Protocol for Footage on Set

[Rocco Schult]

RED alert was supposed to be a temporary app, is that correct ?
I think theres still use for RED alert, even if REDcine is out.
A quick and dirty grading thing (exactly like it is now) and incorporating features needed On-Set, among them the stuff TJ collected and amended.

Just a little note on options and reformatting:

1. checksum source media/footage
2. copy files to one location, option of copying to 2 locations if available, option to reformat and unmount CF after copy
3. checksum destination, note if OK, or question what to do with options-dialogue
4. format CF or drive to record-ready status (incl. reel-number etc. if necessary)
5. write event into logilfe
6. unmount and hint for user if ready

Not meaning to leave everything out others mentioned.

If formatting is done by a RED app, RED could completely control the way and format how its formatted.

For me, REDalert could and should be the "need-2-be-done-on-set" app.

I'm going to add an idea, that will seem wierd but... I think should be taken seriously: In addition to being numbered all your CF cards or drives should have kid stickers on them of various random things. Human memory is aided by association. If you see the cards with the horsey in the reader, you are less likely to confuse it with the other card you have with Mr. Yuck on it.

BTW I'm not convinced about grounding the camera, I think that's overkill.

Ian, this is way funny, but exceptionally right what you're suggesting! I think I might go for a similar way (different icons;-)

And yeah, I think grounding the thing is too much.
Since I am assembling computers, I've never smoked a board.
But if you're shooting on plastic floor it might be worth considering.

 

[Rocco Schult]

Can you do multiple sessions on an LTO?

Yes, you can. Depending on the software, but generally possible.

But sending LTO to the editor is nonsense. You need an LTO drive at/on/in access for the editor instead of using the beenfit of ddirectly importing the files. We send HDDs since years, anf if one should break which of course is possible, you've got the backup.

 

[Chris Kenny]

But sending LTO to the editor is nonsense. You need an LTO drive at/on/in access for the editor instead of using the beenfit of ddirectly importing the files. We send HDDs since years, anf if one should break which of course is possible, you've got the backup.

Sending LTO might not be so crazy, if the expense of buying the editor a tape drive is not a big deal. The editor is almost certainly going to be copying the footage to his own RAID anyway, and LTO-3 and 4 are faster than most hard drives, so using LTO would probably actually make it faster for the editor to ingest footage.

 

[tj williams]

So I made some () and noted the more paranoid parts of the complete listing thanks for the new additions. Please double check and be sure that I am including your ideas clearly and correctly.
TJ

1. mark each media mag (CF or Drop in disk/memory based mag.) with its own mag. number color or symbol. This marking on the media itself, is used to trace any recording failure to the specific piece of media used. This is usually called the "mag number" This along with the "reel number", or sometimes called "roll number" which increments each time media is replaced, should be marked at the head of the recording. They can either be put in as data or by shooting a slate at the head of the recording.

2. record to either single media, or if allowed by RED Technology, to dual card or card/disk or dual disk mag. on camera.

3. (Dit wears a grounding wrist strap) (depending on static, plastic floor, wet etc.) Dit grounds camera to their body during media removal.(this is paranoid but is it paranoid enough?) Possibly Dit asks for the camera to be powered down during off loading? (ditto.) ChronoSync: http://www.econtechnologies.com/site..._overview.html[/url]) may be preferable to check sum. I strongly feel RED needs to provide the data check software as part of RED Cine to adapt a technology which they stand behind. If formatting is done by a RED app, (ie RED Cine) RED could completely control the way and format how its formatted. This also allows them to script the check sum of both the source and destination files and give a simple OK type output, so the Dit doesn't have to do the other step of matching the two.

4. If using single recording, pull partially filled card or drive mag immediately after crucial shots for transfer.

5. immediately pull the re-protect tab on CF media. (if this exists on RED CF media)

6. Place the recorded media into some particular bag or left pocket etc. where you always store the media until you get it out to put into the onset backup storage. (This is subject to the info. that handling is a major source of CF failure.also to avoid the possibility of confusion between the media just removed and the media which is ready, having been dubbed to be reused.)

