Tapeless acquisition -- convince me

Chris Hurd · December 9th, 2007, 10:58 PM

Quote:

Originally Posted by Mike Barber

short record times per card -- if I have to switch-out the card every 8 minutes, may as well go for film, eh!

Where in the world are you getting "eight minutes per card?" You're not stating which tapeless cameras you're looking at. But let's assume the Panasonic HVX200. It can hold a pair of 16GB P2 cards which can record a total of 80 minutes of 720p24 uninterrupted video in the DVCPRO HD format.

Quote:

short record times per card -- again... despite the hot swap-ability, what kind of camera shake am I going to get from swapping the cards out during a shoot?

Again, where are you getting the idea that you're stuck with short record times? Just use a FireStore portable hard drive recorder for several hours of 1080i or 720p HDV, or several hours of 720p24 DVCPRO HD, or 100 minutes of 1080i60 DVCPRO HD.

Quote:

no intuitive or affordable archive strategy -- I'm working on Project M which requires footage acquired for Project G from last year... with tape, I just go to my tape library and grab the source tape; I don't have the money or space to have a 10,000-Terabyte server to store all my footage.

Choose an HDV camcorder then. You can record either 720p or 1080i simultaneously to tape and to FireStore; the tape is your instant archive while the FireStore is your edit-ready video. If you don't choose the HDV format, then simply archive to hard drives as mentioned above (completely economical), or get an LTO-3 tape drive and back up your video to tape. There are a variety of archiving options available to you.

Quote:

These three issues alone tell me that tapeless acquisition is the dumbest thing since auto-focus. But I may certainly be wrong.

Yeah, you certainly could be wrong.

There's nothing dumb about it except not understanding it. In my opinion, tapeless acquisition is just about the smartest move you can make. The dumbest thing is uselessly wasting time capturing video from tape -- why not completely bypass the video capture process and record edit-ready video in the first place.

Bill Ravens · December 10th, 2007, 08:10 AM

I just don't see how anyone can claim IDE hard drives aren't cost effective! On a $/bit basis, about 31 cents per gig, they're the cheapest going, if you figure in time spent writing to tape. And tape ain't that reliable.Solid state?? very expensive storage media at approx. $20/gig.

Now, throw in the cost of a tape reader/writer and figure the depreciation due to head wear and tape looks pretty ridiclulous, if you ask me. Once BD aor HD-DVD gets de riguer, I agree, that will be the storage media of choice.

David Heath · December 10th, 2007, 08:42 AM

Quote:

Originally Posted by Bill Ravens

I just don't see how anyone can claim IDE hard drives aren't cost effective! On a $/bit basis, about 31 cents per gig, they're the cheapest going, if you figure in time spent writing to tape.

Yes, but you may also have to factor in the TIME spent making the backup/archive, whereas tape forms it's own archive. Then again, tape acquisiton may involve extra time spent digitising, which is a reason I find the S270/Z7 very interesting, recording to tape AND Compact Flash. The former for archive, the latter for more immediate use - similar to what Chris suggests with a tape/Firestore combo.

The original post reads more as it's talking not about "tapeless", but P2 in particular. P2 is ideal for some, but I'll accept has workflow issues for most. But not all tapeless, not all solid state even, now means P2, and the recent announcements by Sony about the S270/Z7 and Compact Flash mean solid state practicability has just taken a huge leap forward.

Bill Koehler · December 24th, 2007, 11:15 PM

Granted this is a little over a year out, but I think it will blow away any remaining ideas about "short record times".

Happy reading:

http://news.bbc.co.uk/2/hi/technology/7057717.stm

http://www.tgdaily.com/content/view/34487/135/

Bob Grant · December 25th, 2007, 03:48 AM

Acquisition and archiving are two separate issues. Using camera tapes for archiving is asking for problems, I've been asked too many times to help recover tapes from cameras with guide problems to know how silly this is. DV and DVCAM are far from archival media, digitalbetacam maybe, at least it's robust but who knows for how much longer the VCRs to play it will be around, although we can still playout 2" quad, Umatic is a real problem at times.

