IP Video Task Force Report
IP Video Endpoint (Codec) Evaluation and Recommendations
This section reviews IP codecs according to user needs, performance, and compatibility. Accompanying tables provide summary comparisons of various vendor offerings.
PRIMARY QUESTIONS
The Task Force first examined endpoint equipment to better understand the end-user experience as H.323 implementation proceeds on ITN. Selection criteria thus addressed both questions pertaining to user friendliness and reliability issues as well as more technical questions affecting reliability and compatibility issues. To further refine the criteria and understand the end-user’s point of view, several H.323 terminals (codecs) were evaluated.
User Concerns
- What is H.323 terminal equipment?
- What is a codec?
- Why not use NetMeeting®?
- How many brands of equipment should ITN support and how will the decision affect the end-user?
- What is a room system versus a set-top system versus a desktop system and what are the applications of each type?
Technical Support Concerns
- What are the differences and advantages/disadvantages between appliance-based codecs and PC-based codecs?
- What are the compatibility concerns between different manufacturers and technologies?
ANSWERS
User Concerns
What is H.323 terminal equipment? Terminal equipment refers to any piece of equipment used to call or receive calls on a network. Specifically, H.323 terminal equipment is any piece of equipment used to make or receive calls on an IP based network. The equipment takes the form of a camera (transmits the user’s picture), monitor (displays picture and audio from the far end site), microphone (transmits end-user’s audio to far end site), codec (PC or appliance based, see below), keyboard, and mouse (provide direct interface with the video conferencing unit). This is the only equipment with which the end-user directly interacts when making or receiving a call.
What is a codec? The word codec is an acronym for
COder/DECoder. In short, a codec is a piece of hardware or software that takes
some signal (such as video or audio) and converts it to a format suitable for
transportation using a specific set of protocols. Equipment on the receiving
end must also use these same protocols. The H.323 standard is a collection of
protocols designed to ensure compatibility between products of various manufacturers.
Even though the term “codec” specifically refers to the hardware
or software converter, the term has come to include the entire set of video conferencing
terminal equipment. Someone saying “codec” is most likely referring
to a set of terminal equipment. The terms “video conferencing terminal”
and “codec” will be used interchangeably throughout this document.
Why not use NetMeeting®? NetMeeting® and CUSeeMe are software based
codecs. There is no hardware. NetMeeting® and CUSeeMe are totally dependent
upon the processor’s power to encode and decode audio and video. The more
programs that are open and running in the background, the less processing power
there is for NetMeeting® and CUSeeMe to use. Moreover, NetMeeting® and
CUSeeMe can only send a video stream at 128 kbps. While this may be acceptable
for point-to-point meetings, this will not be adequate for multipoint conferences
on a bridge that requires 384 kbps. Not all bridge manufactures support the
use of NetMeeting® at this time. Therefore, at this time, NetMeeting®
and CUSeeMe should not be supported or allowed into multipoint conferences.
As technology improves, this decision will be reevaluated.
How many brands of equipment should ITN support and how will the decision affect the end-user? The IPVTF determined it would be wise to support products from two manufacturers. Choosing one manufacturer would unduly restrict customer choice. Three or more product lines would be expensive to support. Choosing two allows ITN customers to select the codec that best meets their needs, while allowing ITN support staff to maintain an in-depth knowledge of the equipment. Keeping the number of manufacturers to a minimum will also assist ITN by allowing staff and subcontractors to focus on resolving any issues associated with the products. ITN customers can choose another manufacturer if they wish; however, ITN staff will not be able to support that equipment if the end-user has difficulty.
What is a room system versus a set-top system versus a desktop system and what are the applications of each type? Researching manufacturers made it clear that various vendors interpret video conferencing terminal endpoint solutions differently. To clarify discussion and evaluation, the IPVTF devised the following three classifications for terminal endpoint equipment: Room System, Set-Top System, and Desktop System.
