How little things with Final Cut Server ruin your day

The day was going along pretty well today until I decided to do a small tweak on FCSvr.

I'm freeing up room on our expensive drives by archiving to cheap drives.

That was going just fine. What happens is it copies the file from the main media drives, on to the archive device, checks that it is there, then deletes the file from the main media drive.

Worked like a charm until inexplicably it started refusing to delete from the main drive.  That's weird I thought. But then, considering who uploaded the video, I thought to myself "weird stuff always comes from his computer...probably his fault" 

So as a nascent server admin I immediately blamed a user. It felt really good to do that. Unfortunately it didn't help me.

So I updated the xserve to the lates pro-app and OS upgrade. The server hung during the update, so I pulled the power on it and plugged it back in. It's one of those moments where you pause and think to yourself, this will either work, or I won't see much of my family this week.

It worked. Still couldn't delete though.

Finally I called up the folders on the media drive and compared them to others and sure enough, the permissions weren't set right.

FCSvr is an admin. admin did not have read/write permission on those rogue folders.  Problem solved. Change permissions and it works like a dream. Even better I still get to blame the user because they created the folders in the first place so it is probably their fault.

 Score.

I used to just work on story

That used to be my gig, having good creative ideas and making them into entertaining stories. Now I do that, and I'm a server admin. Go figure. I'm about the most dangerous server admin you can get. 

Updates without backup....check.

Just pushing the power button when hung...check

Blindly entering commands in terminal just because some website tells me so...check

Working with signage software

Even the name itself "signage software" should be enough to scare you away. "signage" ... what a terrible word.

The software is made to sell people things using moving pictures in waiting areas. Waiting for a bus, waiting for a plane, waiting to buy a bag of chips and a slurpee. The software is about as mundane and inspired as those places.

If you want a nice stock ticker (because that's what busy people demand) and a brightly coloured billboard suggesting the soup and sandwich special at the "transportation station snack bar" then this software for you.

This software is decidedly not for public galleries driven by an education and conservation mandate using HD footage.

It works but barely. Whatever you want to do, it will do, but it will make you work for it. Of all the things I want to time on, making presentation software work is not one of them.

Right now we have a touchscreen and a plasma running off of one computer. To get the touchscreen to update the plasma is pretty much impossible. I can do it in the omnivex (yeah I'm calling you out Omnivex) architecture but it's going to make me work for it. Basically it will take about 5-6 hours to set up, then every change thereafter will take about an hour. I don't want that. I want to be able to do it quickly. I want to be able to update often. 

I want the ease of putting a 7 day weather report in there. Thankfully the software makes weather reports easy, unfortunately our cichlid exhibit is not currently in need of weather reports.

Threads n' granules

Sounds like a great breakfast cereal but it's part of a strategy document I'm working on. It's all about multi-platform media development. I think a lot of places get this whole multi-platform thing wrong. They think that it means develop one item for multiple platforms. I think you have to develop many different items for many different platforms. It sounds onerous but if you build them all from the same source, ye olde threads n' granules, then you are okay.

Here's part of what I'm writing...

Traditional approach to multi-platform meant adapting TV for the web i.e.—distribution of “television-type” video content through media such as the web or gallery screens. The DCL is an evolution of that original idea that goes beyond by recognizing the variety of platforms available to us, the uniqueness of these platforms, and delivering custom content to each of those platforms. The custom content is not a force fit of one format into a different format; rather it is the ground up creation of specific content for specific platforms. Blog content is different from Youtube content, which is different from exhibit content and is different from media b-roll content but they all evolve from the same source media granules and messaging threads. The outputs are multiplied but the work required to feed those outputs is not correspondingly increased because the work is done in parallel.

So far Drobo = Love

We're out of room on our xraid. Too many media files. A new promise vtrack array will cost us somewhere in the area of 16k.
I bought a little drobo with 4TB and droboshare for about $1600. The thought is we can delay our array purchase for another year if this works.

The plan is to use the drobo as an archive device on Final Cut Server. Media that we don't access all that often will get archived to the drobo. We can still search for it in FCSvr and view the proxies, it's only when we retrieve it will it take a bit longer.

I haven't tried it yet because I need IS to come up with a backup strategy in case something goes terribly wrong.

