MentalX

We just installed a Flash Media Server at the office. After some introductory issues with getting the server installed, and a few getting it to play nicely with the flvPlayback component, it is now working like a champ. Here is a rundown of how we handled a few things.

First, the skin. We are using the flvPlayback component with a custom designed and developed skin. I opted to design an embedded skin by using the following components: Buffering Bar, Pause Button, Play Button, Seek Bar, and Volume Bar. I also coded a timer that keeps track of hours, minutes, and seconds for both current and total time. These are called by a function, set a variable, and are displayed on stage as dynamic text. The flv itself is loaded through a variable set in the HTML code that calls the swf itself. So, we basically have several different swf sizes for the video player: 320×240, 480×320, 480×360. We also have three sets of corresponding code snips. To use, the flv is uploaded to the FMS, the appropriate swf is dragged into the swf folder at the root of our course sites, the code is placed in the appropriate location within the HTML, and the variable in that code is changed to the rmtp:// path of the flv file. That’s it. Drag and drop custom flv player.

On the server-side, I have created an application for each course that uses the streaming server. It provides a way to distinguish videos from course to course, and seemed more intuitive than have one application with sub-folders. I made one change to the adapter.xml file to outline the port sequence. And added the main.asc file to the root of each of the applications. The logs are fairly extensive, but don’t provide a whole bunch of useful information for us from a technical support standpoint. They will be used primarily as reference, while our standard web server logs (which contain username information) will be our first line of defense. Speaking of the web server, it is being run on a second machine. Both sun boxes. One Solaris, One Linux.

That’s about it. Totally easy. I will write an update in a few weeks after we start pushing FMS a bit.

For as long as I have been working at HerbergerOnline, we have always used audio lectures for our courses. The majority of the classes the utilized audio when I joined the crew were music-based: Bealtes, Pop Rock, Elivs, and the like. It wasn’t until later that we considered expanding the audio lectures into the realm of Art. The biggest problem with art is that you really need to see what piece of art the instructor for the lecture to have impact. My solution was basically to create a digital slideshow, or Downtempo Documentary as I like to call them.

When I started my research into the technological options, I first started with Flash, as I figured it give me the best chance to create something scalable that could changed or updated by simply editing a txt or XML file. While I did create a working model using cue points, I realized it would never be able to offer me the portability that I desired for my users. In other words, the end-user wouldn’t be able to load this thing up on an Ipod and take it for a ride. HerbergerOnline has been built on several ideals, one of them being that an online class (especially one that needs to function on a large scale) should never be tied to a time or location. This flexibility to learn whenever and wherever is part of our success as a whole. With this in mind, I dumped the Flash model and decided to take a more traditional media approach in order to export the videos as both .mov and .mp4 files.

The process for production breaks down like this: The instructor records the audio lectures on an Edirol R-4 with a Rode NTK microphone. During the recording process, the instructor makes a verbal notes indicating where a slide needs to be changed. The audio files are pulled from the R-4 in WAV format and edited in DSP Quattro or Logic Pro (I generally use Quattro for this process as it is much quicker at redrawing the waveform — Logic is pathetically slow at this). During the editing process I create documentation of insert points for each denoted slide change. Generally we receive the images in an easy to follow Powerpoint. Once the audio has all been edited, they are brought into Logic to be bounced to an aif file with effects applied. The audio is now ready for use.

The powerpoints are then painstakingly converted to nice looking 480 x 360 slide in Photoshop and batch-exported as jpgs. These are then imported into Final Cut Pro and paired with the audio. The finished product is exported from Final Cut Pro to a Quicktime .mov file. These are then exported through Quicktime with the appropriate compression: I use H.264 compression, quality of Medium, 15fps (automatic key frames), IMA 4:1 / 44.1 / mono for audio. This seems to be perfect, and (depending on the number of key frames needed) will generate an 8-9 minute .mov file that is always less than 15MB. The original .mov files are next exported as .mp4 files for Ipod playback. These are very good quality and readable if you keep you font size at 10pt or above depending on the font you choose to use.

Lastly, pdf files are made from jpg images using InDesign, and the audio files are compressed to MP3s using Cleaner 6. This is to allow the content to reach everybody even if Quicktime is not installed or working properly.

So, there is a general overview of these movies. They have gotten rave reviews over they last year or two and in my honest opinion are a fantastic learning tool that we use in several of our classes. You can find an example of a these on my site at: http://www.mentalx.com/interaction_02.php