Tapeless acquisition has many advantages, SS media will withstand G forces that'll twist tape and apart from breaking / bending a pin on a card it's very robust. More to the point most of the factors that could cause it to fail are under the users control, the same cannot be said for tape. Sure the limited record times and costs of the SxS cards is an issue but I've also got a M15 HDV deck that lets me record 4.5 hours of HDV from a HDV camera. Yes it needs main power but that's rarely an issue at events where I need continuous recording for long periods. There's also the DR60 HDD recorder and other HDD HDV recorders that will run off batteries. The great thing though with the SxS cards is the higher datarate. Tape creates an upper limit on the quality you can record, tapeless frees us from that.

Storage onto RAID boxes has advantages, one is rapid access, the other is you can clone the whole box and store it offsite. It does cost more, don't forget to factor in the cost of power to keep the boxes runnning over the years. Still the cheapest archival storage is film, as tri-separations if you've got the money. Good article on this here:
http://www.nytimes.com/2007/12/23/bu...=5070&emc=eta1

Jason Hilton · December 31st, 2007, 09:12 PM

The first project I shot involved 12 hours of DV footage shot to tape. With an average log and capture of 2 hours per tape (log and capture time and comp time to actually caputre). So figure 24 hours of work just to get the footage into my NLE, time which the client pays for... Now, consider that against the current (very low) cost of hard drives. In my case, the client will be paying less to purchase 2 tiers of storage than they would be paying the time to log and capture. Even a small project is still cheaper to purchase hard drives. Invest in a hot swap raid array and the cost goes down further still (after the array and controllers are paid for).

I shoot entirely tapeless acquisition these days, mostly HVX200 and hopefully soon Red. Storage per project is now part of the expendables budget. I always budget at least 2 tiers of backup (one main and one backup), the amount based on how much we expect to shoot times our expected shooting ratio. If we're reviewing dailies on set I have a third tier for unwrap/log and transfer and viewing. Everything is then copied to storage on the edit system and the two tiers of onset storage are put into, well, storage. Optionally for the next project, we're investing in a blu-ray burner for final media back-up. Unfortunately BD is not yet cost effective, and at an average of 50 minutes burn time for a 25gb disk, not time effective on set either. But as archiving goes, it's far more reliable than tape, and if you do two sets of blu-ray archive, you can then also reuse your hard disk storage. Assuming blu-ray comes down like DVD-R did, eventually it will be by far the most cost effective option.

Tapeless workflow definitely saves time and energy, especially in the post process. It takes a little more workflow management than you might be used to.

John M. McCloskey · February 18th, 2008, 03:46 PM

Ok, how would you handle this with P2Tapeless or SS cards. 9 camera guys every week in different locations bring in weekly 100 to 120 hours of video each for some one else to edit for the following year shows(in other words the footage doesnt get touched until the following year, we stay a year ahead). They keep this weekly shooting regimine for 11 months out of the year. So 100X9=900 hrs a week of video X 11 months approx.44weeks. 900X44=39,600 hours of footage for 56 different projects each being a differnet episode. Now we label our tapes and archive them on the shelf with a labeling system. I know how important 10 year old footage is in our productions so we must be able to have that feature. Looking at the XDHDCam being able to shoot, put on shelf, and work in proxies seems like the best alternative for us to go tapeless, does anyone have any other suggestions. THX

John M. McCloskey · February 18th, 2008, 04:32 PM

Sorry it isnt 100 to 120 hours a week each, its more like 60 to 80 hours a week each. So instead of 39,600 it would be around 27,750 hours of video to archive tapelessly each year.

Jim Andrada · February 21st, 2008, 11:54 PM

Interesting - if you have this much material, and it was tapelessly acquired, backup/archive to LTO digital tape as used for computer backup these days would be entirely feasible. Current LTO tape cartridges hold 800GB (uncompressed) and the next generation due out around the end of next year would hold 1.6 TB, and 2 years later the capacity is scheduled to be 3.2 TB per cartridge.The LTO consortium (IBM, HP, Quantum) has hit the capacities published in the roadmap right on schedule for 4 product generations so far.

Current sustained data rates to a single tape drive are in the ballpark of 120MB per second, so roughly speaking an hour of video would take in the vicinity of 2 minutes to archive from hard drive to tape, assuming the system were properly configured - you'd probably need raided hard drives to pull this off and keep the data streaming fast enough, but it shouldn't be impossible. Again really rough numbers, but you could (perhaps a bit optimistically) archive your year's worth of video in around 1k total hours - not such a big deal. Again this assumed you were only using a single tape drive but they're pretty cheap so having a few tape drives and a moderate sized Raid disk system I don't see this as a monumental task.

