What spec machine are you using it on at the moment?

A Pentium 4 machine at 2.4 GB with 1 GB RAM and a GeForce4 Ti4200 128 MB running two IBM LCD monitors. It has two SATA disks (that work very nice and fast).

Do you make use of the full suite?

Not at the moment - I only use the switcher and the CG module. But the editor might replace our AVID express some day. I'm looking forward to using Video Toaster[3] with the VTEdit spline editor.

How many of the live morning calls have you done, and how many have been done so far?

We broadcast once a day and soon will be up to twice a day. I have been running the production and managing the entire process since we started, but now we are expanding the group to include a few of my colleagues. It takes about two days/mornings to teach a new person how to run the process.

What's the process for creating one of the programs?

First of all, this is our setup:

We use a Panasonic DVCPro camera, which is connected to the Video Toaster using the composite out. The Video Toaster out is connected to a socket in the wall in our front office. We have three or four locations that we can chose from.

We use an IBM T30 Notepad to administer the encoding.

The Video Toaster's composite out goes to our encoder, located in our secured server room. The encoder is an Intel server with dual 3.2GB Xeon processors. The encoder makes four streams (this will change to six soon) that are sent to Akamai.

Akamai is an internet content distributor that has around 13,000 servers around the internet. We could host the content from Denmark, since we are an ISP. We already have the bandwidth capability or we could simply add it if needed, but we chose Akamai because we expect that the number of visitors will exceed 1,000 simultaneous viewers as soon as we start to market this new product. Building network capacity to support the bandwidth that this amount of traffic would generate is not cost-effective. Also, because our audience is global we would run into a lot of bottleneck issues if we tried to host this locally.

We also have a satellite uplink that enables us to broadcast to other TV stations. Within 20 minutes we can broadcast to any TV station around the world that can receive a satellite transmission. If we want to make a TV interview we just select our TV socket. We have four sockets in our front office (dealing office) that are connected to an encoder-server and a 32Mbit MPEG2 encoder that is located in our server room. The MPEG2 encoder is for regular TV broadcasting.

Now I'll detail the process:

If they are needed, titles, graphics and animations are prepared 30 minutes before we start the process.

1. I start up a web interface that I wrote. The web interface is used to start, stop, monitor our encoders and upload files, calculate time offsets, etc. After the presentation, the files are copied to the Akamai content storage area.

2. The encoders are started.

3. The web page is automatically switched into live mode.

4. Now I move on to the Video Toaster and run a countdown sequence from the DDR.

5. When the countdown has finished I fade into the camera feed.

6. Idents and titles are added using the CG module. Later I will probably use a second DDR to achieve animated titles etc.

7. When the morning call is over I run an outro sequence and fade out to black.

8. Finally I add a CG sign that displays contact information to our research department.

9. Now I can press VOD on my web interface to copy the streams to the Akamai network.

10. While the files are being uploaded the web page is forced into an Upload mode. During this phase a sign will display. "The Morning Call will come up shortly" or something like that.