Flash Incompatibility

[Finner]

I do not believe this has been touched on yet.

I just wanted to confirm that RED was making a flash port that did not just accept "Red" brand flash memory cards. I have always figured that red would accept SanDisk and other memory company's products.

Panasonic made this mistake with the P2 cards and the camera has not taken off as well as it could of if it just took standard flash memory.

Can anyone tell me if the RED is planned to accept flash cards from outside memory companys?

[Pol Turrents]

Since I know, Panasonic has offered the technology to other manufacturers of SD cards. There are plans from other companies to sell P2 cards, but I think it's a slow process, since it's still a small market.

[Blaine Golden]

it's a slow process, since it's still a small market.And that's what most people forget. We're so involved in the this particular market that we forget how small it is in the grand scheme of video recorders.

[Don Woods]

We are a very small market when you think how big the consumer electronic market is. I beleav the plan was to utilze memorey that was already on the market. But I could be wrong. Either way it is still an interesting option for recording media

[Jeff Kilgroe]

We are indeed a very small market. All the more reason why RED should make their FLASH module use standard media. If it can take 4 SD or CF style media cards and interleave them (think RAID 0 stripe), it should work rather well. And we should see 32GB and larger CF style media becoming common by the end of this year. I originally wrote a huge post here about my detailed point of view on Panasonic's P2 and why it is evil, but I moved it to the off-topic section (http://www.reduser.net/forum/showthread.php?t=366).

[Finner]

The film camera companys have things right in this area. They concentrate on making great cameras and improving them and leave the media capture to kodak and fuji.

If Red leaves the media capture to the computer memory companys things will be so much better. We can just go to any local computer store and be able to buy the newest largest fastest best priced flash cards when ever we want or even to have quick ability to replace a card if one goes bad. This way RED can just concentrate on the next new RED improvement and stay ahead of everybody else.

[Greg Voevodsky]

At a seminar, A Panasonic spokesperson said that their tolerances for their flash chips were like 1% or some horrific number. So that they had to choose the top 1% of Flash chips that had little to no errors in order to make their 'perfect' P2 flash cards and that is why they cost so much.

I still wonder to this day, does that mean average flash cards had a lot of drop out? Considering 4k and drop out - lost pixels would be a big deal - unless there is error correcting, etc to compensate for that. Anyway, that is what I was told. If the RED crew could address us, that would be great.

[Finner]

At a seminar, A Panasonic spokesperson said that their tolerances for their flash chips were like 1% or some horrific number. So that they had to choose the top 1% of Flash chips that had little to no errors in order to make their 'perfect' P2 flash cards and that is why they cost so much.

I still wonder to this day, does that mean average flash cards had a lot of drop out? Considering 4k and drop out - lost pixels would be a big deal - unless there is error correcting, etc to compensate for that. Anyway, that is what I was told. If the RED crew could address us, that would be great.

This sounds like panasonic B.S. to me. They use a Fire hard drive right. Is that then as 1% top hard drive system? Panasonics flash excuse sounds weak to me. If there is someone that is extremely flash savy it would be good to get your input. I just don't personally buy the super flash argument, we all know flash is faster and more durable then hard drives so it just doesn't make sense that if you needed the top 1% flash chips how on earth could you build a hard drive that works.

[Jannard]

While errors are a concern... speed is the biggest. The guaranteed minimum write speed has to be greater than 30+MB per second (not just advertised). Speed and reliability are mandatory. If we open to a standard format, say Compact Flash, and someone buys an 8GB card that doesn't work because the write speed is too slow... is it our fault? This is a difficult decision for us. We want an open system. But we also don't want unhappy customers. Two years from now there will be lots of great options. But right now, the options are few.

The good news is that we have the capability to change our flash drive in the future if something great, and open, comes along. Modularity is a good thing.

Jim

[Finner]

I can totally see your point Jim.

High performance cars have the "premium unleaded only" on the gas cap to warn customers. I feel the users of RED will be very tech savy and I think a requirement disclaimer beside the flash port would be enough of a warning.

The ability for RED owners to use high quality 30+ flash straight off the local store shelves would be a huge advantage. i know this may cut you guys out of a few profits but I really do feel in the end the ability for you guys to focus strictly on the camera and leave the flash to memory companys would benifit red in the long run.

[Jarred Land]

I can totally see your point Jim.

High performance cars have the "premium unleaded only" on the gas cap to warn customers. I feel the users of RED will be very tech savy and I think a requirement disclaimer beside the flash port would be enough of a warning.

there is the problem.. i know many, many people with luxury cars that ask for premium gas but the owner puts mid or low grade octane in... Its like cheap DV tape.. nobody every says they use it on there $6000 prosumer cameras, but many people do at one point in desperate times.

[Finner]

Sure Jarred.

You can also put gasoline in a diesel car which will really screw it up, but at some point I feel the responsibility needs to be on the user. I feel given the choice an extremely overwhelming majority of RED consumers would pick the choice of a standard flash card option.

[Blaine Golden]

Sure Jarred.

You can also put gasoline in a diesel car which will really screw it up, but at some point I feel the responsibility needs to be on the user. I feel given the choice an extremely overwhelming majority of RED consumers would pick the choice of a standard flash card option.
That's great as long as it works, but what happens if you get one that doesn't work with the RED?

[Appleton]

All depends on the price, for me. i'm willing to pay a premium (within reason) to get a decent amount of storage and guaranteed reliability .

[Finner]

Hi Blaine

Thats why I mentioned the warning or disclaimer on the flash port. As long as you buy a flash card that is the proper speed rating according to what Jim says it will work fine. Now you could get a dud card from SanDisk or another memory manufacturer but there would be just as big of a chance getting a dud from RED in that case.

Pretty much if you buy a flash card that does not have the speed rating that RED states is needed it is your own fault for not following directions. It's not much different then buying a sony HD900 tape and being upset that it will not work in your panasonic camera.

Lets face it if RED buys Sandisk flash memory and then packages it into a RED flash card it will be inevitably more expensive. I don't think that is something anyone wants.

Personally I think all the owners would be very careful on their flash card choices for their camera. I know mine will be babied. Also if your in a position where you desperately need a flash card ASAP do you really want to have to order one from RED and have to wait out the delivery time. For RED owners out of the country this could be a big hastle and expense.

[Evin Grant]

I somehow doubt that the Red Ram/Flash will be as simple as a single CF card. The stated capacity of 32GB seems to rule that out immediately. I think it's reasonable for Red to use a proprietary flash system in this version 1.0 system knowing that in a year or two memory like CF or Flash Drives will be both fast enough and have the capacity to be real options.

P.S. Sandisk currently has a card, the Extreme IV that can write at a peak of 40MB per second. As well as a 16GB card that is slower but we can obviously see we are not far off from this being a viable solution. However some patience is necessary.

[Finner]

I hope it does not come across like I am trying to stir the pot. I am just addressing this issue as I feel there are a lot of pro's and cons to each side. I am not real flash tech experienced so I am just trying to bring out all the points on this. I just like everyone else am just looking forward to getting the best camera possible.