There are lots of good medium sized tape libraries available (IBM, NEC in Japan, HP,etc) that could completely automate the process - the operator would probably have to remove cartridges from the library and move them to shelf storage weekly or monthly or you could use larger libraries that would hold easily several years or even decades of material acquired at the rates you mention.

This technology is also extremely reliable. An LTO cartridge has a normal lifetime of 5,000 mounts, but I've seen some cartridges that survived 20,000 reuse cycles. The most reliable small library I'm aware of, made by NEC in Japan, was tested to demonstrate 5 Million tape mounts between failures, but I think any available library is rated at a minimum of 2 million.

I think that LTO tape accounts for about 80% of the world's backup and archive market these days so this stuff is entirely for real.

http://www.networkworld.com/news/200...-lto-tape.html

If you have the number of guys shooting video that you mention I think your company is certainly large enough to be looking at LTO automation.

December 10th, 2007, 08:10 AM	#17
Bill Ravens Join Date: Nov 2001 Location: santa fe, nm Posts: 3,264 Images: 10	I just don't see how anyone can claim IDE hard drives aren't cost effective! On a $/bit basis, about 31 cents per gig, they're the cheapest going, if you figure in time spent writing to tape. And tape ain't that reliable.Solid state?? very expensive storage media at approx. $20/gig. Now, throw in the cost of a tape reader/writer and figure the depreciation due to head wear and tape looks pretty ridiclulous, if you ask me. Once BD aor HD-DVD gets de riguer, I agree, that will be the storage media of choice. Last edited by Bill Ravens; December 10th, 2007 at 08:56 AM.

December 24th, 2007, 11:15 PM	#19
Bill Koehler Major Player Join Date: May 2007 Location: Alpharetta, Georgia, USA Posts: 760	Granted this is a little over a year out, but I think it will blow away any remaining ideas about "short record times". Happy reading: http://news.bbc.co.uk/2/hi/technology/7057717.stm http://www.tgdaily.com/content/view/34487/135/ Last edited by Bill Koehler; December 25th, 2007 at 12:05 AM.

December 31st, 2007, 09:12 PM	#21
Jason Hilton Regular Crew Join Date: Jan 2005 Location: United States Posts: 26	The first project I shot involved 12 hours of DV footage shot to tape. With an average log and capture of 2 hours per tape (log and capture time and comp time to actually caputre). So figure 24 hours of work just to get the footage into my NLE, time which the client pays for... Now, consider that against the current (very low) cost of hard drives. In my case, the client will be paying less to purchase 2 tiers of storage than they would be paying the time to log and capture. Even a small project is still cheaper to purchase hard drives. Invest in a hot swap raid array and the cost goes down further still (after the array and controllers are paid for). I shoot entirely tapeless acquisition these days, mostly HVX200 and hopefully soon Red. Storage per project is now part of the expendables budget. I always budget at least 2 tiers of backup (one main and one backup), the amount based on how much we expect to shoot times our expected shooting ratio. If we're reviewing dailies on set I have a third tier for unwrap/log and transfer and viewing. Everything is then copied to storage on the edit system and the two tiers of onset storage are put into, well, storage. Optionally for the next project, we're investing in a blu-ray burner for final media back-up. Unfortunately BD is not yet cost effective, and at an average of 50 minutes burn time for a 25gb disk, not time effective on set either. But as archiving goes, it's far more reliable than tape, and if you do two sets of blu-ray archive, you can then also reuse your hard disk storage. Assuming blu-ray comes down like DVD-R did, eventually it will be by far the most cost effective option. Tapeless workflow definitely saves time and energy, especially in the post process. It takes a little more workflow management than you might be used to. __________________ http://www.tumblemonster.com

