Every tape knows what's on it.
Footage arrives and is watched, listened to and read before anyone asks for it — who appears, what is said, what happens, what it can be used to prove. Your archive stops being a folder of filenames.
One tape, read three ways at once. Open any part to see what that pass contributes.
Most newsroom archives are searchable only by whatever somebody typed into a filename twelve years ago. The material is all there — the interview, the aftermath, the arrival at the airport — but finding it depends on remembering it, which in practice means it depends on one long-serving colleague who is on leave.
Video Studio watches everything as it arrives and writes down what it saw. Not a filename and a date — a description of who is in the frame, what they said, where it appears to be, and which of your editorial categories it belongs to. The point is not the cataloguing itself. The point is that once the archive is described, every other studio can reach into it: the script can ask for pictures, the clip can ask for context, the researcher can ask a question in plain language and get shots back rather than folders.
12 things it does. Open any one for how it works.
Every word of speech in a tape is transcribed with timings, so a phrase you half-remember becomes a way of finding the shot it came from.
A forty-minute rush is broken into its individual takes, and takes that belong together are grouped into scenes — the press conference, the arrival, the walk to the car.
This is what makes the archive usable rather than merely indexed. Handing a producer a forty-minute file and a timecode is not much better than handing them the tape; handing them the eleven-second shot they wanted is a different kind of help.
Each shot is described in words — a crowd outside a courthouse, a damaged building, a signing ceremony — so footage is findable even when the soundtrack never mentions it.
Faces are recognised against your own gallery of people who matter to your coverage. Anyone unfamiliar is grouped as an unnamed recurring face until somebody names them once.
Naming a face once names it everywhere it has ever appeared, backwards through the archive as well as forwards. A researcher who identifies a spokesman in one clip has just labelled every previous appearance of that spokesman.
Name strips, placards, banners and captions burned into the picture are read and become searchable, which is often the only record of who a speaker was.
The editorial summary a librarian would have written, written for every tape: who is doing what, when, and whether a passage is a speech, an interview, a soundbite or general coverage.
It is what lets a script ask for a soundbite and get one, rather than anything where the right person appears — including the shot where they are walking silently to a car.
Footage is filed into the territories your newsroom actually thinks in — politics, conflict, economy, sport, environment and the rest. A tape can sit in up to three, because real stories rarely sit in one.
When the filing is wrong, an editor fixes it — adding a category, removing one, replacing the lot. The change is recorded with who made it and when.
Not everyone should be able to reclassify the archive, and not every correction should be silent. Both are permissions, and both leave a trail, because an archive without a history of its own edits is an archive nobody can audit.
Describe what you are looking for the way you would describe it to a colleague, rather than guessing which words are in the title. You are describing what you want, not guessing at filenames.
Recent material is reached for first and older material stays fully searchable behind it, so a busy day's search is fast without the deep archive becoming invisible.
Publishing a piece of footage on its own comes with a title, a description and tags drafted in Arabic, so a clip going to the site is not filed under its camera number.
Anything found can be opened in the cutting room as it stands, or sent to a script as a matched shot. Nothing is exported, downloaded or copied to get there.
From your cameras, your feeds, your contributors or an upload. It does not matter which; everything goes through the same door.
Speech transcribed, shots separated, faces recognised, picture described, text on screen read. This happens without anybody asking for it.
The contents card is written and the tape is filed into the editorial realms it belongs to. It is now findable by anyone in the building.
They describe what they want in a sentence and get shots back — not folders, not a list of files to open one by one.
A researcher names an unfamiliar face or fixes a category, and the correction applies backwards across everything that person or subject ever appeared in.
You need everything you have ever shot of them, and you need it in twenty minutes. Because faces were recognised on arrival, that is a single search rather than four people opening folders — including the material from before they were well known, filed years ago under an event nobody would now think to look up.
A reconstruction story needs the before and after. Describing the picture rather than the file means you can search for what a place looked like, and find the shot from the coverage of an unrelated story that happened to walk past it.
A viewer disputes how a speaker was represented. The contents card records who spoke, what they said and at which timecode — and the record of who filed the tape and who has changed its classification since is attached to it.
Who matters to your coverage is your decision. Name someone once and the archive knows them from then on, forwards and backwards.
Add a category, remove one, or replace them entirely. Who is allowed to do that is a permission you set.
Which material is reached for first and which sits in the deep archive is an editorial call, not a date calculation.
Working copies are cleaned up on a schedule; original material is not touched. Retention of the source footage is your policy to set.
An unrecognised face is presented as an unnamed recurring person, not as its closest match. A wrong name in an archive propagates into every package built from it.
Once a person has corrected a classification, later processing does not quietly revert it. Human judgement is the higher authority in the record.
Every classification can be opened to see what drove it, and every change to it carries a name and a time.
Describing footage makes it findable inside the building. Nothing leaves the building except when someone sends it.
An archive described only when somebody needs something is an archive that is never described, because the moment somebody needs something is the moment nobody has twenty minutes. Here the work is already done before anyone asks — quietly, once, on material nobody is waiting for — and the day the story breaks, the search is instant.
It runs in the background at whatever pace the hardware allows, and new material is always prioritised over the backlog so today's work is never waiting behind 2019. Most newsrooms find the recent years are the ones they actually search, and those come first.
The watching and describing happen on your own installation. The archive stays where you put it.
An editor corrects it in a couple of clicks and the correction holds — later processing will not quietly revert it. The change is recorded, so a pattern of the same correction is itself useful information.
Yes — mute rushes are described from the picture, the faces in them and any text visible on screen. Silent material is often the most valuable thing in an archive and the hardest to find by conventional means.
The useful demonstration is the one that runs on your footage and your wire feeds. Tell us how your desk works and we will show you the parts that matter — and say so where we do not help.