[Dominique Grenier]

Maybe there's something I don't understand, but if I recall correctly, an 8GB P2 card is really a RAID of four 2GB SD card in a small enclosure with a RAID controller, isn't?

So, if I'm right, that's why you can't use your own flash memory card, right?

Or are you speaking about being able to replace the flash memory inside this enclosure with your own memory?

[Blair S. Paulsen]

I think we all have been the victims of pricing policies on critical items in this business for so long that our radar starts beeping as soon as the word "proprietary" gets used. P2 is certainly a cautionary tale fresh in many minds. At this point I am willing to trust that the Red Leader will be very judicious in marking up the flash media and accept the trade off of more reliability. Jim has already stated that in a year or two a more open standard solution may be available.

Bottom line - in the Red Leader I trust until I have reason not to.

[Greg Voevodsky]

I see 2 solutions: 1. Like Avid, certify the Flash Card with an online REDLIST of approved Flash Cards. and 2. RED Flash - say 20% more but guaranteed money back and tested... I'm sure lots of us would pay a premium for that and RED makes some $$$.... and for the adventurous people, they can try a RED Certified flash card, and if it works, use it. Otherwise, they can sell it on e-bay to all the still photographers.

[MikeCurtis]

If they were to certify....

1.) They'd have to test

2.) They'd have to spend the time and money to test, and then, some other manufacturer gets the benefit of that testing

3.) The speed and reliability issues make it clear to me that we need a Red branded solution - it isn't going to be a stock part you can pick up at Fry's.

Thus I find this thread moot.

[Blair S. Paulsen]

Mike makes good points. Here on the bleeding edge of technology I would rather have a fully Red solution, at least to start out.

[Finner]

Mike I assume you do not find yourself traveling around the world to shoot because if you did I think you would look at this quite differently.

As far as testing goes memory cards hold data if they meet the speed requirement and they are from a reputable company they have already been highly tested. Also who do you think will produce the memory for the RED. I would guess it will be a reputable flash memory company that is making flash memory already with there own products.

If film or hd tapes get lost or destroyed somewhere durring travel to a location you can still buy these products in most places and offices worldwide. If RED uses proprietary memory and you can only order it from RED people will end up stuck on location at some point with no way to shoot and there will be nothing "moot" about that.

[Evin Grant]

Finner, just don't get your Red with the Flash option then. Get the uncompressed out when it ships and hold your water until CF cards are voluminous and fast enough for Red to be sure of thier reliability. Because that seems to be what your advocating anyway.

[Mike Smith]

If they were to certify....

1.) They'd have to test

2.) They'd have to spend the time and money to test, and then, some other manufacturer gets the benefit of that testing

3.) The speed and reliability issues make it clear to me that we need a Red branded solution - it isn't going to be a stock part you can pick up at Fry's.

Thus I find this thread moot.

No offense meant Mike but I don't get what you're saying here. Are you suggesting that Red would sell a proprietary flash drive that has not been tested to meet the full specs of the Red camera? Such tests would HAVE to take into account the minimum data rate of the camera otherwise the device would be useless. They cannot rely on the specs of the memory chip manufacturer. If Red go the proprietary route, they will have to test the resulting flash drive to ensure that the manufacturing process did not damage the internals. You can't just solder together a bunch of memory chips and merely hope that they weren't damaged anywhere along the line of coming in the door and going out as an entire flash drive.

[Finner]

Evin from what you say it may be my fault for not understanding properly.

From what I understand is if RED builds a flash port that is RED specific you will not be able to use standard flash cards. So even when flash cards improve I would be highly suprisied if the had the RED specific port as the port they will have will be what is the standard. Thus you would have to change your RED port at that point but even then there would be old flash cards and inferior ones that would not work right? So just like now you would still have to be selective on your flash card choices.

I am just trying to figure this out. I know some of my clients would appreciate if I could use brand new flash cards on their project and just hand them over after shooting with them just like film stock. Thus the ability to buy reasonably priced sandisk memory off the shelf would be great.

[Jared VanLeuven]

No offense meant Mike but I don't get what you're saying here. Are you suggesting that Red would sell a proprietary flash drive that has not been tested to meet the full specs of the Red camera? Such tests would HAVE to take into account the minimum data rate of the camera otherwise the device would be useless. They cannot rely on the specs of the memory chip manufacturer. If Red go the proprietary route, they will have to test the resulting flash drive to ensure that the manufacturing process did not damage the internals. You can't just solder together a bunch of memory chips and merely hope that they weren't damaged anywhere along the line of coming in the door and going out as an entire flash drive.

I could be wrong, but I gathered that Mike was referring to 3rd party memory certification.

[Mike the beginner]

Here's my take on this:o

Red does indeed require top notch quality control on its product. This ensures the flash media works as expected.

Red ensures that the camera's do NOT require to be shipped back to red if and when a new open stndard fash port is introduced. We simply send our payment to red and they send us a new replcement flash media port or whatever. It is made simple enough for the majority of us to fit it ourselves.

Michael

[Finner]

Its good in theory mike but unfortunately new ports are not introduced very often. Flash memory improves but the ports stay the same for a long time.

Mike if you were given the minimum speed flash card requirements needed for your RED would you go out and buy one that does not meet the standards?

I would think every owner would be careful to buy the right flash cards.

[thomashopman]

Would it be a good idea to have some kind of cartridge where you could install 2 or 4 CF cards? The cardrige would have some interleave or raid function so everybody can use the cf cards with the specific speed and volume they need for the upcoming shoot. If you are going to use your red for HD or 2K mainly you could use slower/smaller CF cards. When needed you add or replace some CF cards and reformat and you will be good to go. When bigger and faster CF cards come to market every couple of months/years you can upgrade.

[Evin Grant]

Finner, I just take jim at his word when he says that an open standard flash device (Most likely CF) will be adopted when capacity and speed are at their requisite levels for reliability.

[PaulClements]

Well Red isn't going to be making their branded Flash memory in house, just as they aren't making their sensors in house, so it can be assumed that wherever they get it from already has a product that is going to be released or already has been released. As such it stands to reason that there will be other products available on the market.

I don't stand by the reasoning that if someone buys the wrong card is Red to blame? Absolutely not, anyone that spends $17,500 on a piece of electrical equipment will be more than capable of discovering the best memory to be used with it.

Whats more the entire Red revolution has been documented on these very forums and just in the same way Evin and others are collecting information on Lenses there is no reason why we cannot do the same for the individual cards on the market as and when they become available. As long as Red supplies a standard Flash drive, whether CF or other, which can write to that type of flash memory faster than the required 30Mbps or whatever it ends up being then the Onboard flash ought to be future proof enough that stripping out the entire drive and replacing it ought be entirely unnecessary.

Making aspects such as flash memory Red Branded is a very backward move in my opinion and is similar to the tactics employed by companies such as Apple and Sony that ends up being overpriced, time-consuming for replacements and annoying in the long run.

