A few years ago, the talk was all about tapeless production and server-based workflows. New installations were designed around the ubiquitous "server". This server, of course, was just a friendly name we used so as not to frighten non-technical users. The real "server" was never a single unit, but typically a large RAID array configured as a Storage Area Network and accesed by several Ingest, Playout and FTP servers and a whole corridor of on-line Edit Platforms.
The technical problems in designing and maintaining these server-based workflows were considerable. So much so that the focus was all on getting it to work and keeping it working. As a consequence of this, archiving was often little more than an afterthought: we'd better have somewhere to dump all this stuff - let's hang a digital tape library onto the SAN.
In other words, the "Archive" was often treated more as a solution to a Media Management problem (how to stop the on-line store form filling up) than as an end in itself. But this has changed, even if some would-be system integrators haven't yet noticed. The Archive, and not the "server", is now, or should be, the heart of the system.
Let's look at some figures. Suppose we set up a News channel. Our on-line storage might be about 500 hours. Typically, we might ingest 100 hours per day from agencies and delete this raw material on a two day purge cycle. That means 200 hours of server time is always owned by ingest. Our typical edited (conformed) clip is around 3 minutes. That's 20 clips per hour. We have 300 hours available storage. Better to call that 250 because we don't want to fill the storage. So, we could have up to 20 x 250 = 5000 on-line clips (including stings, bumpers, intersticials of all kinds).
On Launch Day, we have around 5,000 on-line items and maybe about the same again in our Archive. One year down the line, we still have around 5,000 on-line items, but maybe 50,000 archived clips. And after two years, maybe 100,000 archived clips. Hardware is just hardware; it depreciates from the day you buy it. But the content of your Archive is unique, irreplaceable and growing in value by the day. To put it bluntly, it doesn't take the brains of a wooden rocking horse to see that the Archive, and not the "Server" is the true heart of your station. My example was News, but the argument applies equally to general programming, sports, documentary and probably any genre you care to name.
Q - what's the difference between an Archive and a handbag?
A - you can find what you're looking for in an Archive.
Or can you? The chances are, your clip naming convention (clip metadata) is adequate for managing the 5,000 or so on-line clips, but typical clip metadata consists of little more than Clipname (32 characters max, including the date-stamp), Start of Message (timecode of 1st frame) and Duration. You have no chance of finding a 2-year-old News story among 200,000 items unless you already know the precise date. And of course, unless you know for sure that the item exists.
The Archive, then, if it is to be of any value, cannot just be a vast repository of media files, not simply a deep handbag. The Archive has two components - data (media) and metadata (the index). And of the two, the more important is the metadata. Most people, hearing that for the first time, don't accept it. Surely the media is more important? In practical terms, no, it's not. Imagine walking into a public library where all the books are shelved in random order and with their spines to the wall. And there's no catalogue. Oh, and you've got 45 minutes to research data for an obituary of X__ who's just popped it. Now imagine you have the same assignment, no access to any real books, but an on-line Amazon catalogue complete with précis, reviews and even thumbnail pictures. That's metadata, and metadata is king. Of course, you need both, but metadata is the window on your media. Without good metadata your total media assets amount to not a lot.
So, what is a Digital Media Archive? It is a database describing every media asset (piece of media!) on your system in exactly the terms required by your particular business model. The single most important field in the database is the precise location (URL) of the media. Other vital fields, in no particular order, could be: title (ID), duration, creation date, source, shot location, reporter, event, on air date, etc. The list is long and entirely customised for every installation. There is no one-size-fits-all. There may also be links to thumbnails, storyboards and/or browse quality instances of the entire asset. But for the simple answer to the question, what is a Digital Media Archive?, - it is a database.
Unfortunately, people often run scared from following through an Archive project. They get bogged down in difficult decisions, usually about physical format (data tape, HDD, DVD) or file format (MPEG2, DVCPro, etc). But this is to miss the point completely. It is, in fact, an old-fashioned media-centric view. Yes, it is important to be future-proof by archiving in a file format at the highest data-rate you can afford, but that doesn't come first. First, you design and install your database with all the metadata fields you are going to need. Then you start to populate it by building it into your workflow, so that new media assets are 'automatically' indexed. Once that's in place, you can start to catalogue your existing legacy assets, preferably in an 'as needed' order. Above, we said that the most important field is precise location. There's nothing wrong with an entry here that reads "Room 214, Cabinet 17, Shelf 12, Slot 23, MANDELBROT, SOM 10011506". That is every bit as accurate as "DIVA:\news\lon\2006\KBR_17sep_MANDELBROT.mpg". It's simply recognising the fact that you haven't yet got around to ingesting that item.
From the day you install and start to populate a Digital Media Archive database, you start to reap benefits. At your leisure (joke!) you can start on the parallel process of digitising your legacy media. But the enormity of that task should not deter you from implementing the Media Archive Database. The two projects need not be related, except by a field called "Precise Location" which you update when the time comes to ingest your old tape.
This article introduced Digital Media Archiving in generic terms. However, every facility is different, with unique archiving, indexing and retrieval requirements. TriMedia can help you to see your way clearly through all the options and arrive at the solution you need. Why not ask?