The drobo, which is fully protected storage, has got to be one of the easiest installs I've ever done of any hardware. Seriously. No tools. Just slide the drives in, install the dashboard and away it goes.

It's actually living on our network quite happily without any problems. So far I'm impressed.

We're going to run it for a bit before we formally attach it to FCSvr.

What is the experience

Right now we do this in the big sense. We ask of our galleries what the big experience is. In each exhibit we kind of ask what the experience is. I don't know if we go all the way though. We think about the hook, the core messaging, and what "walk-away" success would be, but I don't know if we ever talk about the experience at that detail level.
If someone plays with one of our interactives, or looks at our video wall, what do they experience?
I can tell you what I hope they walk away with. Sometimes it's success if they walk away with even the lightest grasp on the challenges facing a certain animal. If it is a multi-layered message it can be great if they pick up any of the layers, as long as they get at least one.

But when it comes to the experience they had at the spin-browser, or the experience they had the video-wall or the experience they had at one of the habitats, I don't know if we really put it all together to consider that. I think we should.

How to interview now with more text

Alternate sides of the camera you ask the questions from. Just keep switching sides every time you do an interview. When you edit you'll be able to cut between people on different sides of the frame instead of everyone stuck to the same side.

Ask open ended questions for interesting responses.

Do not ask any question that can be answered with a yes or no.

Train the people to restate the question or answer in a complete sentence. Instead of "yeah Dave totally does that" you'll get "Dave is totally the guy that screws around on his girlfriends". Much better.

Tape off or turn off your tally light so they don't know when you are recording.

Wrap the interview, tell them you are done, physically relax then compliment them on something they said or say "i though you were going to talk about x, thank god you didn't". The trick is, the camera is still recording.
Then they start talking in a super casual manner about all sorts of stuff in an amazingly relaxed tone. Some of the best clips ever come from that trick.

Also just start rolling at the beginning, before you say lets get started. It's not so much about tricking people into saying bad things, as it is about tricking them into being relaxed and themselves in front of the camera.

This is a stupid one, but listen. It's amazing how many don't listen, don't engage in a real conversation, and don't ask the follow-up question.

People often hint at stories that they won't volunteer on their own, but will willingly give up if you ask them directly.

Write your questions down. You don't need to stick to them, but at least you know where you came from and where you are going.

Look them in the eye and be interested.

Don't ever interrupt them.

Let silence hang. Often they will jump in and fill the silence in an effort to make things less uncomfortable.

Nod while talking to them, but don't say anything unless you need to. Their voice is important, not yours.

Always encourage them that they are giving the best interview ever.

If you aren't getting what you need, and you are trying to get something out of them, and you've been at it for awhile, and they are getting nervous, tell them to relax. They already gave you some great answers which now gives you the luxury of getting picky and more in depth. Lies, but it will make them feel comfortable.

There you go. That's over 1000 interviews for camera's worth of experience for you in a lowly, less travelled SA forum.

Too/to bold?

When to use bold and when not to, it's this afternoon's office discussion.
We've just finished new screens for the beluga underwater and we have a nifty new design. It looks good. However, for design sake several words of each phrase are bold, and several are normal font.
It looks good, but the discussion is, why are we bolding those words?
In a full sentence we don't do that because if you need to bold a word to call attention to it, maybe you should just rewrite your sentence to present the concept better instead. If one word is more important than others, maybe you should lose the other words.

However in a phrase like "Grey patch behind dorsal" is it bad if you bold "grey patch"?
It looks good and the non-bold words are still important so we can't get rid of them but do we break our own rules?
When does design trump content? Is that a neutral action? Does design only hurt or help content?
Do we overthink this kind of stuff? All valid questions.

I NEED MORE SPACE

12 Terabytes for only 20k what a deal. I think it is going to be a hard item to get through the budget this year. Buying storage is not sexy. 

We're 500 GB from using up our 10TB of storage. That means we need more room. I'm sure we have a bunch of crap on there, but really we chew up space at the rate of a gig a minute for footage. We need big chunks free, not just a file here and there.

I figure we can keep ourselves to 2.5 TB stored footage per year. That should give a lifespan for the 12 TB array of 3 years or so (RAID 5). So I'm going to have to make sure people understand the yearly cost is about 6K. They are still not going to like that number so I'll also need to explain about falling drive costs. 