[Matt Uhry]

Would it be a good idea to have some kind of cartridge where you could install 2 or 4 CF cards?

YES! That would be a great solution, 4 CF cards. if you were DIY enough you could pull out a set and replace it with a matched group of faster cards. Kind of like recelling a battery. I'm a bit cranky on this topic ( like some others I think ) due to the way Panasonic bungled the P2 card size/pricing. I don't want to see Red go down that path.

Matt Uhry
www.fuzby.com

[Andrew M.]

I don't want RED to go that path ether.
P2 screening explanation from Panasonic is on big BS.
Like picture can not accept one bad memory cell!
Financial programs run off these flash cards, people store documents, one Byte error can freeze your program and what Panasonic can't accept two errors in the picture, wow!! In digital world things work or don't work.

You just tell people what works with RED-FLASH-INTERFACE mounted on the camera (mention 3 brands) and anything else is user own experiment and responsibility.
We have this user-net to exchange info on what works and what doesn't
Like there is no 30 days money back guaranty on the flash-cards:-)

Andrew

[Finner]

Finner, I just take jim at his word when he says that an open standard flash device (Most likely CF) will be adopted when capacity and speed are at their requisite levels for reliability.


Hi Evan I trust Jim a ton as well. In fact people I know that are owners of some of the largest rental companies in LA have met with Jim and talk very highly of him.

It has nothing to do with trust for me. I just want to express that if it is possible to build the RED to accept flash memory from outside companys it would be a big plus for me. At some point responsibility has to be put on us users to choose wisely.

In the end I will probably buy most of my flash memory directly from RED as I will feel more comfortable using it but in a bind the ability to pick up flash memory from a local store would be very helpful.

I have heard that most DVX and HVX shooters will only put panasonic tapes in there cameras but in an emergency when they can't find a panasonic tape they have options. Iwould just like those kind of options.

Truely though in the end if the RED team feel they can not put a standardized flash port on the camera I will be fine with that. I just wanted to express my preference, the reasons behind them and see if others felt the same way.

[Craig Schober]

In the end I will probably buy most of my flash memory directly from RED as I will feel more comfortable using it but in a bind the ability to pick up flash memory from a local store would be very helpful.

flash memory is reliable because it is sold state but speed consistency is always an issue. marketing hype and labels cannot be trusted when it comes to raw mb/sec throughput. some cards write/read fast and some do not but all claim to have great sustained rates. who is supposed to test all these products and combinations? normally i would assign the task to red but i prefer a reasonably priced proprietary solution that they can guarantee. it's not like cheap dv tapes that are mostly interchangeable.

[PaulClements]

As long as that proprietary system doesn't alienate the possibility of using third party products I'd agree. People will post here on these forums if a cheaper, popular and more readily available product works fine with the Red One. Sticking one in the camera to see if it works isn't going to break the thing and the ability to buy flash locally rather than ordering it from the other side of the world is a massive plus point.

Whilst labels and other text written on card packaging may be misleading (Although in my experience this has never been the case) there are a host of good review sites for such products, the main dealers of them on the web give very thorough specifications and anyone with half an ounce ought to be able to see if it is of an exacting standard to work with Red One.

By all means have a Red Branded version but don't make it incompatible with other cards, afterall even if the other card that was being used was slower it could still possibly be used to shoot 1080p or 720p instead of 4k and probably work fine. The lessened speed would probably mean that shooting at lower resolutions was cheaper and more efficient, ideal for eng efp etc - lightweight and low cost.

[Chris Kenny]

It's likely that a REDFLASH mag is going to use multiple memory chips in parallel to improve performance, so allowing commodity third-party memory cards would require making it possible to open the mag itself and install your own cards.

Frankly I'm not sure this would be very useful. For one thing, it would raise costs, as it would require memory in the mag to be on memory cards; there's no reason not to just use bare chips in this application. More relevant to this discussion... RED would still effectively control pricing, because you'd still have to buy the mag from RED at RED's price.

I think at this point, we should probably just trust that RED isn't going to screw us on pricing. They haven't yet, have they?

It would be nice if they documented how the camera communicates with flash mags, so third-parties could also make them, but if RED's pricing is as aggressive as I suspect it will be, I wonder whether anyone else would bother getting into that market.

[RobRoySyd]

I think flash memory is a lot like a hard disk. It's not 100% flawless, parts of it maybe defective or become defective and the unit as whole is still OK. Those defective areas are simply spared out. So long as the amount spared out doesn't become excessive all is well.

However the process of redirecting read/write cycles obviously means a speed hit. Not normally a problem, maybe once in while it takes a little longer for you still camera to cycle or a file to save or retrieve.

Where it might become a problem is writing 8GB of data that needs a sustained data rate e.g. video. In all fairness to Panny I think this might be where they're coming from with their P2 card story, probably they've dumbed the story down too much so the marketing types can remember it.

I'm really guessing now, perhaps for applications such as RED the flash memory will need to be qualified as meeting sustained data rates rather than average data rates. Alternatively perhpas a large RAM buffer would obviate the problem of the odd glitch in the sustained data rate, all that needed to be met was an average data rate. Maybe the system should / could have a status warning that the flash tolerance was getting too low. Decades ago I worked on memory systems with ECC that did this sort of thing. Start getting too many corrected errors in a time window and the OS raised an alarm.

[Craig Schober]

As long as that proprietary system doesn't alienate the possibility of using third party products I'd agree. People will post here on these forums if a cheaper, popular and more readily available product works fine with the Red One. Sticking one in the camera to see if it works isn't going to break the thing and the ability to buy flash locally rather than ordering it from the other side of the world is a massive plus point.

i have an hvx200 and refuse to buy a p2 card (i use firestore which works fine for me) until pricing is more reasonable but i also understand that we're not dealing with dv/dvcpro data rates here. it's the old mac vs/ pc argument. mac user experience will always prevail so long as they maintain control over the hardware and software developers. not the best deal but the best user experience. and when you're dealing with a substantial investment like redone (unlike a flash based camcorder), i'm willing to take the hit and pay a little more for the assurance. i'm perfectly fine with sticking in a few cards to test them on red but that won't tell me anything because those cards could fail or drop frames a few hours later when i'm in the middle of a take on location. then what? i can't complain to red. all i can do is scratch that card maker off my list and try again. so the few hundred i saved i just lost on one bad take and i still have no assurances for the next card experiment.

[Finner]

All the concerns about getting a bad card or having problems makes no sense to me. Just because you have the option to buy flash memory from a local store does not mean you have to. This is about choice. Anyone who would not trust a flash card other then REDs thats fine don't trust it use only RED flash cards, no one is telling you not to.

I feel the main thing that is being discussed here is the option to use flash memory other then RED. I personally doubt that anyone with a RED camera was put in the uncomfortable situation of run out and buy non RED flash cards or don't shoot would choose to buy the non RED cards and use them until they could get the RED ones.

