't recall this ever being established, nor can you be sure that this users problems are also caused by the same thing as your own.


Perhaps you could explain to the OP as to what you have done to prevent it happening since (assuming this is the case) and they can test it out too?

Re: WARNING TO URSA OWNERS

PostPosted: Wed Jul 29, 2015 3:01 pm
by Guy Fiorita
Hey Matt -

What cards were you using?

Which firmware?

Recording format?

Resolution and Frame rate?

Re: WARNING TO URSA OWNERS

PostPosted: Wed Jul 29, 2015 3:13 pm
by Nigel Daniels
This is a known issue. BM is supposedly working on a fix.

Re: WARNING TO URSA OWNERS

PostPosted: Wed Jul 29, 2015 4:44 pm
by David White
Tom wrote:Have you ever had to create a data integrity process for video content before? Or are you just guessing that it shouldnt be hard.


Yes I have. I am a professional software engineer. Since the data corruption covers a wide area instead of just a few pixels, reading a few bytes should be enough for a confidence check. For Prores recording, the file can be opened for both read and write access.

Tom wrote:Before pointing fingers and implying that this issue could have been avoided if the designers had implemented something apparently easy - let's first try to diagnose where the fault is.


I didn't intend to 'point fingers'. I'm an engineer and wanted to offer a suggestion. If there is a problem with the camera's firmware, their engineers will find the cause and fix it. There have also been problems with cards and a low impact confidence check might avoid spoiled recordings. I was just trying to be helpful.

David

Re: WARNING TO URSA OWNERS

PostPosted: Wed Jul 29, 2015 8:20 pm
by Ian Cresswell
David White wrote:
Tom wrote:Have you ever had to create a data integrity process for video content before? Or are you just guessing that it shouldnt be hard.


Yes I have. I am a professional software engineer. Since the data corruption covers a wide area instead of just a few pixels, reading a few bytes should be enough for a confidence check. For Prores recording, the file can be opened for both read and write access.

Tom wrote:Before pointing fingers and implying that this issue could have been avoided if the designers had implemented something apparently easy - let's first try to diagnose where the fault is.


I didn't intend to 'point fingers'. I'm an engineer and wanted to offer a suggestion. If there is a problem with the camera's firmware, their engineers will find the cause and fix it. There have also been problems with cards and a low impact confidence check might avoid spoiled recordings. I was just trying to be helpful.

David



This post makes me smile.

Re: WARNING TO URSA OWNERS

PostPosted: Wed Jul 29, 2015 8:47 pm
by GaryHansen
I had the same problem but caught it early as we were doing interviews for the day. I turned off and the powered on the camera recorded a few seconds and then played it back... seemed to be fine. During the interview process I was checking each shot while the producer was asking the question... The BlackMagic URSA is fast and is very nice for this function. We use the URSA as a tool and we never trust any tool that is created with software and little to no moving parts. We live in a new time, no more shooting on 16mm, chain it to 3/4 and edit to BetaCam for the evening news. (ok that showed my age a little) So with these minor problems here and there no big deal just continuously check your progress and hope the Engineers at Blackmagic continue making our life easier.

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 7:20 am
by Tom
David White wrote:Yes I have. I am a professional software engineer.


Camera software? as in firmware?

David White wrote: Since the data corruption covers a wide area instead of just a few pixels, reading a few bytes should be enough for a confidence check. For Prores recording, the file can be opened for both read and write access.


This assumes the corruption occurs in a specific part of the process. How would a "dumb" system identify what is corruption and what is not? It has no way of comparing it such as a disc writer can, as it has the original data in digital form and the written data. Such a system might work if the corruption occurs towards the back end of the chain - but it would fail to identify it towards the front end - so it wouldn't be a reliable system.


David White wrote:
I didn't intend to 'point fingers'. I'm an engineer and wanted to offer a suggestion. If there is a problem with the camera's firmware, their engineers will find the cause and fix it. There have also been problems with cards and a low impact confidence check might avoid spoiled recordings. I was just trying to be helpful.

David


I appreciate that, but it implies that the reason such a system doesnt exist already is just a result of either lazyness or lack of imagination - which is a bit insulting.

I acknowledge your positive intentions but when people propose ideas to remedy a situation on here, they seldom consider the actual complexity of what they are suggesting. It's similar to when people claim that X is possible because they did it with magic lantern - so it must be easy.

Sorry I snapped a bit, but often someone proposes something, something completely unfeasible (not necessarily suggesting that this is true here) - and no one challenges it, and so it gets repeated and quoted and thrown about as if its a guaranteed solution and why arent blackmagic doing it.



Also, I want to say how its good to see a software engineer guy on here, I look forward to more of your contributions (as long as I don't find them so objectionable ;-) )

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 9:28 am
by Aaron Scheiner
Everyone's an engineer, especially those that write HTML.

I always try and use backup SDI recorders and where possible we regularly dump footage (with md5sums) to an external storage array for verification and safe keeping.

I've learnt that buying and/or pre-ordering a new BMD product is a really bad idea. Wait a few months and you'll be able to pick the same item up second hand, for a great price and it'll have all sorts of bug fixes and features it wasn't originally released with.

Lastly, I always fill new storage media with random data (under Linux). I do a minimum of three full writes before using new storage devices in the field. This process forces the device to detect (and re-allocate) unreliable sectors and in general places the storage device under considerable stress. A couple of devices have failed on me during this process.

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 9:28 am
by Kristian Lam
Hi,

We are aware of this issue and have a fix for it that is being tested. Some customers who have contacted our support offices have been provided with a pre-release software build that addresses this issue.