Room System
A room system is a vendor’s top of the line product designed to provide medium and large corporations, government, and educational institutions with custom room configurations. These systems are used in situations requiring high-quality video performance and extensive conferencing capabilities for applications such as distance learning, boardroom conferencing and high performance multi-site conferencing. Performance is based on providing IP data rates as fast as 2 Megabits per second. These systems typically support multiple video and audio inputs such as document cameras, computer video converters, and VCRs. Support for one to three video displays is also common. Complete systems may come with two or more 32” viewing monitors or can be used with larger projection systems. These systems can be permanently built into a room or set up as semi-mobile units that can be taken to another room if necessary. Some units even have built-in streaming services. The main advantages to these systems are video quality and built-in multipoint services. The main disadvantage is price; these systems start in price at $14,000 list and go up from there.Set-Top System
The term set-top system usually refers to a vendor’s mid-level product designed for small to medium sized conference rooms. This system can look just like the room system with monitors and carts, or it can be a smaller unit that sits on top of a monitor. These systems typically will perform the same as the room systems at lower data rates; however, the set-top systems typically do not support data rates above 768K and do not have the built-in multipoint conference capability. The list pricing for set-top systems can be in the $3,000 to $12,000 range. One advantage to the set-top solution is that products on the lower end of this pricing scale can give a user excellent conference room or classroom performance as long as they do not require the extra services or options offered by the room systems. The disadvantages are that multipoint services would need to be provided elsewhere (such as IHETS) and that higher data rates (more than 768 Kilobits per second) are not attainable.Desktop System
A desktop system, simply put, is a video conferencing terminal, either hardware or software, that interacts directly with the personal computer on a desk. It is designed to be a one-person unit, making it unsuitable for a conference room environment; but it can be very useful for an office or classroom/lab-station solution. It will either connect through the USB port or will have a hardware card to be installed. The speed of the computer processor will directly affect performance. Desktop systems list from $599 to $4,995. The advantage of such a unit is its low-cost focus on an individual user, with no special room or setup needed. Within minutes of installation, a user can talk face to face to someone using another desktop system or someone with a $50,000 room system. The disadvantage to this type of system is that it is not designed for conference rooms.
Technical Support Concerns
What are the differences and advantages/disadvantages between appliance-based codecs and PC-based codecs? The fundamental difference between the two is that PC-based systems rely on a personal computer as a platform for the user interface, and in some cases the entire system, while appliance-based systems are self contained and do not use MS-DOS or Windows.
PC-based Video conferencing Terminal
A PC-based video conferencing terminal is completely integrated into a PC (not necessarily a desktop system, this could be a room system). The designers of such systems started with a Windows platform and added the necessary hardware and software components from that point. The primary advantage is that these systems can give the user immediate access to PC display capabilities in order to show PowerPoint slides or other software screens. These units typically support a method of application sharing as defined by the T.120 application-sharing standard. This allows an end-user to collaborate with an endpoint by electronically sharing a document or whiteboard and sharing control of the document or whiteboard. The main disadvantage of a PC-based system is that they operate on Windows 95/98/2000, all of which can lock up or crash. Reboot times are typically measured in minutes, significantly disrupting an important conference or class. Another disadvantage to a PC-based system is that not all multipoint conference units support T.120 application sharing, thereby limiting application sharing to point-to-point calls. Multipoint application sharing would have to be performed using third-party software such as NetMeeting. (Note: While NetMeeting is capable of H.323 communication, the fact that it uses a software-based codec makes it extremely unreliable, to the point that it has caused H.323 network components such as MCUs to crash. Therefore, the IPVTF does not consider NetMeeting a viable H.323 option. This includes other NetMeeting class software such as CUSeeMe. However, the T.120 functions of NetMeeting make it an extremely useful tool worth mentioning. Some of the PC based platforms actually use NetMeeting to support their T.120 functionality)Appliance-based Video conferencing Terminal
An appliance-based video conferencing terminal does not require a personal computer to function, being self-contained with proprietary hardware inside. Units contain their operating software in an internal flash-type memory bank, measuring boot times in seconds. Since these units are designed from the ground up to be videoconference units, the manufacturers do not need to make compromises to accommodate a Windows type of operating environment. User navigation menus are highly graphical with multiple levels of on screen help available. The user interfaces are highly intuitive and are adapted to easily by the novice. A main advantage is ease of installation. With a PC-based system, the installer needs to possess the basic skills necessary for PC installation in a network environment. For an appliance-based unit, the user typically needs to hook it up to one or two video monitors, supply a network connection, power up the unit, and follow the on screen instructions for software setup. A key drawback of these units is their lack of T.120 application sharing support. Also, the lack of modularity means that when they fail, the entire unit needs to be sent in for service, rather than a small component of the system.
What are the compatibility concerns between different manufacturers and technologies? Vendors will always want to maintain a competitive edge over their rivals, so as technology improves, vendors apply their own interpretations to a recommended industry standard. When a technology is young there are many interpretations as to how the technology should be implemented. As the technology matures and standards are applied, there is more compatibility between vendor’s products. However, with the vendor’s attempt to be unique, the product may have a component that is proprietary, causing incompatibility with other products and limiting repair solutions. Purchasers should always ask vendors what products are compatible with their products early in their research.