If the option to use outside companys memory is not there then RED owners better be prepared to purchase at bare minimum 3 forms of memory retreval. Either 3 red flash cards or 1 red drive and 2 flash cards or 3 red drives. Production shooting costs (sets, crew, locations, set decoration, other departments rental costs) are all very expensive and I have learned thet you should always have a backup for your backup. So if RED tells me that it is not possible to let the camera use outside memory fine. But as far as people being so scared to use outside memory that they feel they have to tell others what they are allowed to use well who are you to tell anyone.

[Andrew M.]

all i can do is scratch that card maker off my list and try again. so the few hundred i saved i just lost on one bad take and i still have no assurances for the next card experiment.

Few hundred you saved you didn't loose, you post your experience on dvx forum or here and hundreds of other users will know what flash memory works and which doesn't, so if you lost on this one you will not loose on some other thing that you will read in these forums all about:-)

Andrew

[Jeff Kilgroe]

I think all the paranoia over common FLASH memory is totally unjustified. Sure, there are variations in quality and speeds... Even with the same manufacturer on typically good products. Hard drives have the same issues too, but we all buy those off the shelf -- or at least most of us do. I've bought tons of defective components over the years, many of which were hard drives or other media, even FLASH media. It happens... Test out the media a time or two before using it on a critical production. Can't tell me that even if RED produces a proprietary and certified FLASH media format that there won't be defective ones every now and then. I've got a defective P2 card here that Panasonic let me keep, so I took it apart to study it. ..It happens.

I truly see no reason why the RED FLASH module couldn't house 4xCF cards in a striped RAID configuration and get the proper throughput. Each card could be formatted or initialized via the camera and given a sequence number. On later use when cards are re-inserted later for additional reading/writing, the card sequence numbers could be referenced and it wouldn't even matter which CF slots on the module they were plugged into.

To me this seems like a no-brainer solution that can scale and grow with FLASH technology and with the users' needs. 4x8GB CF cards would be easy to buy just about anywhere right now and we should see 32GB CF cards on the common market by the end of this year. RED could sell their own label of certified CF media too... Why not?

[Andrew M.]

CF card is good reliable way to go.
I rather see 8 slots instead of 4 slots for CF cards there.

If we are lucky enough, by the time we see camera production, more of these will be available.
http://www.theinquirer.net/default.aspx?article=36841

eSATA interface would be more in line with the interfaces that camera will have.
But as I see it coming, the RED FLASH MEMORY INTERFACE for the camera will be plugging in to the RED-DRIVE SATA connector, so anything goes in terms of the interface enclosure for flash memory, as long as it will interface with the SATA connector on the camera
Remember SATA connector on the camera body is modified to supply the power since the original SATA spec do not have provision to deliver power to the SATA device.

Andrew

[Craig Schober]

While errors are a concern... speed is the biggest. The guaranteed minimum write speed has to be greater than 30+MB per second (not just advertised). Speed and reliability are mandatory. If we open to a standard format, say Compact Flash, and someone buys an 8GB card that doesn't work because the write speed is too slow... is it our fault? This is a difficult decision for us. We want an open system. But we also don't want unhappy customers. Two years from now there will be lots of great options. But right now, the options are few.

The good news is that we have the capability to change our flash drive in the future if something great, and open, comes along. Modularity is a good thing.

Jim

no one is paranoid about failing cards. flash memory is very reliable. it's about price/performance. there's a reason why p2 is so expensive and why jim hasn't committed to an open redflash standard publicly. it seems that off-the-shelf (usually cheapest) flash memory cannot currently guarantee required speeds. and we're all tempted to pick up the cheapest flash we see so you're just asking for problems which will become red problems. it's like running non-supported drives on your non-linear system, they might have great specs and you might save thousands and you might never have a problem but you can only blame yourself when your editing system fails.

[Finner]

no one is paranoid about failing cards. flash memory is very reliable. it's about price/performance. there's a reason why p2 is so expensive and why jim hasn't committed to an open redflash standard publicly. it seems that off-the-shelf (usually cheapest) flash memory cannot currently guarantee required speeds. and we're all tempted to pick up the cheapest flash we see so you're just asking for problems which will become red problems. it's like running non-supported drives on your non-linear system, they might have great specs and you might save thousands and you might never have a problem but you can only blame yourself when your editing system fails.


I'm not tempted to pick up the cheapest flash memory. In fact I have stated that I would prefer to use the RED offered flash. I dont want it to be red specific in case of an emergency situation. Having Red overnight you flash memory cards does not help when its 9AM and you could just send a PA to a computer store to grab you some flash cards. This is not about what is the best choice. To me that is clearlry the RED flash card. This is about the option and the choice to use other Flash cards in an emergency situation.

[Jeff Kilgroe]

When I was talking about paranoia, it was about sustaining proper performance throughout the entire FLASH structure. Not just reliability. And I was saying that hard drives have this very same issue. Saying that off the shelf flash isn't a usable solution because it can't sustain the advertised the advertised rates is a bunch of nonsense... I don't know of any recording media that can. What matters is that the FLASH memory can at least sustain the rates required by RED, and I say that it can in the proper configuration.


there's a reason why p2 is so expensive

Really? Is there a truly good reason it's so bloody expensive? I know what it would cost to manufacture P2 cards... We almost started doing it here, but at the time it wasn't a good fit for us since it was early in the game. Now we're not really interested due to the direction the whole P2 format has taken (or should I say not taken). There is absolutely no good reason why an off the shelf 32bit PCMCIA cardbus enclosure, 4 x 2GB SD style memory chips, an off-the-shelf 33MHz Broadcom interleaving memory controller/DSP combo chip and an LSI with 256KB EPROM costs $1200. That is insanity. Sorry, but P2 is being sold for 4X what it should be and if Panasonic doesn't push the 16GB cards out of the factory for less than $800 within the next two months, then P2 will be about 12X what it should be for the capacity. And no, I'm not basing my prices off of junk memory. I'm comparing with the same Samsung or Siemens SD style FLASH memory that is zero-fault rated. The very chips already being used in P2 cards. P2 doesn't have to use 4x SD either... The P2 spec does not govern what type of RAM or what RAM config is used inside. Someone could very well license the P2 spec and create P2 compliant PCMCIA adapters that hold single CF cards as long as it meets the bandwidth criteria, which isn't anything special and most current 8GB 120X or 150X CF media is more than half way there. The max advertised transfer rate of P2 is 640Mbps, which no current cards offer... It would take 4 * SD chips rated at 180X to reach that. Current cards are manufactured with 133X SD chips and give a bit less than 500Mbps throughput. The max throughput of a 32bit cardbus interface is 1033Mbps (129MB/s).

Utilizing 4 * CF media rated at 120X shoult theoretically give 576Mbps (72MB/s). 8GB 120X rated with zero-fault chips run about $160 each. Conventional without the ZF testing are about $120 each in retail pack.