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 2:32 pm
by Rakesh Malik
Tom wrote:
David White wrote: Since the data corruption covers a wide area instead of just a few pixels, reading a few bytes should be enough for a confidence check. For Prores recording, the file can be opened for both read and write access.


This assumes the corruption occurs in a specific part of the process. How would a "dumb" system identify what is corruption and what is not? It has no way of comparing it such as a disc writer can, as it has the original data in digital form and the written data. Such a system might work if the corruption occurs towards the back end of the chain - but it would fail to identify it towards the front end - so it wouldn't be a reliable system.


That amused me also, since it was some rather backward logic. If the corruption covers a wide area, reading a few bytes is clearly not sufficient for a confidence check, because statistically you'd be reading good pixels most of the time anyway.

It also fails another major criterion: how do you know that the data you're reading are good pixels rather than bad? On disk, they're juts bits. You can't know whether they're the correct bits simply by reading them.

The solution is to either require enterprise-grade media like Red and AJA do, or use some variation of ECC, both of which are expensive. To be reliable with consumer media requires a proper validation step, not just a hokey "read a few pixels" check.

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 3:59 pm
by rick.lang
Kristian Lam wrote:Hi,

We are aware of this issue and have a fix for it that is being tested. Some customers who have contacted our support offices have been provided with a pre-release software build that addresses this issue.


Kristian, thanks for the update, good news.


Sent from my iPad using Tapatalk

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 5:52 pm
by Mattwright
Sorry for the slow reply on this.

Although most of the questions have been answered with Kristian's post.

This bug is known to BM, and certainly I have had this issue with firmware 2.3 and previous versions. I may be wrong with this information but I am sure Kristian will correct me if I am, But I don't believe the bug is to do with the writing of the data to the CF Cards at all, as in, I think it can happen on any cards, and frame rate. I believe the problem is to do with the sensor passing bad information to the cards to be written, thus why I could not try and fix the files I ended up with. Nor could BM.

Hopefully this new firmware will allow us to shoot without fear.

It does't happen very often, we have shot maybe 40 hours of material over 10 shoots, and it has only happened on 2 of the shoots. But even so, in my case, they were critical shots.

Fingers crossed that this resolves this issue.

Matt

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 6:44 pm
by Rakesh Malik
Mattwright wrote:It does't happen very often, we have shot maybe 40 hours of material over 10 shoots, and it has only happened on 2 of the shoots. But even so, in my case, they were critical shots.

Fingers crossed that this resolves this issue.

Matt


I'm with you, because I'm awaiting an Ursa Mini, and because I want to see BMD gain a reputation for dependability, so that they become more of a selling point. *I* know that the camera I use is less significant than my own skills as a cinematographer, but the bottom line is that I want to get hired... and most people who would hire me don't give a blert about the cinematographer, they care only about the camera that the cinematographer is bringing to the table... so BMD getting a better reputation is a plus for my marketing, in addition to being a camera I am looking forward to using for aesthetic reasons.

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 6:55 pm
by rick.lang
Perception is everything. A knowledgable acquaintance simply dismissed my interest in BMD cameras because he had heard they had a black sun. He knew just enough to know he didn't want to go there. And he may never learn that it isn't an issue on the URSA Mini but I did send him the webpage! We all know the saying, "You only get one chance to make a first impression." That's why I am hoping the URSA Mini 4.6K hits a home run in its first at bat. It may be delayed, but eventually worth the wait.

BMD has sometimes made a poor first impression to those who don't want to look any further. For those who do take a second look, many simply fall in love!


Sent from my iPad using Tapatalk

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 7:01 pm
by Guy Fiorita
Hey Matt -

What card were you using out of curiosity? The other two times I've seen posts about this on the forum its been the 64 GB lexar.

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 7:05 pm
by Rakesh Malik
rick.lang wrote:BMD has sometimes made a poor first impression to those who don't want to look any further. For those who do take a second look, many simply fall in love!


They have indeed. Hopefully, they've learned from that, and will do a much better job with the Mini. I don't mind if they have some functionality that's coming later (like raw recording on the Pocket) as long as they're dependable out the gate.

Hopefully, in addition to CFast2 media becoming less expensive, it will also be getting more reliable...

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 7:27 pm
by Andrew Deme
The pink frame looks like the sensor is out of sync and the bars look like memory...either way clocking.

Just a stab in the dark but shouldn't be hard to sort.

Re: WARNING TO URSA OWNERS

PostPosted: Thu Jul 30, 2015 9:19 pm
by Mattwright
Hi Guy

On both occasions, we were using the Lexar 256GB cards, but as mentioned I believe it is a sensor error rather than a card error, the data on the cards is fine, its just what has been sent to the cards is scrambled.

I have to say in BlackMagic's credit, they have contacted me since my initial post and have now told me that new firmware which should resolve this issue is very close to hitting public release.

Matt

Re: WARNING TO URSA OWNERS

PostPosted: Fri Aug 07, 2015 9:54 pm
by Kevin DeOliveira
Hey all,

I'm still following this issue to see how it pans out.

Keep us posted as to whether or not the new firmware helps.

Re: WARNING TO URSA OWNERS

PostPosted: Sat Aug 08, 2015 12:08 am
by Norbert Bielan
Every camera will be different.

I have never had this problem after shooting more than 200 hours on the URSA, but I don't doubt the OP's issues for a second.

Re: WARNING TO URSA OWNERS

PostPosted: Sat Aug 15, 2015 9:01 pm
by James Parker
Just experienced this issue for the first time on a San disk card. Never have experienced this with our lexar. Ours might be the card. I can no longer format the card in any way. Using the camera, or on our computer. Calling Sandisk to verify the issue


Sent from my iPhone using Tapatalk