Jun 27 2005
Recent experiments by me with a Sony digital video camera and Adobe Premiere have pushed me into an obsession with the production of web-ready video in the expectation that video will soon go the way of audio and follow an RSS-path towards something akin to podcasting called vidcasting.
With high hopes set for devices such as Sony's PSP to finally build a platform for portable video, I give it a year before vidcasting begins to surface as a meme.
The question is one of gnarliness -- podcasting can be as time-consuming as one wants to make it depending on relative degrees of obsessive compulsiveness and the complexity of the production tools, but I found the learning curve on opensource products such as Audacity signficantly reduce the turnaround time to about a four-to-one ratio of post-production to capture time.
Video is a different matter. Thanks to the speed of a firewire connection, D/V can be sucked off of the camera and into the PC very easily. Editing tools I've been playing with are Adobe Premiere -- which is way too feature-laden for my uneducated tastes (this bed is too hard), the video editor that is bundled with Windows XP (this bed is too soft) and Adobe Premier Elements which fits the bill nicely thanks to its ability to output onto recordable DVDs (and this bed is just right).
I recorded some rowing races earlier in the month and due to the keen interest by the rowers and their parents have had to find some time on the last two weekends to edit the raw footage, title it, and encode it for web viewing off of Churbuck.com in .wmv format. This past weekend I encoded the files into one 25 minute flick, complete with a DVD menu system, and burned the results onto discs for distribution to parents.
It was, all in all, a good experience and Premier Elements was adequate for my purposes. I now need to read a good book on videography to teach myself the do's and don'ts of over zooming, panning, using a monopod to cut down on handshaking, and what to do to override autofocus so the lense doesn't autoseek on something in the foreground when I'm trying to capture the background. I'd put the production to capture ratio at roughly eight-to-one, mainly because I was trying to figure out hairy add-ons like scrolling credits that really aren't necessary.
Battelle's scoop last week that Google was launching in-browser video playback today, drove me to Google's existing video submission service, where I opened an account under my gmail name and uploaded one of the rowing clips. It was an easy process and the file is currently under review before being posted. I was asked by Google if I wanted to charge users for viewing (which I did not), so evidently there is a market to be made. Now my concern is hosing my bandwidth allotment for Churbuck.com. Once the clip is approved I'll be interested to see how it is tagged for search finds as there was no tagging facility offered in the upload process. Google, according to Om, is using an opensource player, VLC, which I will download and checkout as part of some webcasting due-diligence I'm now performing for CXO.
Key insight learned from these recent video experiments: the size of the image plays a huge role in helping the viewer assign sounds -- conversation -- with people. The smaller the image, the more disconnected the audio. Amazingly stupid insight, but nevertheless something I'll keep in mind when building future web videos displaying in small apertures -- keep the number of people to a minimum, the images are too small for the user to assign voices to faces.
Google Video either turns into America's Dumbest Digital Videos or really gets some quality and makes some people some money. Wonder what Google's cut of the action is? Google Video deserves some attention.