Adventures in the land of Grass Valley

Just back from a conference job in Rome – usual brief: a big conference has keynote presentations filmed, these need to be captured and edited down to their bare essence for viewing on the web.

Conferences at this level go on for days. Picture the scenario: a conference may last four days, with 25% of the time in ‘keynote’ mode: four cameras record a presentation given to 2,500 people in an auditorium, whilst presenters do their stuff either solo or in groups. For the other 75% of the time, the 2,500 delegates split into, maybe, 25 groups. Every hour, there are 25 presentations happening, and this lasts for three days, eight hours a day.

So let us leave aside the ‘breakout’ presentations and concentrate on the eight hours of keynote presentations, each of which includes a presenter or two, quite a lot of powerpoint slides, and maybe a video or a software demonstration.

Ideally, as soon as a presentation finishes, the highlights are made available, but that can’t happen.

The impression I am trying to create is of a machine that generates huge amounts of content which need digesting and cutting down before sharing on the web. How can a month’s worth of presentations be broken down to provide an accurate summary of the benefit of attending such an event?

Welcome to the world of conference video.

So the bottom line is that, in the old days, we’d record a vision mixed feed onto tape. Somebody who understood the content and its politics would sit with somebody who knew what timecode was and understood how important it was to note the in- and out- words of a good sound bite, would furiously concentrate on creating a shot list to hand over to an editor who would take the tape, shuttle through, pick the sound bites out and cut it all together.

But of course it takes time to shuttle through a 180 minute tape, and errors invariably crop up in timecode, or – worse still – you get a list of ‘he said something interesting after he talked about penguins’. That’s when you realise you’ll have to shuttle through a 60 minute presentation at double speed at least twice to find out what they’re talking about whilst you survive on a drip feed of espresso.

So wouldn’t it be nice if we ditched tape and went for a hard disk solution? Enter the Grass Valley Turbo – it records very high quality video to hard disk, and it can handle huge amounts. They have been cropping up on lots of the events I cover. You could run a television station using just two of these beasts. They can record and play back at the same time, you can shuffle playlists during playout, they are serious toys.

But for an editor, they are an expletive nightmare. Sure they record high quality, but it’s all at 8 megabit MPEG2 in GFX format, which means your Mac won’t play it in anything other than VLC. You can only play video from certain points, which may be minutes away from each other. You can only note down the approximate timecodes of the bit you want. Then you need to open up the whole GFX movie in something like Episode Pro, and convert the bits you want using the TimeCode in/out settings. Get it wrong, and Episode will have a hissy fit.

So you get the Turbo files into a format you can edit, and you realise that you need more than the bit you want. So rinse and repeat. Or the client wants to see more of it so you need to load up the GFX in VLC, and tell them to go get a coffee whilst you do the unmentionable.

The Turbo will convert to DV, but the exchange format is usually a USB disk, so you can get an 8 gig file in FX format, or a 24 gig file after a LENGTHY process in DV format off a turbo. So we get the MPEG files.

Quite frankly, I think I prefer the bloody tape.

So this time, we had a new toy: the AJA KiPro.

The Ki Pro is basically a tape deck without tape, recording to hard disks using Apple ProRes. The disks are special, in that they push into the deck like a tape, and when you pull them out, you find a little FireWIre 800 connector in the back which means you start editing straight away (it’s bus powered too). Or you can copy them off to your editing hard disk at FW800 speeds.

It was quick, direct, and easy – three words I DO NOT associate with Grass Valley Turbo.

It was even higher quality. The KiPro was set to up-rez the Standard Def it was fed via SDI to 720p, which it did marvellously.

The best bit is that a Ki Pro, even with lots of those special disks, costs a LOT less than an ultra-broadcast Grass Valley Turbo. It all happens at ProRes and FW800 rather than MPEG2 and USB.

I doubt Grass Valley see the Ki Pro as competition, but I’ll want one of these puppies for conference record in the future.

Gone in a flash

So Steve Jobs doesn’t like Flash.

Flash has always had a chorus of catcalls and boos from off-stage, way before Mr Jobs started his campaign. It dates back fifteen years ago, in fact: http://www.useit.com/alertbox/9512.html and http://www.useit.com/alertbox/20001029.html

Nevertheless, the reason why Flash became so popular in the Corporate video world was that Mac based video generators found WMV a hard format to publish in, and WMV wasn’t the nicest progressive download format around. QuickTime was a bit of a no-no at the time, with a 40 MB download and cumbersome install (from the viewpoint of conservative IT departments). Flash played nice on both Mac and PC, and was ‘as standard’ on corporate PCs.

Now… imagine a world where Microsoft adopted QuickTime (that’s never going to happen, but just imagine), would we be messing around with Flash? Sure, Flash works, but the playback is prone to stuttering and feels gritty in all but perfect playback environments. And even then, a dropped frame would never occur in the same place.

I used to use QuickTime for web based work. It was easy to integrate, provided smooth playback, looked great and worked well on the PC – so long as you installed QuickTime, which went from 7 MB to 42 MB (mandatory iTunes install) in the days before ubiquitous broadband. So QuickTime was out for client-facing stuff.

I adopted Flash, learned to like and to use flash, because the alternative was so unappealing (convincing Corporates, NGOs and the like to adopt QuickTime.

Well, hell’s closed for skiing and formation pig-flying:

http://blogs.msdn.com/ie/archive/2010/04/29/html5-video.aspx

We think H.264 is an excellent format. In its HTML5 support, IE9 will support playback of H.264” — Microsoft.

Flash gave lots of us video guys a solid foundation on getting video on the web as a reliable, easy standard that any website could benefit from.

Then Mr Jobs comes along, starts a war, and it’s out with Flash, in with HTML5 if you want to play in his little iGarden.

Don’t get me wrong – Flash is going to be around for some time yet. Many corporates do not use HTML5 compatible browsers, but give it a couple of years and Flash for video publishers will fade to black.

So it’s time to get good at H.264. For those of us publishing corporate video, we’ve got to get to know new settings, new wrinkles, new ‘chops’ that get even better results. New gamma, new keyframes. Maybe new software, or new plug-ins. New workflows.

And more importantly, new hardware. H.264 is not a quick codec to encode to. Whether it’s raw horsepower with an Octocore Mac, a mid-end solution like the Matrox MXO2 or an Elgato Turbo264 HD, we’ll need hardware help for a while yet. It’s not like encoding to the On2 codec!

And there’s a transition period. Remember, H.264 works in Flash now, and that’s pretty much the bleeding edge as corporate web video goes. The safe route has been On2’s Flash 8 codec, but I for one will be moving on to become H.264 based.

Until, of course, the next great codec comes along.