NOTE: Some codec components must be assessed by the person making the comparison. For example, how the picture looks on one monitor versus another is as much a matter of subjective judgment as it is of objective or quantitative measures. The Task Force’s criteria, therefore, have avoided the qualitative in favor of network design and management impact.
Criteria for Choosing Systems to Support
The IP Video Task Force formulated three criteria on which to base recommendations for recommended video conferencing systems. The criteria take the end-user into consideration by choosing a system that is simple to operate and provides high-quality video.
The first criterion requires that the manufacturers we support have a wide offering of products. This translates to a range of products from large room systems down to the small desktop systems. The IPVTF believes that if the manufacturer has a wide offering, end-users can choose a unit to fit their needs as closely as possible. This criterion benefits both the end-user and the ITN technical support structure; the customer has many choices, and the support staff has fewer technical and training issues to resolve than if ITN supported many manufacturers.
The second criterion requires ease of set-up and use, especially for non-technical end-users.
The third criterion requires the units to provide high-quality video. Of course, the problem here is that quality cannot easily be quantified. The IPVTF therefore focused attention on items such as resolution, motion artifacts, and standards support to help assess quality.
Room System Descriptions
POLYCOM
Viewstation VS4000 H.323 $14,999 List
Viewstation VS4000 V.35* $16,998 List
Viewstation VS4000 PRI* $18, 998 List
Viewstation FX H.323 $14,999 List
Viewstation FX V.35* $16,998 List
Viewstation FX PRI* $18, 998 List
Viewstation H.323* $6,999 List
Viewstation 128* $5,999 List
Viewstation SP 384* $5,999 List
Viewstation SP128* $3,999 List
NOTE: Codecs with an asterisk were not evaluated. They are included to show other models in the vendor’s product line. The additional codecs are mentioned in Appendix E.
The Polycom Viewstation video conferencing terminals are all appliance based. Installation for all units with the exception of the VS4000 consists of removing the package from box and placing unit on top of a video monitor. The VS4000 mounts in a standard 19” equipment rack and uses standard Sony PTZ type cameras. The rear connections of all units are color coded to match the cabling, making installation simpler. A picture card is provided showing where the appropriate cables need to be attached on the unit and the monitor, making installation nearly foolproof. After turning on the unit for the first time, the user interface is accessed by moving the cursor to the small blue information button. Information needed by the end-user or set-up technician to get started is: IP of the unit, IP of the Gatekeeper, IP of the network’s gateway. Units can also be used in a dynamic IP environment where the DHCP network protocol is used for IP assignment, thereby simplifying the setup process. Units must be restarted when changing gatekeeper or other IP sensitive information; however, this usually takes less than 30 seconds. The user interface is straightforward and intuitive making extensive use of large descriptive icons. Firmware upgrades can be performed via the network using a web browser interface. This web interface also gives system administrators complete administrative control of the unit. Units can be controlled by hand held remote or by web interface. QoS support is built-in and easy to configure based on the network requirements. Polycom Viewstation FX and VS4000 units can also act as four person MCUs. However, bandwidth of each call is divided by the number of participants in the conference. For instance, a four-port conference call would be limited to 384 kbps.
PICTURETEL
970* $20,490 List
960* $13,990 List
680* $6,995 List
No Picturetel units could be obtained for evaluation.
VCON
MC8000 Model 3 $14,995 List
MC8000 Model 1* $11,495 List
VCON Media Connect 8000 (MC8000) is a PC based codec that operates on the Windows NT platform as well as Windows 2000. User interface is easy to use but not as intuitive as Polycom’s Viewstation. Meeting point works with Microsoft’s NetMeeting in the background for T.120 application sharing support. QoS is supported by the VCON PacketAssist and is proprietary to VCON. Multicasting can also be accomplished with VCON but this is also a proprietary standard.
TANDBERG
7000* $44,990 List
6000 $23,990 – $34,490 List
2500 $15,990 – $24,990 List
800* $9,990 List
500* $4,990 - $7,490 List
Tandberg 2500 and 6000 codecs are appliance based. The menu system is not intuitive for the end-user, and very difficult to navigate. Tandberg’s B3 software allows for cascading multiple codecs for the purpose of creating a pseudo-MCU. Using three codecs, a multipoint conference with up to nine other users can be achieved. QoS is supported within the software settings.