December 25th, 2007, 03:48 AM	#20
Bob Grant Trustee Join Date: Nov 2005 Location: Sydney Australia Posts: 1,570	Acquisition and archiving are two separate issues. Using camera tapes for archiving is asking for problems, I've been asked too many times to help recover tapes from cameras with guide problems to know how silly this is. DV and DVCAM are far from archival media, digitalbetacam maybe, at least it's robust but who knows for how much longer the VCRs to play it will be around, although we can still playout 2" quad, Umatic is a real problem at times. Tapeless acquisition has many advantages, SS media will withstand G forces that'll twist tape and apart from breaking / bending a pin on a card it's very robust. More to the point most of the factors that could cause it to fail are under the users control, the same cannot be said for tape. Sure the limited record times and costs of the SxS cards is an issue but I've also got a M15 HDV deck that lets me record 4.5 hours of HDV from a HDV camera. Yes it needs main power but that's rarely an issue at events where I need continuous recording for long periods. There's also the DR60 HDD recorder and other HDD HDV recorders that will run off batteries. The great thing though with the SxS cards is the higher datarate. Tape creates an upper limit on the quality you can record, tapeless frees us from that. Storage onto RAID boxes has advantages, one is rapid access, the other is you can clone the whole box and store it offsite. It does cost more, don't forget to factor in the cost of power to keep the boxes runnning over the years. Still the cheapest archival storage is film, as tri-separations if you've got the money. Good article on this here: http://www.nytimes.com/2007/12/23/bu...=5070&emc=eta1

February 18th, 2008, 03:46 PM	#22
John M. McCloskey Major Player Join Date: Jun 2005 Location: West Point, MS Posts: 313	Ok, how would you handle this with P2Tapeless or SS cards. 9 camera guys every week in different locations bring in weekly 100 to 120 hours of video each for some one else to edit for the following year shows(in other words the footage doesnt get touched until the following year, we stay a year ahead). They keep this weekly shooting regimine for 11 months out of the year. So 100X9=900 hrs a week of video X 11 months approx.44weeks. 900X44=39,600 hours of footage for 56 different projects each being a differnet episode. Now we label our tapes and archive them on the shelf with a labeling system. I know how important 10 year old footage is in our productions so we must be able to have that feature. Looking at the XDHDCam being able to shoot, put on shelf, and work in proxies seems like the best alternative for us to go tapeless, does anyone have any other suggestions. THX

February 18th, 2008, 04:32 PM	#23
John M. McCloskey Major Player Join Date: Jun 2005 Location: West Point, MS Posts: 313	Sorry it isnt 100 to 120 hours a week each, its more like 60 to 80 hours a week each. So instead of 39,600 it would be around 27,750 hours of video to archive tapelessly each year.

February 21st, 2008, 11:54 PM	#24
Jim Andrada Inner Circle Join Date: Feb 2007 Location: Tucson AZ Posts: 2,211	Interesting - if you have this much material, and it was tapelessly acquired, backup/archive to LTO digital tape as used for computer backup these days would be entirely feasible. Current LTO tape cartridges hold 800GB (uncompressed) and the next generation due out around the end of next year would hold 1.6 TB, and 2 years later the capacity is scheduled to be 3.2 TB per cartridge.The LTO consortium (IBM, HP, Quantum) has hit the capacities published in the roadmap right on schedule for 4 product generations so far. Current sustained data rates to a single tape drive are in the ballpark of 120MB per second, so roughly speaking an hour of video would take in the vicinity of 2 minutes to archive from hard drive to tape, assuming the system were properly configured - you'd probably need raided hard drives to pull this off and keep the data streaming fast enough, but it shouldn't be impossible. Again really rough numbers, but you could (perhaps a bit optimistically) archive your year's worth of video in around 1k total hours - not such a big deal. Again this assumed you were only using a single tape drive but they're pretty cheap so having a few tape drives and a moderate sized Raid disk system I don't see this as a monumental task. There are lots of good medium sized tape libraries available (IBM, NEC in Japan, HP,etc) that could completely automate the process - the operator would probably have to remove cartridges from the library and move them to shelf storage weekly or monthly or you could use larger libraries that would hold easily several years or even decades of material acquired at the rates you mention. This technology is also extremely reliable. An LTO cartridge has a normal lifetime of 5,000 mounts, but I've seen some cartridges that survived 20,000 reuse cycles. The most reliable small library I'm aware of, made by NEC in Japan, was tested to demonstrate 5 Million tape mounts between failures, but I think any available library is rated at a minimum of 2 million. I think that LTO tape accounts for about 80% of the world's backup and archive market these days so this stuff is entirely for real. http://www.networkworld.com/news/200...-lto-tape.html If you have the number of guys shooting video that you mention I think your company is certainly large enough to be looking at LTO automation.