7. Set computer and Esata or Firewire 800 raid are set up on an anit static matt.

8. download card to either dual LTO's, two separate drives, redundant raid. or perhaps some combination of slow raid/ fast raids/LTO's or seperate drives as seems good for the particular production. (S2 docking system. may be the best option at this time for uncompressed 4K as there is not yet a better system from RED)

9. RED cine automatically does check sum (or perhaps faster or better data check) and automatically gives a simple answer that data is copied correctly and checked and then and only then gives the option to reformat the card or disk mag.

10. RED cine reformats the disk with the same project settings. and automatically, writes event into log file, and unmounts media from computer. (or the Dit can do each of these tasks manually if RED does not support this.)

11. Assistant places the blank formatted media into some repeated storage bag or box and or perhaps keeps it in their right
pocket until it is reloaded into the camera (see () above)

12. media is installed onto camera and (if RED will support this) A menu item flashes on the camera asking for the reel number. The time code continues from the last media recording so there is not time code gap between reels. This will assure us that there is no reel missing on the backup, or that reels have not been copied to backup twice.

14. Deliver one to post house. *(when one of these must be shipped consider shipping LTO tape as it will probably survive better than disks in rough handling.) (It's necessary to consider that whoever receives the LTO esp. if a post house also has access to a compatable LTO playback machine.) (Editors preference to receive HDD or LTO probably good prep to work with Editor on short segment to see which works best in their workflow.)

15. Using a seperate delivery method deliver one to safe storage at a separate location from the post house.*
__________________
TJ

 

[I Bloom]

1. mark each media mag (CF or Drop in disk/memory based mag.) with its own mag. number color or symbol. This marking on the media itself, is used to trace any recording failure to the specific piece of media used. This is usually called the "mag number" This along with the "reel number", or sometimes called "roll number" which increments each time media is replaced, should be marked at the head of the recording. They can either be put in as data or by shooting a slate at the head of the recording.

It's possible that just like P2 cards CF cards each have a unique serial number or ID that is detectable by the software. In this case, tracing Mag Numbers can be automated. The same might be true for Red-drives.

The time code continues from the last media recording so there is not time code gap between reels. This will assure us that there is no reel missing on the backup, or that reels have not been copied to backup twice.

This is really important right, and it begs an interesting question that I think may have already been answered somewhere in the forum. It seems like every clip should be recording three channels of timecode:
1. A Freerun (jam synced to sound or other cameras)
2. A Record Run (useful for the purpose you just outlined)
3. A Time of Day (just generally useful see 1)

It's possible that the camera could always record all three all the time (tell me this is already the case.)

It might be nice then to visualize the order of recorded clips in REDCine, so that you can quickly solve problems or find missing footage in your left pocket.

Ian

 

[I Bloom]

By the way TJ and all, great work!

I put this up on the wiki:
http://www.redhax.net/wiki/Footage_Protocol_On_Set

I need to continue to rewrite it but I think we've done some really constructive work!

Ian

 

[Greg M]

So I made some () and noted the more paranoid parts of the complete listing thanks for the new additions. Please double check and be sure that I am including your ideas clearly and correctly.
TJ


__________________
TJ

You forgot to add:

1. Cover Post House windows with aluminum foil to prevent et transmissions from changing data on disks.
2. If crop circles appear in vicinity of post house, do not ship drives on third Thursday.

 

[tj williams]

DF
Thats silly crop circles only affect analog electronics.
TJ

 

[I Bloom]

I wish I was still shooting film. The problem with digital is the need to wear this stupid tinfoil hat

IBloom

 

[Rocco Schult]

Sending LTO might not be so crazy, if the expense of buying the editor a tape drive is not a big deal. The editor is almost certainly going to be copying the footage to his own RAID anyway, and LTO-3 and 4 are faster than most hard drives, so using LTO would probably actually make it faster for the editor to ingest footage.

You can plugin the drive and work from it, you can't with a LTO. Putting the data to tape is again another step as you can't write directly to the tape. You need to have the data laid down first and then archive them.
LTO is a backup medium and more than approproate at that in my eyes. But it takes double time compared to disc if worked from the disc.
All my current projects request the fastest way from acquisition to edit.
And if there should be a failure, yes, THEN theres a backup. ANd you can still dump the data onto the raid at night, after editing, but still gain the effect of profiting from the quick start.