VTEL
Galaxy $9,995 - $34,995 List
Galaxy 2.02 is the current model evaluated at the time of this document. It is a PC-based video conferencing terminal operating under Windows 98 platform. At this time, the Galaxy OS software does not support Windows 98 SE. Internet Explorer upgrades are also not supported. Any automatic updates will cause the machine not to work and the OS will have to be reinstalled. Version 2.1 is the current version being shipped. Version 2.2 should be released by the end of the fourth quarter 2001. Version 2.2 will have the OS and IE fixes to allow for upgrades to the latest Windows 98 offerings.
The interface is significantly different from the ESA model currently used by ITN ATM video endpoints. There will be some adjustment to the placement of the features for those operators switching from the ESA to the Galaxy. The Galaxy comes with a wireless keyboard, a mouse, and a hand-held touch screen unit. Any of the peripherals can be used to operate the codec. As with other PC based codecs, changes to IP sensitive information will require a reboot for those changes to take effect, which could take several minutes.
NOTE: Some of the video conferencing systems covered here come with a stand and monitor. Some do not. Basic pricing is stated below for additional components if required.
27” NTSC Sony $950
27” Cart $1,000
29” XGA Monivision $2,000
32” NTSC Sony $1,500
32” Cart $1,100
34” XGA Monivision $3,250
Desktop System Descriptions
POLYCOM
ViaVideo $599 List
Polycom ViaVideo is a desktop video conferencing that sits on top or to the side of the PC’s video monitor. The ViaVideo connects to the PC’s USB port. Setup includes loading the software onto the PC, plugging the unit into the USB port and the sound card, and providing power. There is no opening of the PC case to insert a hardware card, as might be required by some other manufacturer’s offerings. Therefore, it is conceivable that this unit could be installed and set up by the end-user. Information required by the end-user or set-up technician to get started is: IP of the unit (if it will differ from that of the PC) and IP of the Gatekeeper to which the unit will be registered. The user interface is straightforward and intuitive and designed very similarly to that of the Viewstation line of offerings from Polycom. This would facilitate an easy transition for those who might use both in their operations. Upgrades are performed by downloading the latest software from the Internet and installing it on the PC. ViaVideo uses the PC’s processor for decoding while the camera unit handles the encoding. This makes the unit a much more effective unit than all software based programs such as NetMeeting or CUSeeMe. The unit also works well with MCUs. It is recommended, however, that the PC not be running a lot of other software while using the codec, as performance will decrease. ViaVideo’s performance can be directly related to processor power and amount of RAM available. ViaVideo has a plastic lens versus a glass lens. This tends to show a slightly grainy but clear picture on the monitor. This graininess does not impede the interaction, sharpness, or distinctness of the experience. QoS is fully supported and easily accessible through the user interface.
PICTURETEL
Picturetel 550* PC Based codec. $1,295 List
No Picturetel units could be obtained for evaluation.
VCON
ViGO Standard* $699 List
ViGO Professional $1,199 List
ViGO Executive* $1,995 List
Cruiser 150* $1,295 List
Cruiser 384 $2,995 List
Cruiser Executive*$4,995 List
Escort 25 $899 List
VCON has three products in the desktop arena. All VCON products are PC hardware codecs. The Escort and Cruiser require one PCI slot. The ViGO has its own base that connects to the PC via USB. All cameras are glass lens. The cameras sit either on top of the monitor or on top of the ViGO base. All VCON products come with VCONs PacketAssist QoS technology and multicasting capability. Depending on type of model purchased, the microphone will be incased in the camera or as an optional tabletop microphone. Installation is straightforward by inserting a CD and then following the instructions. Inserting the IPs for the gatekeeper and the system the unit is running on is done during setup. When changes are made, the software must be restarted for the changes to take effect.
NOTE: Expanded views of Table 1 and Table 2 can be seen in Appendix C and D respectively. Appendix E has a synopsis of codec cost (List Price) by vendor.
Table 1
Comparison of Desktop Videoconference Terminals by Vendor Model
Shaded region indicates a corresponding room system in that product line.
