PärPod by Claude
PärPod by Claude
PärPod by Claude
Green Hippo in the Kona: The Afterlife of Broadcast Gear
19m · Oct 12, 2026
Green Hippo in the Kona: The Afterlife of Broadcast Gear

Green Hippo in the Kona: The Afterlife of Broadcast Gear

A Kona Full of Black Boxes

On a grey Monday in October, a Hyundai Kona left Kungsholmen in Stockholm loaded to the roof. In the back seat and the boot sat two aluminium Mac Pros, two black rack-mounted computers with a lime green hippo on the front, a pile of video cards with green stripes and round twist-lock connectors, reels of Ethernet cable, a Matrox box still in its packaging, and a tangle of DVI breakout cables that looked like a family of octopuses had given up. The total cost, bidding fee and VAT included, came to eight hundred and thirteen kronor.

The man from the TV studio who helped load it all said, more than once, that this was an insane deal. He also said the brand names with a particular tone of voice. Not the tone people use for computers. The tone people use for an old car they once drove across Europe, or a guitar amplifier that never let them down on a Saturday night. Datapath. Green Hippo. Blackmagic. Matrox.

This episode is about why people talk like that about boxes of electronics. It is about where this stuff came from, what problem it solved, why it lived inside a television studio for a decade, and what happens to professional gear when it stops being professional and starts being a pile in a Kona. It is also, near the end, a plan for tomorrow evening, when the boxes come out of the car in Kall and the question becomes whether you own a treasure, a radiator, or a very ambitious science project.

Let us start with the hippo, because the hippo is the strangest thing in the car. A Hippotizer is not a normal computer. It is a media server, and to understand why anyone needed a media server, you have to go back to a time when putting video on a stage was genuinely hard.

When Video on Stage Was a Separate Department

Picture a big concert or a television entertainment show around the year 2000. There is lighting, and there is video, and they are two different tribes. The lighting people have a lighting console, a large desk full of faders and buttons, and they program every cue of the show so that moving lights swing and change colour exactly on the beat. The video people have something entirely different. They have tape machines, or DVD players, or a computer running playback software, and a vision mixer, and projectors. When the chorus hits and the lighting designer wants the whole stage to turn red, the video operator has to press play on the right clip at the right moment, by hand, and hope.

This split caused endless friction. Lighting designers wanted video to behave like lighting. They wanted to say, at cue 47, the screen behind the drummer shows the fire clip, at seventy percent brightness, slowly rotating, tinted orange to match the lights. And they wanted that cue to fire every night, identically, triggered from the same desk that ran everything else.

The language lighting desks speak is called DMX. It is an old, simple, wonderfully stubborn protocol from the 1980s. A DMX universe is 512 channels, and each channel is just a number from 0 to 255. One channel might be the brightness of a light, another its colour, another which way it points. Every moving light on a rig is controlled this way. Later came Art-Net, which carries those same DMX numbers over ordinary Ethernet, which is one reason a TV studio throws away reels of network cable along with its video gear.

So the idea that changed everything was this. What if a computer full of video clips pretended to be a lighting fixture? Channel 1 picks the folder. Channel 2 picks the clip. Channel 3 is brightness. Channel 4 is rotation. Channel 5 is colour tint. Suddenly the lighting designer could program video from the lighting desk, exactly like a moving light. The video stops being a separate department and becomes part of the rig.

Around the turn of the millennium several people had this idea at roughly the same time. One of the early ones was Catalyst, created in the UK and later sold through the American lighting company High End Systems. It ran on Macs and let lighting desks control video as if it were a fixture. In London, a small company called Green Hippo built its own take on the idea, and called the product the Hippotizer. Other names followed over the next decade. Pandoras Box from Germany. Avolites Ai. And d3, later renamed Disguise, which made its big public entrance around the time of U2's enormous 360 tour at the end of the 2000s.

The category got a name that stuck. Media server. Not because it serves web pages, but because it serves media to the stage on command.

Version 3 and the Rise of the LED Wall

The Hippotizer went through several generations. What sits in the Kona carries a V3 sticker, which places it in the era roughly from the late 2000s into the early 2010s. The Windows XP Pro licence sticker on the lower unit fits that picture. These machines were built when XP was the safe, predictable, every-driver-works choice, and in the world of live production safe and predictable beats new and shiny every single time.

What made that era special was the LED wall. In the early 2000s, big video on stage mostly meant projection, and projection has problems. It needs darkness, it needs throw distance, it washes out under stage lighting. LED panels were the answer, modular tiles of tiny red, green and blue diodes that could be bolted together into a wall of any shape. At first they were expensive and coarse, with big gaps between pixels. Then, through the second half of the 2000s, prices dropped and pixels got closer together, and suddenly LED walls were everywhere. Concert tours, conferences, car launches, and television studios.

