Author Archive
The file-based problem
Posted by Dylan in General Ramblings on December 24th, 2008
File-based video formats offer huge productivity benefits in terms of faster-than-realtime ingest, increased metadata and data efficency, but there are also some huge drawbacks around the less tangible aspects of video production and post-production.
When we shoot on tape we always have the tape (as long as we look after it anyway) – we ingest what we need, at a lower resolution if necessary, and then put the tape on the shelf. If we’re short on discspace we can delete what we don’t need, and can always get it off tape again if we need it. If a drive dies we have the tape. Tapes are inexpensive, easy to source, easy to store and fairly easy to index and track. Files and memory card are much less so.
And as much as this is a problem in professional fim-making, it concerns me much more in the consumer market where flash-based and harddrive cameras have become increasingly common. While as professionals we can be reasonably expected to have the knowledge and skills to preserve our footage, the same can’t generally be said for consumers.
Are the next generation going to be able to rummage through boxes in the attic for the old home movies of them as kids? I doubt it. Think about it, have you got many or any of the files your treasured on your computer 10, 15 or 20 years ago? Of course, there are similar issues with consumer tape formats – can you still play Video8 tapes? Probably not, but at least if you have the tapes you can search for a player.
Are there any solutions being offered to consumers at large to make the memorys and moments they capture on their file-based camera anything more than temporary? I’m certainly not aware of any.
The moments we’re capturing now might not exist in any form at all within years unless we are especially careful to retain them for the future, and we take that responsibility upon ourselves. For professionals this is hard enough – have you taken reasonable steps to backup your P2 or XDCAM EX footage? Certainly for many the answer is no. And if we’re shooting file-based homevideos are we even half as careful with them? Would we expect our parents, siblings or friends to take the care and initiative to ensure their videos survive longer than their current computer?
When people ask me for a video camera recommendation now I don’t have an answer, but generally I’m still predisposed to something that shoots on a tape of some sort. It may be more limited in many ways, but there is some comfort in seeing a tape sitting on a shelf.
Title Safety
Within a PAL 16:9 video frame there are three distinct common title safety areas, they are:
16:9 Safe – basically the full width of the frame, suitable for 16:9 widescreen, and 16:9 letterbox viewing.
14:9 Safe – 14:9 is the middle ground between 4:3 and 16:9 – it is a slight crop, so the safe area is not the full width of the frame.
4:3 Safe – This means that graphics will remain fully visible after a 4:3 center-cut crop of the image. It is only around 65% of the frame width.
I’ve created a few PSD files that illustrate these safe areas (for PAL and HD frame sizes). They can be downloaded here framesizes.zip
These guides should conform to BBC and EBU recommendations. Other broadcasters may have their own specifications.
Labelling Tapes
Posted by Dylan in General Ramblings, Video/TV on October 29th, 2008
It’s boring and pretty easy to overlook, but when editing any video, at all, I strongly believe in the idea of serialised labeling. Nothing is more frustrating when revisiting an old project, or stealing footage from another job than to have the NLE ask you for “Tape 1″
Even for the small jobs, the one offs, I think it’s worth using simple and unique labelling that will enable you to find a tape again if you need it.
My guidelines for tape identification would be:
- Not related to content – a tape label reel id is just about finding the tape, it doesn’t need to tell you whats on it. You have clip names for that.
- Not repeated – this is the biggest one. You should be able to find the tape by knowing nothing more than the ID of the tape.
- Not too confusing – consider things like O and 0 and anything else that could lead to confusion later.
- Digitise nothing without an ID – stick tape ID labels on tapes before you capture them.
In choosing an ID scheme I would advise using only numbers and letters. Have at least 3 zero-padded digits in conjunction with letters, or at least four digits if using only numbers.
You could create prefixes you understand, perhaps a two-letter code for each job or client. The important thing if you go that way is to avoid repeating codes.
Go to the local stationery store and get the smallest label sheets you can find. Then using Word or OpenOffice or something make up sheets for of tape ID labels. Print them out and keep them handy. Whenever a new tape arrives, whack a label on it.
When you capture, use the tape ID either alone, or at the beginning of the tape name in the NLE. So rather than “Interviews GS0021″ make it either “GS0021″ or “GS0021 Interviews” – that will ensure the relavant portion of the ID survives any EDL or Metadata exports that might truncate it.
That’s it really. Just do whatever you can to avoid the dreaded “tape 1″ syndrome.
Setting up OpenFiler
Posted by Dylan in General Ramblings on September 4th, 2008
I’ve been setting up an OpenFiler server lately. I have mixed feelings, it is a slick and well put-together package, but if you go for the open-source version then you’re pretty much on your own. Most of the documentation is either available to commercial clients only or sold commercially. The open-source version has all the features, but none of the support or documentation.
So far here’s a few things I’ve had problems with…
My discs didn’t show up. They had previously been used in another system. It turns out that the backend script that pulls disc information got a little confused when it tried to read the partition information and didn’t like the setup of the disc, it threw an error and the OpenFiler GUI couldn’t see the disc information.
That was resolved by the ‘dd’ trick in this forum post: Hard Drive missing – Better error checking needed in list-disks.pl
The next problem (and one I’m just starting to tackle now) is LDAP authentication from the local LDAP server. If you’re like me and don’t really understand LDAP then none of it really makes sense, but this post is helping: Quick setup guide install with local ldap
Fingers crossed now, new fileserver should be setup soon.
RED Expectations
Posted by Dylan in High Definition, RED on July 30th, 2008
The following is the text of a post I made on a forum as a general response after someone was asking about how best to set up budget/low-end 2K monitoring and finishing for the FCP system..
I reckon Jim Jannard and RED have a lot to answer for when it comes to messing up expectations.
I love the RED One camera, I have now worked on set and in post on a number of RED shoots (all destined for no greater than 1080 in the end). It’s a camera for $18,000 (we’ll ignore lenses and accessories) that can shoot 4K!
Every indie filmmaker in the country is now convinced they can cut their opus in 4K in FCP and have it projected at the local Multiplex for less than it costs to sponsor a kid in Africa.
They are very mistaken. Yes, they can shoot RED, in 2K, 3K or even 4K. Yes they can cut it in FCP (or Avid, or Premiere, or whatever if they are game to play around). But the 2K or 4K finish, there’s the rub. FCP can, in theory, work with 2K resolutions. But it’s far from simple at that point, it’s not like video. You can’t just spit it out. There are huge issues to consider about LUTs, delivery, compression, audio mastering.
My advice to anyone thinking they ought to get into 2K on the cheap is, don’t. Wait. Stop. Think about it. FCP and Color and all the rest are good tools, but this isn’t really their strongest area, and chances are it’s not yours either. There are product and people who really know what they are doing and do it well. Those people and products are what you need. Shoot 4K, edit the proxies in FCP or Avid or whatever, but when it comes to the finishing, if you want any bigger than 1080 then it’s time to find a friendly expert. Hire/beg some time on a Scratch or Pablo or Smoke with someone who really knows their stuff. The future nightmares will be reduced and the quality of the end product will be increased.
2K will come to your desktop in time, but wait a while.