Our post-production facility has many networked computer-based editing and media management systems. In the administration of these bread-and-butter production systems, I have one hard-and-fast rule — no Internet connections. Oh, sure, we have lots of “connected” workstations, separated from the production workstations with physical space. Only “sneakernet” is used to transfer files between these two worlds. I had an Internet-connected DVD workstation that was destroyed by an e-mail virus a few years ago, which confirmed my policy. And, our office network has been attacked on numerous occasions, requiring us to implement some of the firewall and anti-virus techniques described in October's issue.
I would like to see an article on ways to enable isolated networks to still share files with Internet-enabled networks, on a selective basis. As an example, a client sends us a logo graphic file to use in a spot. We download it on the office network, dump it to a disk (or jump-drive or firewire drive or whatever) and connect that to the production network. How about a separate connection to that office network, with a way to temporarily enable a file-transfer connection, just for the amount of time needed to transfer the info? This, effectively, isolates the production system from the Internet, while still enabling us to be “connected” to the outside world.
Brad Gilmer responds:
There are several things you can do to “cautiously connect” your broadcast network to the Internet. At a high level, one approach would be to build two independent networks, one for the broadcast network and one for the business/Internet. You can then control what crosses over between the two networks based upon things like protocol, MAC address of the source or destination computer, or some other parameters.
Another thing you can do is use dual-homed hosts — servers with two NIC cards. You plug one NIC card into the broadcast network and one NIC card into the business/Internet network. The server can be configured to make its data available on both networks. Also, most servers can be configured to be a router, routing permitted traffic between the two networks. While you can use the server to do both jobs, I suggest you use the server to deliver its content to both networks, but use a separate router to control the traffic moving between the two networks.
Dear Michael Robin,
I bought both of your books, but I'm struggling against the pixels. Are the pixels that you sample really displayable pixels, or just samples?
If they are samples, can you never reach more real display pixels than half the samples (Nyquist)?
Michael Robin responds:
If you are sampling an analog video signal, the horizontal resolution, (expressed in LPH) is equal to the maximum video frequency multiplied by the resolution factor. You can't do better! The CCIR 601 maximum allowable frequency is 5.75MHz (lower than the 6.75 Nyquist frequency) so the resulting resolution, at best, is approximately 455LPH. If, on the other hand, you have a concatenation of conversions back and forth between analog and digital, as in a broadcasting studio, the end result cannot be better than what the resulting analog frequency response allows. This will definitely be lower than 5.75MHz.
If you are talking about samples, CCIR 601, 720 active horizontal samples equals 360 sinewaves during an active line, or a resulting frequency of 6.75MHz. So, because the maximum video frequency is 5.75MHz, we have no hope of ever displaying so many pixels on an active line. What we can hope to see is about 613 pixels per active line.
Bear in mind that we live in a mixed analog/digital world!
Q. Name the company and product name of the first videodisk recorder for ENG applications.
A. NEC Diskcam
Test your knowledge!
See the Freezeframe question of the month on page 8 and enter to win a Broadcast Engineering T-shirt.
Send answers to email@example.com