Yesterday we had a meeting using some videoconferencing. In a situation I find fairly common, the setup was a meeting room with many people, and then a small number of people calling in remotely. In spite of this being a fairly common situation, I have had trouble finding conferencing systems that do this particular task very well. I have not been looking in the high-priced end but I believe the more modestly priced tools should be able to focus on this and make it work. Yesterday we used Oovoo, one of the few multi-part conference systems to support PC and Mac, with some good but many bad results.

The common answer, namely a speakerphone on the meeting room table and a conference bridge system, is pretty unsatisfactory, though the technology is stable enough that it is easy to get going. The remote people are never really part of the meeting. It's harder for them to engage in random banter, and the call fidelity is usually low and never better than PSTN phone quality. They usually have trouble hearing some of the people in the meeting room, though fancier systems with remote microphones help a bit with that.

The audio level

The next step up is a higher quality audio call. For this Skype is an excellent and free solution. The additional audio quality offers a closer sense of being in the room, and better hearing in both directions. It comes with a downside in that tools like Skype often pick up ambient noise in the room (mostly with remote callers) including clacking of keyboards, random background noises and bleeps and bloops of software using the speakers of the computer. While Skype has very good echo cancellation for those who wish to use it in speakerphone mode, I still strongly recommend the use of headsets by those calling in remotely, and even the judicious use of muting. There's a lot more Skype and others could do in this department, but a headset is a real winner, and they are cheap.

Most of these notes also apply to video calling which of course includes audio. * Skype and some others offer a video display which shows who is speaking at any given moment. That's very handy and it also lets you immediately identify who is broadcasting background noise and needs to mute. * When people leave or are disconnected from a call, reconnect seems to always have some pain. Most PSTN conference systems make you go through an identification rigmarole with a code number, instead of just seeing the caller-ID and putting you right back in -- terrible. * Re-calling in Skype and other tools is usually one easy click, but annoyingly this often means a distracting ring to the parties on the call and a conference master has to approve them. For meetings of more than a couple of people, automatic re-entry for lost members is a must in my view. Most calls probably should happily let any buddy join, but even modest-security calls would do OK letting a person join without any conference master action if the person has already been on the call and not manually removed by the conference master. * In fact, any sudden disconnect due to anything but explicit action by a caller or conference master should probably lead to immediate (silent to others) attempts at reconnect for a while, unless that's deliberately aborted. * Usually in the conference room, the audio will be on big speakers and on loud. So there should be special care taken that the PC not play any sounds (except perhaps at a very muted level) during the call, particularly incoming rings and notification bloops. If these must be played, the typical volume of a quiet speaker should be measured and used for the sounds, not their normal volume. Because of the way volume levels work, we found each bloop and ring caused people to physically jump from surprise -- this was a dealbreaker. * If other software is going to make sounds out a PC's speakers, the conferencing software should see this (by tracking output sound levels from all programs) and mute the microphone during these bleeps and bloops if the party is not talking. We don't want to hear every time a remote party gets an instant message or email. * One solution to the above that many of the systems support is to get a cheap USB audio adapter and put the conference audio on that sound device, while all signals and other programs use the "regular" device which can be kept low. * The use of an external mic in the conference room is highly encouraged. They are not expensive. The mic can be placed in the center of the table, and away from clacking keyboards. The mic in a laptop is usually a bad idea, especially if somebody then needs to use the keyboard or trackpad on that laptop. * It would be super nice if the conference software supported the use of multiple mics. USB microphones are cheap so you can often have many on a single computer. Another option, since most people in meeting rooms have laptops running, would be networked software using all the mics of the laptops in a clever fashion. (While laptop mics are poor, 10 laptop mics working as private mics for the people right next to them is another story. People with laptops using a cheap clip-on mic would do even better and make the audio superb for the remote folks.) * The ability to bring in people on PSTN phones is of course nice in all these systems, and a number of them do have it (at a cost.) You still can get useful features like display of who is talking. However, everybody is still lacking in areas like easy join and automatic rejoin unless you dedicate a number to just these calls.

Video notes

All the audio notes apply to video tools as well. However, there are special notes for multi-party video tools when it comes to a meeting-room-to-many conference.