With any luck that will convince people. In all fairness I did say that storage costs would be an ongoing expense.

The workflow of it all

Final Cut Server really is an amazing piece of software. It remarkably customizable and though complex, even a hack like myself can start to move around it pretty easily.

That said, the real complexity is not even in the software. The real complexity lies in your workflow and the personalities that drive it.

Right now I'm trying to figure out what our workflow is (it's pretty much all over the map) and what it should be.

The problem with an awesome, strategic, comprehensive workflow is people. People actually have to do the things you want them to. They have to enter the right data, in the right spots in a consistent manner.

That is the tricky bit and that is human nature. You have to design a path that will actually be walked. If you do anything else, and you kid yourself that you'll make people stay on the path, you're in for a surprise. You can't make anyone stay on the path, they have to decide to stay on the path.

We had a huge control vocabulary that would bring our cataloguing up to library standards. I slashed and burned the thing. There is no way anyone would ever go to that level of detail when uploading data. They'd go around it and we'd have a totally flawed system.

So I've simplified. Hopefully the simplified system will work out. It doesn't ask for much effort, just consistent effort. We'll see how well that works.

After all the system to have all our stories have a simple 3-4 word title that is consistent from raw footage to final delivery, is hardly a success, and that is about as easy as you get.

quit emailing me

So I thought I'd be smart and have FCSvr email me whenever someone submitted a master for review. It worked. Wooo.

Except that FCSvr emails me about 10 times for every one piece that is submitted. It's very excitable this Final Cut Server.

It appears that once I have FCSvr email me about one status change, it emails about every status change. There is no email once and stop.

Update breaks final cut server

Well, go figure, the update of awesomeness that is supposed to fix everything no problems ended up breaking things. Go figure.

Yesterday I submitted a project for upload just as I was leaving, thinking it would run overnight. Welp, no such luck.

FCSvr stopped creating video proxies for the uploaded footage and left me in the lurch. What good is having video in there if you can't view a preview.

The old e_sum_param error which no one has ever heard of didn't really help me troubleshoot compressor.

That was last night and the reason I was late for dinner.

However this morning, with a clear head and a stick-it note left by me on monitor saying "check server" I checked the server and found the problem.

During the update it changed the render cluster to nothing. Nothing was selected for that. This morning I changed it to "this computer" and just like that, proxy heaven.

Now I don't have any idea what I really told it to do. I think I said rather than sitting there wondering where you should be rendering this, go ahead and use your own cpu and get 'er done.

A lifetime of closed fist button mashing has prepared me well for my FCSvr admin duties...Just keep pressing buttons and restarting things until it works or you've really, really, really broken it.

But wait there is more Final Cut Server craziness

At this point I feel as though I'm accelerating along my descent into madness that Pro Training book so eagerly embraces.

Another word for word quote from the Advanced Administration chapter:

"With greater control comes greater complexity, so it's best to caffeinate heavily before diving into these pages"

Good lord. What am I in for? I do like the honesty of this book though. No messing around pretending everything is going to be okay. It states it pretty clearly that you are on a likely suicide mission, but if by some chance you survive there will be great reward.

I consider myself warned.

Final Cut Server is hilarious

Clearly someone was losing their mind in the course of writing the Pro Training book for Final Cut Server.

I'm dutifully going along, reading the book, trying to figure out what the hell I'm doing and how I'm going to deliver what I promised when I come across this word for word gem...

"Now that users can log in, and assets are beginning to show up in the client application, you can take advantage of all the distracted oohing and aahing to knuckle down and design the true gems of your already thankless job: automations."

I don't know what it seems like to you, but to me it sounds like that giddiness that starts to creep in when over-tired and facing punishing deadlines.

Content Inside

Content, content, content. Lots of content. Just spray it around, fill some buckets with it and do it fast. 

It's a word that is thrown around so casually. "We'll just put the content in there" or "We need some content".   Many people look at content like it is a coat of paint. You build the thing, then just decide on the paint and slap it up. Content doesn't work like that, as much as people think it does.

I had someone upset with the time it was taking to come up with the right message for something. Their response to us taking so long was to grab some facts off wikipedia and slap them together...done. 

It's hard to get someone to recognize what it is you do when basically many consider the job a matter of finding and replacing all the "Lorus ipsum" with some of them english language words.