I know Jim hasn't made any public commitments with RED FLASH, but I'm assuming will learn this in March or at NAB. And I will be very surprised if it's a proprietary format. Rather I expect it to be based on a common form like CF media. However, RED will most likely contract with a good supplier like SanDisk, Transcend, Etc... and will ship properly ZF tested media with the RED label. Both those manufacturers I just mentioned showd 32GB CF cards at CES with sustained write speeds in excess of 50MB/s for a single card and they should be shipping later this year. There is absolutely no reason to invest in proprietary interfaces or formats.

[Jannard]

You guys are speculating professionals... 4 CF cards? Striped together? :-(

We are bench testing a wide variety of flash and I think you all would be very surprised with the results.

Sleep tight. RED is awake.

Jim

[Jeff Kilgroe]

Speculation is my middle name.

Thanks for the input, Jim! As always, your words are encouraging... Now I have to throw away the CF array idea. :)

[Jannard]

I can't imagine the CF card shuffle... and if you get the right capacity with the correct write speed in the convenience of a single unit for the right price... what does it matter if it is proprietary or not? Not syaing it will be, but we should focus on what really matters.

Jim

[Jeff Kilgroe]

I can't imagine the CF card shuffle... and if you get the right capacity with the correct write speed in the convenience of a single unit for the right price... what does it matter if it is proprietary or not? Not syaing it will be, but we should focus on what really matters.

All true. Proprietary systems aren't necessarily a bad thing, as long as they can continuously deliver what the user needs for the right money and be readily available. I just don't want to see RED FLASH become another P2.

Anyway, in RED I trust... I can't wait to see what you guys come up with.

[tj williams]

JIm

Thanks for the reasurance: I'm sure the RED flash cards will be excellent and work as we always hope all "certified" parts will work.

I wish you had committed to open standards:

Finner quote: "I dont want it to be red specific in case of an emergency situation. Having Red overnight you flash memory cards does not help when its 9AM and you could just send a PA to a computer store to grab you some flash cards."

In a community of INdy, overseas, remote location workers, who have only so much money for backup, there will be instances where availability of drives and cards, immediately, will be critical. Certainly sometimes these non certified parts, may fail, but much more often they will save the day, and bring reliability kudos to the RED.

We lost the viewfinder(it fried) on our HDcam in Zimbabwe Africa several years ago. "Hey it's just not an overnight place." We ordered it "overnighted" from NewJersey and just managed to meet the plane. two days later, in Zambia. If this was an open source computer part like cards or disks we could have had it that afternoon from Capetown.

[Draccan]

While errors are a concern... speed is the biggest. The guaranteed minimum write speed has to be greater than 30+MB per second (not just advertised). Speed and reliability are mandatory. If we open to a standard format, say Compact Flash, and someone buys an 8GB card that doesn't work because the write speed is too slow... is it our fault? This is a difficult decision for us. We want an open system. But we also don't want unhappy customers. Two years from now there will be lots of great options. But right now, the options are few.

The good news is that we have the capability to change our flash drive in the future if something great, and open, comes along. Modularity is a good thing.

Jim

Wouldnt it be easy to find a manufacturer to partner with who already make a standard, say Compact Flash, and then have them print a logo on "approved for red" (for the ones with the right speed)? Catering to the RED crowd.

I mean surely that wouldnt cost them much and you only have to guarantee those?

I think it would be great with a common format.

[Jarred Land]

you find me a Compactflash card that can write over 30MB/s by itself and I will put those stickers on myself.

[Jaime Vallés]

you find me a Compactflash card that can write over 30MB/s by itself and I will put those stickers on myself.

You will put the stickers on the Compact Flash cards, or on your own body? ;)

[Kyle Mallory]

According to the Wikipedia: "The CF 3.0 and 4.0 standards have also been released which supports up to 66 Mbyte/s data transfer rates, and a number of other features. The CF 4.0 standard supports IDE Ultra DMA 133 for a maximum data transfer rate to 133MB/sec."

To further, the CF standard, as it stands, supports the largest storage capacity of any standard (typical?) flash media at 137GB for CF 1.0. According to the source (CompactFlash Association), CF 3.0 supports 66MB/s. Revision 4, bumps the speeds from 66MB/s to 133MB/s.

The downside of CF... No write-protect switch. Once you've pulled the card, there's nothing to prevent you from trashing your media once you've inserted it into whatever (camera, printer, pc, pda).

My two cents on this topic... Ensure that the camera-side of things regarding flash is at, or exceeding, the specs for the format. I've got a dozen still cameras that use CF and SD media... but imagine my frustration when I discovered that my new high-speeds cards have a faster read/write speed than the camera can support.

[Rob Lohman]

According to the Wikipedia: "The CF 3.0 and 4.0 standards have also been released which supports up to 66 Mbyte/s data transfer rates, and a number of other features. The CF 4.0 standard supports IDE Ultra DMA 133 for a maximum data transfer rate to 133MB/sec."

Support unfortunately does not mean it actually is available / works reliably

[tj williams]

This is what I draw from the posts so far:

First it seems clear: RED developers will choose a card/card array that works at all data rates. RED developers, will certify that this card/card array will work when purchased from RED.

RED will purchase cards from some supplier, and test them, label them and mark them up for their trouble and to make a profit. All Good! and I for one am willing to pay such a markup, the same way I buy the most premium tape, or new (in date not short ends) Eastman film. Failure elimination is good!

If you make available the mfg. model. number of the card you use, then in an emergency we can order up such cards where we are, when we need cards NOW!

If we cannot find the specific card, then we can take a chance on other cards that appear to be fast enough. We can reduce to 2K whatever to be able to shoot then. This is better than not having media that day because it was: crushed, dropped overboard, lost in luggage, set afire, lost in shipping, held up in customs, etc etc.

[jbeale]

Support unfortunately does not mean it actually is available / works reliably

so true... for example the original SATA spec is 1.5 Gb/sec and the newer SATA is 3 Gb/sec. Since SATA uses 8B/10B encoding, the theoretical max bitrate is 1.2 Gb/2.4 Gb respectively (150 MB/s or 300 MB/sec). http://en.wikipedia.org/wiki/Serial_ATA

Of course that is just the interface, not the media. Does an external hard drive labeled "SATA II" have a true sustained transfer rate of 300 MBytes/sec? Nope.

I have one 500 GB eSATA drive which I measured at 54 MB/sec, but no single drive on the market even approaches the SATA theoretical max.

[Finner]

This is what I draw from the posts so far:

First it seems clear: RED developers will choose a card/card array that works at all data rates. RED developers, will certify that this card/card array will work when purchased from RED.

RED will purchase cards from some supplier, and test them, label them and mark them up for their trouble and to make a profit. All Good! and I for one am willing to pay such a markup, the same way I buy the most premium tape, or new (in date not short ends) Eastman film. Failure elimination is good!

If you make available the mfg. model. number of the card you use, then in an emergency we can order up such cards where we are, when we need cards NOW!

If we cannot find the specific card, then we can take a chance on other cards that appear to be fast enough. We can reduce to 2K whatever to be able to shoot then. This is better than not having media that day because it was: crushed, dropped overboard, lost in luggage, set afire, lost in shipping, held up in customs, etc etc.

