Page 287 - The ROV Manual - A User Guide for Remotely Operated Vehicles 2nd edition

276 CHAPTER 10 Video
satellite networks, video can be cost-effectively transmitted real time with little time delay or loss in resolution.
Typically, a composite video feed is brought to the surface, converted to MPEG-4 (or equiva- lent) digital video, and transmitted to a multicasting station for rebroadcast. Each node on the net- work (e.g., the MPEG-2 converter box) is assigned an IP address whereby either the video is assigned to a port on the multicasting station or an individual may log into the converter at the field (i.e., ROV surface console) location.
Remotely (typically from the client’s office ashore) logging into a remote video server aboard the vessel of opportunity (e.g., the converter located within the ROV shack) and then viewing the video individually is certainly acceptable. This is termed “pulling” the video from the converter. The challenge with this method is the video pipeline is limited to the one user as the video requires a certain bandwidth in order to maintain a certain nominal image quality (e.g., 540 kb/s). If a sec- ond user logs in remotely to the same port on the converter, the bandwidth is cut by half, which destroys the video feed for both individuals. A far superior method for multiple viewing of video streams is to assign the video feed to a port on a multicasting service whereby the signal is “pushed” to a port on the server of an Internet Service Provider (or within the ROV service com- pany) with the signal then multicast to the population of viewers logging into the multicast session for their viewing pleasure.
The challenge is then to change the outbound video signal by adjusting the frame rate, frame size, and video resolution to conform within the bandwidth constraints of the video connection.
Two methods of video transmittal are used depending upon the tolerance for latency time between video event occurrence and display at the final viewer’s location—Transmission Control Protocol (TCP) and User Datagram Protocol (UDP). TCP allows for data packet proof of receipt circle-back for confirmation of packet receipt. The benefit of this method of transmission is that the picture quality is of the highest quality due to resubmission of packets should they be lost in trans- mission. The problem with this method is that with a dodgy connection, the packet loss is high and the bandwidth is taken up by lost, then retransmitted, packets causing an inordinate amount of latency time between the transmission of the video and the final receipt. The UDP protocol, on the other hand, transmits its packets at full rate with no confirmation of packet receipt by the end receiver. So, TCP is characterized with high video quality accompanied by a certain (perhaps sub- stantial) amount of latency time between picture capture and end display while UDP allows for full data rate video transmission of varying (perhaps very poor) image quality.
10.7 Video documentation
The difference between a film produced by a Hollywood production company and an amateur vide- ographer is less about the equipment and more about the technique. What separates the amateur from the pro is the attention to detail that allows a complete portrayal of the subject matter through moving images. It is not about the image, it is about the message.
It has been said in many different industries: “If you didn’t get it on video, it didn’t happen.” The meaning of this saying indicates that video and still-camera recordings are an integral part of a professional documentation package, forming part of the deliverable at the end of a project. Part of