Television studios loved LED walls for a specific reason. A news set or an entertainment studio used to be built from painted flats and printed backdrops. Change the show, rebuild the set. With an LED wall behind the presenter, the set becomes software. The morning show can have a sunrise, the quiz show can have flashing graphics, the election special can have a map, all on the same physical wall, changed in seconds.

But an LED wall is a dumb display. Something has to feed it the right image, in the right resolution, with the right content mapped onto the right tiles. If the wall is an odd shape, a curve, an L, a ring around a stage, the content has to be warped and sliced to fit. And it must never, ever freeze during a live broadcast. That is the job the Hippotizer did. When you looked at that pile in the car and guessed it was gear for powering LED screens, you were right in spirit. Not power in the electrical sense, but in the sense that nothing would have appeared on the wall without it.

There is also a reason these boxes hung around in that studio long after newer versions existed. Broadcast engineering has a deep culture of leaving working things alone. A machine running a ten year old operating system, never connected to the internet, never updated, doing exactly the same job every day without a single failure, is not seen as outdated. It is seen as proven. The day you replace it is the day you introduce risk. So V3 Hippos, XP installs and all, stayed in racks well into the 2020s in plenty of places. They only leave when the whole studio changes, when there is a renovation, a move, or a decision to go to a new generation all at once. Which is, presumably, exactly what happened in Stockholm.

The Cards People Talk About With Love

Now the green-striped cards. Several of them are labelled things like composite, and analogue or digital input, and some have rows of BNC connectors, the round metal bayonet plugs that broadcast video has used for decades. These are almost certainly capture cards, the cards that let a media server take a live camera feed or a computer signal and put it into the mix, so a presenter's laptop or a studio camera could appear on the LED wall alongside the pre-made clips.

The green branding suggests they shipped with the Hippos, and Green Hippo is known to have worked with cards from Datapath, a British company based outside Derby. And this is where the affection comes in.

Datapath built a reputation that people in control rooms still talk about. Their capture cards and multi-screen display controllers became the default in places where failure is not allowed. Security control rooms, air traffic displays, broadcast galleries, stock trading floors. Their selling point was never flash. It was that the drivers worked, the cards ran for years, and when you rang them with a problem, someone who actually knew the hardware picked up. In a world full of consumer gear that dies after warranty, that kind of company earns loyalty that lasts for decades.

The same goes for the other names in the pile. Blackmagic Design, founded in Melbourne by Grant Petty around 2001, started with a capture card called DeckLink, and their whole story is about taking broadcast-grade video tools that used to cost as much as a car and selling them for the price of a weekend away. A lot of broadcast engineers built their careers on cheap Blackmagic cards, and they remember that. AJA, from Grass Valley in California, made the Kona line of capture cards, which means that if one of those is somewhere in the pile, you literally drove a Kona home in a Kona.

[laugh] Matrox, from Montreal, has been making graphics hardware since the 1970s. The box in the car is very likely a DualHead2Go or a TripleHead2Go, a small external box that takes one computer output and splits it across two or three screens, so the computer thinks it has one enormous display. Small, unglamorous, wildly useful, and still something people buy secondhand today.

What these companies have in common is that they make tools that disappear when they work. Nobody in the audience knows the card exists. The engineer knows. And because the engineer remembers every late night when something else failed and this didn't, they say the name with that particular warmth.

The Cheese Graters

Then there are the two silver towers. The aluminium Mac Pro, launched by Apple in 2006 and sold until the end of 2012, nicknamed the cheese grater because of its perforated front and back. It came out of the earlier Power Mac G5 design and kept the same industrial look, two big handles on top, feet underneath, a side panel that comes off with a single latch, and inside a beautifully organised layout of drive sleds, memory trays and PCI Express slots.

For video people, those slots were the whole point. The cheese grater was the last Mac for a long time that you could fill with cards. A capture card here, a fast graphics card there, a RAID controller, a fibre channel card for the shared storage. Final Cut Pro 7 era editing suites were built on exactly this machine, often with a Blackmagic or AJA card inside, and studios all over the world ran them for years.

Then in 2013 Apple replaced it with the cylindrical Mac Pro, the one people called the trash can, which had no internal expansion slots at all. For many post-production houses this was a genuine crisis. Their whole workflow depended on cards. So instead of upgrading, many simply kept their old cheese graters running, upgraded their memory, swapped in new graphics cards, and some of them carried on for nearly a decade. Apple eventually admitted the mistake and brought back a slotted tower in 2019, but by then the old silver machines had become a kind of legend.

There is an important detail for tomorrow. Not all cheese graters are equal. Look at the back for a model number. If it says A1289, the machine is from 2009 to 2012, and those are the ones people still want. The 2010 and 2012 versions in particular can take a lot of memory and fast processors, and an enthusiast community has kept them running newer versions of macOS than Apple ever intended. If it says A1186, it is from 2006 to 2008, and those are mostly nostalgia, beautiful to look at, slow to use.