TJ

The problem mentioned throughout this thread is if RED puts a "red" designed port on their flash card system we will be forced to only use the red type cards. Thus the option of off the shelf cards would not even be possible because they will have a standard industry port that will not fit in the camera. I also would be happy to pay a little more for a RED recomended flash card sold by them. I just would like the port to be a computer standard version so in the emergency situations you can buy off the shelf flash.

When you say if red tells us who the manufacturer is... this will make no difference because if the port is red specific they will not have flash that fits. Computer memory companies mass produce standard port memory. As far as specialty type they rarely do, especially if they do not have the rights to and even if they have the rights to they charge way way more for the same amount of memory if it has a specialty port.

[Chris Kenny]

There is no memory card standard which reliably handles the data rates Red needs. This isn't a matter of Red putting a standard port on the camera, or a proprietary version of basically the same thing. REDFLASH is almost certainly going to be a sort of RAID using multiple memory chips/cards. The interface between the camera and the digital magazine is going to be something which supports that. There is no generic version of that.

[Stokestack]

But as I see it coming, the RED FLASH MEMORY INTERFACE for the camera will be plugging in to the RED-DRIVE SATA connector... SATA connector on the camera body is modified to supply the power since the original SATA spec do not have provision to deliver power to the SATA device.

Is the existence of this port on the Red confirmed?

And I couldn't believe eSATA failed to provide for power, considering it's a standard from 2004. What a ridiculous omission. People never learn.

[Jeff Kilgroe]

There is no memory card standard which reliably handles the data rates Red needs. This isn't a matter of Red putting a standard port on the camera, or a proprietary version of basically the same thing. REDFLASH is almost certainly going to be a sort of RAID using multiple memory chips/cards. The interface between the camera and the digital magazine is going to be something which supports that. There is no generic version of that.

...Although, several new FLASH products were just shown at CES that could handle the data rate.

IMO, the more I have thought about it the last two days, I think RED should adopt an ExpressCard interface and use standard EC storage media. Then we could pop the card right out of the RED FLASH unit and put them right into a Macbook Pro or newer PC notebook or other system that has an ExpressCard slot. Essentially taking Panasonic's P2 design to the next level. Also don't cripple the standard by making the cards and/or interface proprietary like P2. There will be lots of EC storage devices based on FLASH memory and many of those due out over the next several months will have transfer rates plenty good enough. SanDisk showed both CF and EC media at 32GB and 64GB at CES that had write speeds in excess of 30MB/s. Read speeds were nearly 50MB/s. Off-the-shelf FLASH products that can cope with RED's I/O requirements are literally a couple months away.

[Jeff Kilgroe]

Now, out of curiosity, while we're on the subject of FLASH and mini-RAIDs and all that. What file system(s) will be used on the RED FLASH media and the RED DRIVE? Anyone at RED able to comment on that yet? Has it even been decided? Hopefully not FAT32 so we don't have to put up with multiple partitions on RED DRIVE. NTFS is great, but OSX can't write to NTFS volumes and HPFS requires third-party software on Windows. Hmmm... I guess the FLASH and DRIVE units could use a common format and then provide translation for one platform or the other. ZFS would be cool to use, but I'm not sure on the licensing from Sun... OSX 10.5 will support ZFS. Windows does not natively... Hmmm.

[Chris Kenny]

The file system question is a good one.

ZFS would be cool, because it checksums everything transparently, and in REDFLASH, where there will probably be a sort of mini-RAID setup with multiple chips/cards internally, you could even conceivably use RAID-Z for redundancy. It would also be really slick if you could use ZFS features on external RAID arrays. Just hook up a box of SATA drives (with a port multiplier), and set up any combination of striping/mirroring/RAID-Z across them. I suspect, however, that the overhead and complexity of ZFS might be rather high for an embedded system.

NTFS isn't writable out of the box in OS X, but is this a huge deal in this application? The only real consequence of this would be that you couldn't reformat drives/cards on OS X, but the camera could presumably format them, so this wouldn't be a huge problem, right? On the other hand, I'd be really afraid to erase a digital mag if I weren't the one who'd just personally witnessed it downloading, so maybe the inability to erase it directly on the machine that downloaded it would be a big deal. MacFUSE (http://code.google.com/p/macfuse/) implements writable NTFS on OS X, but I don't know how well it works.

Apple's HFS+ isn't conveniently readable on Windows systems, so that's probably out, unless Red wants to license some of the existing software that does this and bundle it with the camera.

Red probably doesn't want to write new file system drivers for two or three operating systems, so something custom is probably out.

This basically leaves FAT32....

Volume size with FAT32 isn't actually a huge problem. It can go up to 8 TB with 32K clusters (which wouldn't waste a lot of space in this application.) The 32 GB limitation a lot of people are familiar with is actually just a limitation in the Windows disk formating utility. OS X and numerous third-party Windows utilities will create larger FAT32 volumes, which Windows can use with no problem.

The real killer for FAT32 is the 4 GB maximum file size. This would mean any shot over about 2.4 minutes (at the REDCODE 4K data rate) would have to be broken into multiple files, which would be a big hassle in post. REDCINE could have a feature where it recombined them automatically as appropriate while copying them to your computer, though, which would make it tolerable.

[Jarred Land]

Is the existence of this port on the Red confirmed?

And I couldn't believe eSATA failed to provide for power, considering it's a standard from 2004. What a ridiculous omission. People never learn.


yes... the sata port on the back of the camera goes to the RED DRIVE and records 4k Redcode.

No power on E-sata is indeed seriously retarded. It makes Firewire 800 look like gold.

even more retarded about Sata is that the power port uses so many damn pins.. 3x the amount it uses for data.

[Chris Kenny]

There will be lots of EC storage devices based on FLASH memory and many of those due out over the next several months will have transfer rates plenty good enough. SanDisk showed both CF and EC media at 32GB and 64GB at CES that had write speeds in excess of 30MB/s. Read speeds were nearly 50MB/s. Off-the-shelf FLASH products that can cope with RED's I/O requirements are literally a couple months away.

And there are now SATA flash drives with similar performance. But 30 MB/s is really cutting it too close for REDCODE RAW 4K (27.5 MB/s) + four channels of audio. The interfaces support more, but the products currently on the market don't. True, they probably will in six months.

[Jarred Land]

Fat32, as much as it sucks, is the best solution for the application... and is why other systems like Panasonic's P2 uses it.

Multiple clips are not necessarily a bad thing.. i would rather have files save and new files open every 2 minutes while i'm recording long form.. because if a file gets corrupt, its only one small piece.

[Jarred Land]

And there are now SATA flash drives with similar performance. But 30 MB/s is really cutting it too close for REDCODE RAW 4K (27.5 MB/s) + four channels of audio. The interfaces support more, but the products currently on the market don't. True, they probably will in six months.

most of the specs manufacture's release are bogus.. they are at best measuring small bursts of speed.

There were some nice drives announced at CES that look to be fast enough.. but none of those are shipping, and wont be available for some time.