ViGo |
Escort/Cruiser |
ViaVideo |
PICTURETEL |
|
H.320 – ISDN |
X |
X |
||
H.323 – IP |
X |
X |
X |
X |
H.281 - Far end camera control |
X |
X |
X |
|
H.261 - Video codec for audiovisual services at p x 64 kpbs |
X |
X |
X |
X |
H.263 - Video coding for low bit rate communication |
X |
X |
X |
X |
G.711 - Pulse code modulation (PCM) of voice frequencies |
X |
X |
X |
X |
G.722 - 7 kHz audio-coding within 64 kbps |
X |
X |
X |
X |
G.728 - Coding of speech at 16 kbps using low-delay code excited linear prediction |
X |
X |
X |
X |
Acoustic Echo Cancellation (AEC) |
X |
X |
X |
X |
Automatic Gain Control (AGC) |
X |
X |
X |
|
T.120 - Data protocols for multimedia conferencing |
X |
X |
X |
X |
Quality of Service (QoS) |
X |
X |
X |
|
Frames per Second: fps |
30 |
30 |
30 |
30 |
Windows 98 |
X |
X |
X |
X |
Windows NT |
X |
X |
||
Windows 2000 |
X |
X |
X |
X |
Windows ME |
X |
X |
X |
|
Processor |
P II |
P |
P with MMX |
P II |
Max Lan Speed |
1.5 Mbps |
1.5 Mbps |
384 |
384 - ISDN 768 - IP |
USB or Card (ISA/PCI) |
USB |
PCI |
USB |
PCI* |
Camera (Manual/ PTZ ) |
Both |
Both |
Manual |
Manual |
* One free PCI slot
** Unit in degrees
Table 2
Comparison of Room System Codecs by Vendor
Shaded Region indicates a corresponding desktop system in that product line.
VCON |
POLYCOM |
PICTURETEL |
TANDBERG |
VTEL |
|||
MC8000 |
ViewStation FX |
600 |
900 |
2500 |
6000 |
Galaxy |
|
H.320 – ISDN |
X |
X |
X |
X |
X |
X |
X |
H.323 – IP |
X |
X |
X |
X |
X |
X |
X |
H.281 - Far end camera control |
X |
X |
X |
X |
X |
X |
X |
H.261 - Video codec for audiovisual services at p x 64 |
X |
X |
X |
X |
X |
X |
X |
H.263 - Video coding for low bit rate communication |
X |
X |
X |
X |
X |
X |
X |
G.711 - Pulse code modulation (PCM) of voice frequencies |
X |
X |
X |
X |
X |
X |
X |
G.722 - 7 kHz audio-coding within 64 kbps |
X |
X |
X |
X |
X |
X |
X |
G.728 - Coding of speech at 16 kbps |
X |
X |
X |
X |
X |
X |
X |
Acoustic Echo Cancellation |
X |
X |
X |
X |
X |
X |
|
Automatic Gain Control |
X |
X |
X |
X |
X |
||
T.120 - Data protocols for multimedia conferencing |
X |
X |
X |
X |
X |
X |
X |
Quality of Service (QoS) |
X |
X |
X |
X |
|||
Frames per Second: fps |
30 |
30* |
30* |
30 |
30 |
30 |
|
Windows NT |
X |
X |
|||||
Windows 2000 |
X |
X |
|||||
Camera (Fixed/PTZ) |
PTZ |
PTZ |
PTZ |
PTZ |
PTZ |
PTZ |
PTZ |
Focus/Brightness/ White Balance |
Auto |
Auto |
Auto or Manual |
Auto or Manual |
Auto |
||
VISCA camera support |
X |
X |
X |
X |
X |
||
Voice activated camera positioning |
X |
X |
X |
||||
Built-in MCU CAPABILITY |
X |
X |
X |
||||
Max Lan Speed |
1.5 Mbps |
2 Mbps |
2 Mbps |
||||
Web access for control of codec |
X |
X |
X |
||||
*30 fps at 256Kbps and above
**Units in degrees
Recommendation: The IP Video Task Force recommends that ITN support products from the following manufacturers. The Task Force also recommends that ITN customers purchase from these choices in order to have access to a larger knowledge base for troubleshooting. These recommendations are based on how well the manufacturers meet the criteria based on product comparisons described in the previous section and the feature comparisons in Tables 1 and 2.
Polycom:
Appliance Based Room Systems
- Viewstation FX
- Viewstation VS4000
- Viewstation MP
Appliance Based Set-Top Systems
- Viewstation H.323
- Viewstation SP128
- Viewstation SP384
- Viewstation 128
PC Based Desktop System
- ViaVideo
VCON:
PC Based Room Systems
- Media Connect 8000 model 3, MC8000
PC Based Desktop System
- ViGO
- Escort
- Cruiser