The graphics card visible in the first Mac had DVI and DisplayPort outputs, which points more toward the later generation. Not a guarantee, but a hopeful sign.

What Old Professional Gear Is Actually Worth

Here is the odd economics of professional equipment. When it is new, it is extremely expensive, because it is sold to a small number of customers who need it to never fail and who are spending a client's or a broadcaster's budget. A media server with capture cards could easily cost as much as a new car when these were bought. Then it depreciates in a very particular way. Not smoothly, but in cliffs.

The first cliff comes when a new generation is released. The rental houses, the companies that supply gear to tours and events, move to the new version because their clients ask for it by name. Overnight, the old version is worth a fraction. The second cliff comes when the operating system or the software is no longer supported. And the third comes when the equipment ends up at a general auction, listed with blurry photos and no model numbers, where the bidders are mostly people looking for cheap office computers. At that point, it is worth whatever the scrap metal and the hope are worth.

That is how you get eight hundred and thirteen kronor for a car full of it.

But there is a fourth group of buyers, and they are the reason this kind of haul can be more valuable than the auction price suggests. There are small rental companies, hobbyist VJs, community theatres, churches, schools and festival crews who would love a working media server and cannot afford a new one. There are people repairing their own older Hippos who need a spare capture card. There are collectors of Apple hardware. Those buyers search with exact names. Hippotizer V3. Datapath VisionRGB. Mac Pro 2012. The auction never used those words, which is part of why the price stayed low.

There is one more thing that affects value, and it is the licence. Professional media server software often ties itself to the hardware, either through a dongle, a USB key that must be plugged in, or through a licence bound to the specific machine. A Hippo that boots up with its software and licence intact is a complete working system. A Hippo that has been wiped is just an old rack PC. The difference in value between those two states can be very large, which is why the first rule tomorrow is simple. Do not wipe anything until you know what you have.

Tomorrow Evening in Kall

So here is a plan for when the Kona is finally unloaded.

Start with safety and patience. These machines have been sitting in a studio, then an auction warehouse, then a cold car. Let them reach room temperature before you plug them in, so no condensation forms inside. Open the side panels first and look. Look for dust, for swollen capacitors on the motherboard, little cylinders with bulging tops, and for any loose cards rattling around. Take photos of everything before you move a single card, because the way the capture cards are arranged in a Hippo is part of how it was configured.

Then find the labels. Every card has a model number printed on it, usually on a white sticker near the bracket. Photograph them all. The label on the lower Hippo, which looks like it reads something like HIPPO VG 3015, is likely a serial or asset number, and it is the number Green Hippo support would want if you ever ask them about the unit. Look inside for a USB dongle as well. If there is a small key plugged into an internal or rear port, leave it exactly where it is.

Then power one Hippo on, connected to a screen and nothing else. No network. If it boots into Windows XP and the Hippotizer software opens, you have a working media server. You now own a piece of equipment that can drive multiple outputs, map video onto odd surfaces, take live inputs, and be controlled from a lighting desk.

And then the fun starts. One Hippo kept intact as a real media server opens up projection on the side of a barn during Västgård Game Fair, or a visuals package for PartyPär events. The second Hippo, once you know the first works, is a candidate for transplant surgery. Wipe it, install Linux or a newer Windows, and the capture cards might turn it into a multi-camera streaming box, the kind that could carry a byalag meeting, a local event, or an Årebladet live broadcast. Or it could drive a wall of screens showing a live dashboard of the whole municipality, weather and roads and whatever else, which would be the most Årebladet sentence ever spoken.

The Mac Pros, if they turn out to be A1289, make excellent quiet file servers, or simply very handsome machines to run old software on. And if any of the loose cards in the pile turn out to be graphics cards, check how much memory they have. Anything from that era with less than 8 gigabytes is not going to help with modern AI models. That is not a judgement, just arithmetic. The cards were built for pushing pixels to walls, not for running neural networks.

And the Ethernet reels? In a village of 120 people where a fibre connection is still something you wait for, cable is cable. They will find a job.

Why the Affection Matters

There is a reason this story is worth a whole episode and not just a shopping list. The tone of voice the studio man used, the affection for brands, is a kind of history. Every one of these companies, Green Hippo, Datapath, Blackmagic, Matrox, AJA, became loved because for years, in rooms where nothing was allowed to fail, their gear did not fail. Engineers remember the gear that saved them as clearly as some people remember a teacher.

When a studio clears out, that memory usually goes in a skip. Most of these things end up melted down for their copper and gold, and nobody ever turns them on again. This time some of it ended up in a Kona heading north, with a driver who wants to see whether a 2010 media server can still make a barn wall glow.

That is a better ending than most professional equipment gets. Drive gently. Wedge the heavy boxes low. And tomorrow, when the first Hippo boots up and the old XP start screen appears, take a moment, because somewhere in Stockholm an engineer would be pleased to know it still works.