[Stuart English]

Jarred beat me to this post - above all don't believe what the spec says - test it!

Our experience is that few of the data write speed claims stack up in practice, so you would need a very healthy margin between needed speed and claimed speed before you could trust that off-the-shelf (non certified) media would always be up for the task. And there is nothing more critical than reliably recording your images.....

[Stuart English]

Multiple files per clip could be a pain in post, but its really a question of how they are handled / presented. There is no doubt that some of the existing IT based camcorder systems have failed miserably in that regard.

[Stuart English]

Perhaps as we mature the camera system over the coming months we could enable a mode where we can measure the speed of a piece of off-the-shelf media that you may need to use in an emergency situation. We'd have to write then read back test signals to the media and based on those results either lock out certain recording resolutions / frame rates or work out how to lower the recorded image quality via a higher compression setting.

If it were possible, is it a desireable function to have?

[Jeff Kilgroe]

Fat32, as much as it sucks, is the best solution for the application... and is why other systems like Panasonic's P2 uses it.

OK... I figured as much.


Multiple clips are not necessarily a bad thing.. i would rather have files save and new files open every 2 minutes while i'm recording long form.. because if a file gets corrupt, its only one small piece.

True, but such breaks can be introduced without resorting to FAT32 file and volume size limits to impose them. Ironically, if a more robust file system were being used (like ZFS, HFX, XFS, EFS, UDF etc..) data corruption would be a lot less common. FAT32 isn't very robust at all and the only reason people still use it is because, well... I honestly don't know.

Seriously, here's a consideration. UDF. The UDF file system is supported by OSX, Windows, Linux, BSD, Sun and even legacy Irix. It's not limited to 4GB file sizes or 32GB per disk/partition. It's the standard being used for both HD-DVD and BluRay media. Also used by various magneto-optical formats, prototype holographic units, high-capacity cross-platform FLASH products like some of those demoed at CES, XDcam media and Iomega REV drives. UDF is an open standard. And if you download one of the various UDF format utilities out there, you can format an external hard drive as UDF and it plugs into a Mac, Windows or Linux system and works just fine.

Sorry to pick at this, but FAT32 is just so 1998. It works OK for something 32GB and smaller, but the thought of FAT32 and having 10 partitions on a RED DRIVE with all video files capped at 4GB no matter what isn't all that appealing.

[Finner]

Perhaps as we mature the camera system over the coming months we could enable a mode where we can measure the speed of a piece of off-the-shelf media that you may need to use in an emergency situation. We'd have to write then read back test signals to the media and based on those results either lock out certain recording resolutions / frame rates or work out how to lower the recorded image quality via a higher compression setting.

If it were possible, is it a desireable function to have?

This is massively desirable. I would want to use red flash memory and would pay extra for the peace of mind but it would be terrible not to have any back up options if a person ran into an emergency situation.

You just mentioning looking at the option makes me feel a lot more comfortable with this workflow.

[Jarred Land]

OK... I figured as much.

True, but such breaks can be introduced without resorting to FAT32 file and volume size limits to impose them. Ironically, if a more robust file system were being used (like ZFS, HFX, XFS, EFS, UDF etc..) data corruption would be a lot less common.

Sorry to pick at this, but FAT32 is just so 1998. It works OK for something 32GB and smaller, but the thought of FAT32 and having 10 partitions on a RED DRIVE with all video files capped at 4GB no matter what isn't all that appealing.

I agree with you 100%.. Fat32 was the cool new kid on the block in friggin 1995 for christ sakes. But unless some kinda christmas miracle happens, I think we will still see Fat32 floating around in many devices and drives for the next decade.

[Chris Kenny]

Sorry to pick at this, but FAT32 is just so 1998. It works OK for something 32GB and smaller, but the thought of FAT32 and having 10 partitions on a RED DRIVE with all video files capped at 4GB no matter what isn't all that appealing.

See my earlier post. FAT32 volumes go up to 8 TB. The 32 GB thing is just a limitation of the default Windows formatting utility. It's not a limitation of the file system itself or of its implementation in Windows, OS X, or Linux.

[Jarred Land]

You just mentioning looking at the option makes me feel a lot more comfortable with this workflow.

Welcome to RED :)

[Jeff Kilgroe]

I agree with you 100%.. Fat32 was the cool new kid on the block in friggin 1995 for christ sakes. But unless some kinda christmas miracle happens, I think we will still see Fat32 floating around in many devices and drives for the next decade.

It's sad, but I know you're right...

[Jeff Kilgroe]

See my earlier post. FAT32 volumes go up to 8 TB. The 32 GB thing is just a limitation of the default Windows formatting utility. It's not a limitation of the file system itself or of its implementation in Windows, OS X, or Linux.

Er... Microsoft chose to limit FAT32 to 32GB partitions for a reason. I've got the whole FAT32 spec in a binder on a shelf here. Yes, it can be extended to much larger volumes -- actually beyond 8TB if you really wanted to. That doesn't mean it should be done. Capacities aside, there are plenty of good reasons to NOT USE FAT32.

FWIW, Windows isn't the only operating system (or disk management software) out there that limits FAT32 volume sizes.

With FAT32, there is no internal garbage collection. No real error correction or data verification. No drive surface mapping functions to monitor for bad or currupt data areas, etc.. FAT32 is just a way to dump bits of data to a storage volume and use the allocation table (over-glorified table of contents) to find everything.

I don't even trust NTFS or HPFS for my SAN volumes in excess of 1TB over RAID-5, why would I want FAT32 in anything that size or larger?

[Chris Kenny]

Jarred beat me to this post - above all don't believe what the spec says - test it!

Our experience is that few of the data write speed claims stack up in practice, so you would need a very healthy margin between needed speed and claimed speed before you could trust that off-the-shelf (non certified) media would always be up for the task. And there is nothing more critical than reliably recording your images.....

Well, those flash hard drives drives I mentioned get pretty close to the 30 MB/s, and of course they're stolid state, so they're a lot more consistent than mechanical storage. Take a look here: http://www.tomshardware.com/2006/09/20/conventional_hard_drive_obsoletism/page4.html#data_transfer_diagram

But, like I said, this is pretty iffy for REDCODE RAW 4K + audio. If that line were close to 40 MB/s instead of 30 MB/s this format would be more plausible. Although, it's probably not exactly the form factor you want.

[Chris Kenny]

Er... Microsoft chose to limit FAT32 to 32GB partitions for a reason.

[...]

I don't even trust NTFS or HPFS for my SAN volumes in excess of 1TB over RAID-5, why would I want FAT32 in anything that size or larger?

Frankly, in this application, the answer is because it probably beats messing with 10 partitions on a 300 GB REDDRIVE. Unless REDCINE does, in fact, hide all of this nastiness from the user, in which case smaller partitions would be fine.

[Jeff Kilgroe]

Unless REDCINE does, in fact, hide all of this nastiness from the user, in which case smaller partitions would be fine.