Its another thing with uncomressed, there I agree with LTO being better, as you can't work directly from any drive-container (if its not a codex...)

By the way TJ and all, great work!

I put this up on the wiki:
http://www.redhax.net/wiki/Footage_Protocol_On_Set

I need to continue to rewrite it but I think we've done some really constructive work!

TJ and Ian: You did! Thanks for that!

 

[Cail Young]

I wish I was still shooting film. The problem with digital is the need to wear this stupid tinfoil hat

IBloom

Tape cans shut
Keep stock in fridge
Label EVERYTHING - including exposed film as 'EXPOSED' with the stock code even though it's encoded in the roll anyway.

I could go on. There are rites and rituals associated with shooting film, it's just after decades of them, they don't seem so odd any more.

 

[Colin Sheldon]

FYI The RED CF cards do not have a write protect tab on them (Which would be really nice!)

The other thing that I have found to be a good idea for RED and for P2 is to at least scan through the footage, or at least circle takes.

To me this is one of the real advantages of digital. I was able to catch soft focus, exposure issues, etc while there was still time to fix it.

I've also had P2 crap out in the transfer - the file was the correct size, but the footage was unusable - And the RED camera has that same potential. Watching at least some of the footage ensures that what you're backing up so carefully is actually usable.

Another thing to note about the RED is that when the camera formats a card it puts two small files on them. All you have to do is delete these files after you have copied the data and the camera will prompt you to format the card. That is how we ensured that the cards were empty. If the camera did not prompt, then the card was rejected and sent back to the DIT for rechecking.

Also, every time the camera rolls a new folder is created which includes the roll number. So on the first CF card all of the folders start with A001, the second A002, etc. This made it easy to determine if a card had been missed without resorting to time code.

 

[Clint Johnson]

FYI The RED CF cards do not have a write protect tab on them (Which would be really nice!)

That surprises me very much, I figured that switch would be a given. The 1st AC who removes the CF card is the only one who can write protect the card and the DIT is the only one who can write enable the CF card one they verify the transfer and formatting was good.

Without that little switch, the 1st AC would be relegated to doing something like throwing a piece of gaffers tape over the contact end of the card to make sure that it isn't accidentally over written before it is transfered.

But maybe I'm just paranoid.

 

[Colin Sheldon]

I agree, a way to mark the cards would be nice. The system we used, however, did work. If the camera did not request the card be formatted, it was sent back to the DIT. Also, the camera will not overwrite files, it will simply show up as "Media 57%" or whatever the case may be. If you choose to record on the card it will simply use the remaining space.

 

[tj williams]

Nice that the camera won't overwrite files. Clint your paranoia is a beeeeutiful thing. There is never enough paranoia!

 

[I Bloom]

FYI The RED CF cards do not have a write protect tab on them (Which would be really nice!)

I agree but I think it might not be neccessary. I'm not sure but I think RED is just rebranding CF cards that they know will work. Adding a write protect switch means more custom stuff which means added cost I'm assuming.

However if you look carefully at the workflow that TJ's outlined you can see it's not neccessary for the card to have that, since the only time a card is formatted is after it's been verified as completely copied. This creates kind of a one way street for footage.

On the other hand IMHO the write protect tab actually adds complexity and doesn't totally eliminate the possibility of human error. You can for example, forget to flip the switch, and a card will appear ready to format when it still has hot footage on it.

IBloom

 

[david farland]

TJ,
I'm still a little worried about the static procedures.
Say you're running round on synthetic surfaces (carpet etc) wearing synthetic clothing (non cotton/wool) with rubber soled shoes building up hundreds of kilovolts. If your camera is isolated ( not grounded to mother earth) and you swap CF cards in the camera you should be fine strap or no strap. Then you go over to your computer etc sitting on a static mat. Hopefully the next thing you do is clip your static strap onto the static mat which is electrically connected to the frame ground of the computer OR touch the computer frame. If you don't do all this there is a good chance you'll discharge those hundreds of kilovolts thru the CF card and thru the first piece of frame ground it touches (remember if you can feel a static discharge it's over 60K volts!). The best advice I can give is if you're doing DIT duties, don't wear synthetic clothing/shoes. Make sure after you move over static generating surfaces i.e. carpet or wriggle your arse on a synthetic (preferably leather) chair, you should touch the metal frame of your equipment before anything else! And realise if it's not earthed your won't discharge anything. Also note that long walks particularily with synthetic clothes/shoes over synthetic surface build up charge. It's as you learned in science rubbing different materials together is what builds up charge.

