CTO - Did You Know?: Streaming Video
Richard Blackwell, Chief Technology Officer, IVCi
In the literal sense ALL video is streamed, so it is somewhat a misnomer that the term is applied not to video conferencing (which is certainly streaming video), but only to web-based applications.
The question then becomes "When to Stream or Not to Stream?", and by this I refer to stream video or to use a multi-point device (software or hardware MCU) in order to reach many participants. How do you decide what is best based on the defined needs? Use the scenarios below to help you when faced with this question.
To begin with let us make some general statements;
- When you need real-time video (minimal latency), an MCU is the better answer due to the inherent latencies of streaming video (15 seconds to several minutes is typical in streaming) and no conversation can be held with even a one minute delay.
- You can reach practically unlimited numbers of end users with a streaming server at a very small incremental cost increase per user when compared to an MCU. When you have a hundred or more end users in a presentation, an MCU may be cost prohibitive. This would generally be a streaming application.
- If delay (latency) is of little or no concern, the quality of streamed content can be quite good, even over the Internet. On the other hand, if you need low latency, the Internet will greatly degrade your video as there is no time to deal with typical lost packets, jitter or out of order packets.
- When video must be stored and made available for replay, a video on demand streaming solution is the answer. Recording a multipoint call and making available for playback on demand is only now becoming possible in some limited solutions (Polycom iPower now supports recording a call).
- You can mix the solutions.
Now let's share some Scenarios:
Customer is a school system that needs to give dozens of sites access to a class session. Storing the session would be nice.
Proposed Solution for Scenario #1
Being a school system, the budget is limited and the large number of required sites suggests a streamed session (instead of an MCU). Limited feedback is a nice-to-have feature and low latency is needed to be truly interactive (subjectively I suggest < 20 seconds). Conferserv would work nicely here, but may be expensive, so this may also be the Starbak solution.
CEO speaks weekly to top 10 lieutenants. Meeting is highly interactive and CEO wants to see all participants.
Proposed Solution for Scenario #2
In this case the CEO wants to see the faces of all attendees. Streaming video is a "one-to-many" solution (they all see the CEO, but he sees no one) where multipoint video is a "many-to-many" solution, which in this case, is the need of the CEO. This is a bridge solution.
A human resources department must deliver training to all employees on various topics. The topic is provided real time to classrooms around the country and feed back is expected.
Proposed Solution for Scenario #3
The desire is for quality video and the ability for replay to those that missed the class. Conferserv is the premium answer here as the quality of the video is very good. Latency is not low (30 seconds) but acceptable in a teaching environment. Feedback from the students is recorded and seen at the time of replay on demand.
The president of a company speaks to all employees once a quarter. The three department heads are to be seen with the president for real-time Q & A but are located in various cities.
Proposed Solution for Scenario #4
This is a classic mix and match. A low cost MCU can be employed (perhaps embedded) to connect the president to the three department heads and a streaming server is used to capture the four sites and send to the desktops of all employees. This could be a Tandberg/Polycom embedded streaming solution (depending on the number of employees) or Starbak/Conferserv.
Video Conferencing Etiquette
Ministry of Education, British Columbia
The following are some suggested practices to ensure a successful video conference for all participants and presenters.
- Arrive 15 minutes before the video conference. Arriving early provides an opportunity to find a seat, conduct introductions among participants, and organize materials for the meeting.
- Ensure you can be heard; move closer to the microphone if necessary. Avoid shuffling papers or tapping objects near the microphones. Keep your microphone on mute at all times until the chairperson/host asks your site to respond. This improves the audio quality for everyone in the video conference by removing extraneous room noise (chairs, paper shuffling etc.).
- Avoid side conversations that would exclude the far-end sites from your conversation. Interactivity between all locations is the key to a successful video conference.
- Ideas and comments from each remote site should be encouraged and discussed within the timeframe of the agenda.
- Once the connections have been made with the sites in the conference, all participants should introduce themselves.
This helps to establish appropriate audio levels in all the meeting rooms. Remember, just because the other participants are not in the same room, proper business protocol should still be followed. When your video conference has concluded, leave the room in a timely fashion as other conferences may be booked following yours.
Procedures for Site Hosts/Facilitators/Presenters
Familiarize yourself with the room, keypad and document camera or other special equipment. Presenters that will be using a number of devices (document camera, vcr, computer) may want to arrange a visit to the room a week or so before the video conference, to practice operating the equipment. If the event will be large, complex or critical we highly recommend a rehearsal session.
- In most sites with two monitors the left television monitor is the preview monitor and will show what the camera sees.
- The right television monitor is the monitor that will show the other sites.
- Mounted on the top of the right monitor is the camera.
- The switching from site to site is voice activated. It takes four seconds of continuous noise from your site for the video to show your image to the rest of the conference.
The keypad that controls the equipment is usually located on the conference room table. A participant in the room will operate the keypad for the session. The document camera is used for sending images of printed materials, much like an overhead projector, and is also capable of displaying 3-dimensional objects. Participants should remember to pause for reaction. In order to coordinate the audio and the video coming from a site there is a 1/4 second audio delay.
The video picture is compressed before being transmitted and decompressed at the receiving sites. There is a choice of two data rates, 112 kilobytes per second (low speed) and 384 kbps (high speed). The higher data rate produces a higher quality visual image. If there will be a lot of movement or video tapes shown, it is best to use the 384 kbps data rate. However, if the presentation doesn't contain a lot of movement, the 112 kbps data rate is suitable.
At present, there are over 100 video conferencing sites within British Columbia. Not every site has the capability of using the higher data rate. This should be taken into account when planning your presentation.
The compression process will make the other sites move on the monitor in a jerky motion. This distortion is called artifacting or pixelation.
Movement should be kept to a minimum to reduce this effect. With the higher transmission rate, the jerky movement is less noticeable. Most participants grow accustomed to this effect quite quickly.
For additional information regarding video conferencing, or to speak to a representative, please contact IVCi at 800-224-7083, or click here to have an IVCi Representative contact you.
All trade names, trademarks, and/or service marks herein are owned by their respective holders.