Hopefully it does hide such things... I would expect it to actually because we have to transfer our footage from REDCODE RAW to REDCODE RGB or other editing format (unless I've misunderstood something along the way). And I doubt anyone is going to want the brain damage of managing their footage in clips no larger than 4GB.

[Chris Kenny]

Hopefully it does hide such things... I would expect it to actually because we have to transfer our footage from REDCODE RAW to REDCODE RGB or other editing format (unless I've misunderstood something along the way). And I doubt anyone is going to want the brain damage of managing their footage in clips no larger than 4GB.

Right, but it's conceivable it doesn't have a 'copy' mode, but only a 'transcode' mode. If you can just copy footage straight off a single partition in appropriately sized files, this is fine, because you can just use the file manager to copy your 'digital negative'. But if there are going to be multiple partitions and split files, you'll want to do copies though REDCINE to hide all that stuff.

Actually, you probably want a 'copy' mode even if you don't have to hide file system limitations, so you can automate the process of copying, verifying and possibly reformatting. Red is probably ahead of me on this stuff anyway, though.

Actually... here's a request for a less obvious feature. Give REDCINE a copy mode that lets it write footage to two targets at once. This should let you end an ingest with your footage already backed up, with no significant time penalty (if you're writing to at least two physical drives, of course). Yes, there's RAID mirroring, but that doesn't protect against all the failure modes that writing to two independent volumes does.

One reason why this is desirable is because it means even the cautious can erase their digital mags and return them to shooting immediately after ingest (otherwise they're going to wait for a backup of the ingested footage to finish before they wipe the digital mag). And, actually, if you could do things this way, you could confidently set REDCINE to automatically erase a mag after ingest/verification. This would mean you'd almost never have to manually format mags, which would be desirable because if erasing them is a manual process, someone, somewhere, is going to have a really expensive "oops" moment.

[Rob Lohman]

Frankly, in this application, the answer is because it probably beats messing with 10 partitions on a 300 GB REDDRIVE. Unless REDCINE does, in fact, hide all of this nastiness from the user, in which case smaller partitions would be fine.

As indicated before the 32 GB is an artificial limit, and yes, I have the FAT32 spec as well. It will be a single partition. Obviously file size limit is still there.

Writing and installing file system drivers is a whole other ballgame. UDF has similar problems since it is not supported on a hard drive out of the box on some Operating Systems. Yes, this has been tested and verified.

I don't get the earlier remark about not trusting NTFS on a large volume (I can't speak for HPFS). It's an extremely reliable journaling file system. I've used it for over 10 years and the server market trusts it for a good reason.

We all feel the same as you do, but it is what it is. We need a system where you can copy data off without the need to install file system drivers and whatnot. This has all been thought out, discussed and debated for months. It's not a simple decision.

Finally: we are looking at the complete workflow so it will be as smooth and easy as we can make it. There will be tools to assist. Everything about the product is modular so we can upgrade (on both ends) if something better comes along. For now there's just not much choice.

[Blair S. Paulsen]

I really appreciate Rob giving us the scoop on how much thought was put into the file system selection. It seems to me that the biggest headache is that coding RedCine will be tougher. RedTeam coders, I hope you get to see your families again someday...

One feature request that slipped into this thread that I want to echo is for running multiple backups of the DCN from RedCine. The RedTeam is probably all over this but the file system discussion got me thinking just how many issues could crop up in the very critical backing up task. Making back-ups simple and solid is a deal-maker/deal-breaker issue during the transition to the no tape, no film workflow - bungle this and we got BIG trouble.

[Rob Lohman]

I had to say goodbye to my family when I joined RED, but I got a new one in return! ;)

REDCINE coding is not really complicated by this. It will load a sequence of camera files as one movie (or not).

[Jeff Kilgroe]

As indicated before the 32 GB is an artificial limit, and yes, I have the FAT32 spec as well. It will be a single partition. Obviously file size limit is still there.

OK. I turst you guys to do your thing... So I will also assume that REDCINE will include an interface to format drives and FLASH media?Windows requires command line intervention on the users' part to format larger than 32GB.


Writing and installing file system drivers is a whole other ballgame. UDF has similar problems since it is not supported on a hard drive out of the box on some Operating Systems. Yes, this has been tested and verified.

I personally haven't run into any systems that don't support UDF on hard drive out of the box for general use, but then again, I'm just using it with my own XP/Vista installs, OSX 10.4 (10.5 beta)... Some Linux distributions include UDF formatting abilities -- I do my UDF formats with SUSE Linux. But you're right, this is an issue.. You would have to provide your own format utility and other disk tools. UDF support in OSX and Windows is fairly minimal. In fact, Windows requires a registry mod to use UDF as you would a normal hard drive. Otherwise it tries to treat it like packet-write optical media.

Ideally, it's not a good situation just yet. I do think UDF will replace FAT32 in the coming years unless something miraculously better comes along that is cross-platform. Microsoft is fully committed to NTFS (which is fine I guess), but it's too bad they won't share full licensing with everyone else, citing security as the reason.


I don't get the earlier remark about not trusting NTFS on a large volume (I can't speak for HPFS). It's an extremely reliable journaling file system. I've used it for over 10 years and the server market trusts it for a good reason.

I don't trust NTFS compared to something more robust like ZFS. NTFS is a fine file system and has excellent security features and journaling abilities. But it also lacks a lot of the self-managing features that go well beyond journaling that are available in a system like ZFS. Having managed a couple SAN systems in a large datacenter a couple years back, I can say that the Sun servers running ZFS for the SAN were far less problematic and required a lot less user interaction than Windows server systems with NTFS. HFS/HPFS on OSX is pretty equivalent to NTFS... It lacks a lot of the secuirty features, but I think the journaling is more efficeint and it's a lot easier and friendly to implement on other platforms (never mind Microsoft doesn't release license to NTFS for any purpose other than read-only). But Apple is incorporating ZFS support on a basic level within OSX 10.5 and on a much higher level with upcoming OSX server versions and XSan software.


Everything about the product is modular so we can upgrade (on both ends) if something better comes along. For now there's just not much choice.

I know I'm beating a dead horse with this discussion... And I'm mostly just fantasizing about the way it should be rather than the way it is. I know you guys will make the workflow as efficeint and reliable as possible.

[GlennChan]

And I doubt anyone is going to want the brain damage of managing their footage in clips no larger than 4GB.
You know what? It's not that hard. Having to deal with the annoying PC/NTFS versus OS X/HFS issue... FAT32 works. The only thing that doesn't work is when you try to make files that go right up to the 4GB limit. Vegas will try to save marker and region information in the originating AVI file, but can't because there's no room left (this is my guess anyways; slightly lower than 4GB seems to fix things).

Other than that, dealing with a bunch of clips instead of one isn't hard at all.

[Poi Boy]

I could care less what kind,how conveniently available, what brand,what standard,...does it work ? that is all I care about. I think you all are forgeting what we are talking about here. 4K for $30K or so...all these details fall into the whatever catagoty.
Aloha
-A