One last rant point not about static but keeping you alive! If you're gotta touch a piece of live (120/240V) equipment/metal....and you're not sure the frame is live or not, always touch it with the BACK of your RIGHT hand as the muscles in your hand will contract away from the live conductor and most of the charge 'hopefully' will pass down the right side of your body and not as much thru your heart.

Pardon the hijack!
Dave

 

[Brent J. Craig]

I don't believe a write protect switch is part of the CF standard as it is with SD cards. I have never seen CF with a switch, although that doesn't mean some enterprising manufacturer couldn't start adding them if enough of us requested it.

As for all this talk of grounding the camera, wearing wrist straps, etc - If the camera has a video-out line attached, and the video assist system isn't using any ground cheaters (which should be illegal on a properly-run set) then the camera is already grounded.

One thing we have done with P2 cards during the dry/staticy Canadian winters was to never pass them directly from person to person because of the potential for exchanging some free electrons (sparks). We either sealed them in their anti-static cases, or laid them on the table in front of the person doing the downloading. Probably a good practice for any flash media.

 

[I Bloom]

TJ,
I'm still a little worried about the static procedures.
Say you're running round on synthetic surfaces (carpet etc) wearing synthetic clothing (non cotton/wool) with rubber soled shoes building up hundreds of kilovolts. If your camera is isolated ( not grounded to mother earth) and you swap CF cards in the camera you should be fine strap or no strap. Then you go over to your computer etc sitting on a static mat. Hopefully the next thing you do is clip your static strap onto the static mat which is electrically connected to the frame ground of the computer OR touch the computer frame. If you don't do all this there is a good chance you'll discharge those hundreds of kilovolts thru the CF card and thru the first piece of frame ground it touches (remember if you can feel a static discharge it's over 60K volts!). The best advice I can give is if you're doing DIT duties, don't wear synthetic clothing/shoes. Make sure after you move over static generating surfaces i.e. carpet or wriggle your arse on a synthetic (preferably leather) chair, you should touch the metal frame of your equipment before anything else! And realise if it's not earthed your won't discharge anything. Also note that long walks particularily with synthetic clothes/shoes over synthetic surface build up charge. It's as you learned in science rubbing different materials together is what builds up charge.

One last rant point not about static but keeping you alive! If you're gotta touch a piece of live (120/240V) equipment/metal....and you're not sure the frame is live or not, always touch it with the BACK of your RIGHT hand as the muscles in your hand will contract away from the live conductor and most of the charge 'hopefully' will pass down the right side of your body and not as much thru your heart.

Pardon the hijack!
Dave

I don't think this is a hijack David, I think it's part of what we've been discussing. I think a lot of what you are suggesting can be solved by just touching the computer with your hand before you touch it with the CF card. Likewise on the camera. You could also touch them first with your tinfoil hat... that's my method

If the shell of the CF card was a conductor, rather than plastic, the problem could also be solved since the actual electronics would be inside a Faraday cage.

Has anyone actually fried a CF card with static though? I'm just looking for some real world examples.

IBloom

 

[Jack James]

This is all very interesting, there are some very good points made here. However, on the checksum front, there are a couple of very important things to remember:

1. Unless the checksum is created by the camera itself during the shoot it is almost a moot point. Generating a checksum on a laptop requires reading the data in the same way as making a copy, and will therefore not be an indication of a robust file. In other words, if the file is already damaged by the time it gets to the laptop, you won't know about it. The checksum only ever "guarantees" (if that can be said about it) data from the moment it's generated.

2. A checksum doesn't ever provide any way to actually recover the data if it's later found to be bad. An infinitely better approach is therefore to generate parity: http://www.surrealroad.com/research/archives/tag/par/