• You really want a remote webcam on a wire, and you want to see if you can put it up on the wall or a tripod at or beyond the end of the table so it looks down from about the height of a standing person. That way the camera is able to see everybody and show the dynamic of the table. A laptop at the end of the table doesn't see a lot of people, and a those closer in block those behind them. USB webcams are quite inexpensive, and the slightly more expensive ($50) models are way better than the ones built into computers. An ideal location for the camera might well be up on the wall at the end of the table, with nobody sitting at that end.
• As noted, even though the camera should be a bit away from the table to get everybody, the microphone needs to be closer to people. The speakers might need to be near the mic for good echo cancellation in some systems, otherwise they belong with the screen.
• A big screen, such as a flat-panel HDTV or monitor, is a big plus. The bigger the people are on screen, the more presence they get in the room, the less they are ignored. An HDTV, mounted on the wall at the end of the table, with a USB webcam, seems like the right approach. The HDTV may also have decent speakers.
• Indeed, most computers can drive multiple monitors, and all laptops can drive a 2nd monitor along with their built-in. Software should detect this, and handle it (ie. put things on the right screen) and even better, make use of it, putting some things on one screen and others on the 2nd. With monitors and TVs being so cheap, this should even become a standard way of doing it.
• Of course, if the meeting is small and more dedicated to video, it also works well to have everybody on one side of the table facing a camera/display on the other side of the table or on the wall.
• The view of the conference room has many people and thus needs to be high-resolution. Cameras showing single remote participant talking heads don't need to be high resolution at all -- even 160x120 is sufficient for this. As such a mode which makes the main conference camera transmit higher quality, and allows it to be bigger compared to the other participants makes a lot of sense. A fancier system would figure out who is speaking and temporarily switch them to better video if they talk for a long enough time.
• While most webcams are 4:3 and some are 16:9, when showing a single talking head, it can make sense to switch to 3:4 in a multi-party conference, as you can then fit more people on screen at higher resolution. Do this at the sending end of course.

Oovoo has a feature that lets a camera request higher quality, which at first seemed great, but turned out to be broken. When this was enabled, it switched the sender's screen to a mode where their own video was large and the other callers were tiny, with no apparent way out. In the main conference room you want to be sending big video but you don't want the self-view to be large if you see it at all. It definitely should not make the other callers even smaller.

Second camera

Consider a second camera in a multi-party video system. This could be a 2nd camera on the video call computer if it supports that, or a second computer with a camera. The former, if designed for, could be more efficient with one combined video stream sent to the conference bridge, though recipients might split it apart.

• The second camera can be fixed, simply pointing at a different angle to make sure there is good coverage of everybody in the room. This can also allow all parts of a room or all seats at the table to be used, as they can be seen from one of the cameras.
• The second camera can also be moved around, pointed by a willing helper at the table or an official camera operator at whoever is speaking, or at slides. This is quite nice and lets the remote people get a good view of the speakers as well as the sense of the whole room. The wide view shows people mostly as blobs so the 2nd camera provides facial communication.
• A fancy product would be a camera that sits in the center of the table and automatically figures out where sound is coming from and points to that person if they speak for more than a few seconds. Slightly less fancy would be a camera with presets for various positions around the table, and an interface to let the remote callers pick where to point it. One could also have a standard pan/tilt camera (These, like the Logitech Orbit, can be had for under $100) controlled by one of the remote callers. (You can have local control, but trust me, the remote callers are more keen on this.) I have always been disappointed that even though Skype made the Orbit one of its 3 officially supported HQ cameras they never put in support for pan/tilt/zoom.
• Possibly simpler again would be use of cheaper fixed cameras, which can be had for under $20. Just make a tree of about 6 of them to cover a circle. You may not have the bandwidth to transmit them all, but remotes or sound direction sensing could turn one one as needed.
• Finally, the cameras in all the laptops around the table could be activated and selectable by the remote particpants for views of individual speakers who have them and have loaded in some software.
• Desktop share is of course useful for playing slides to remote participants.

Particular problems

As noted, the crew were not pleased with Oovoo. We liked the concept with multi-party video arrayed on the screen, but the execution had problems as noted above. In addition, there is a hard truth about any software that requires futzing during the meeting. While the remote people are usually very keen to improve their perception of the meeting and their presence in it, at least one person in our meeting grew livid at any interruptions or delays caused by adjusting or playing with the AV technology. People will get frustrated when the tech dominates the meeting and wish to switch back to the inferior tech, at the expense of others who are gaining benefit from the better tech.

OOvoo is sadly not stable, and it was not unusual for callers to get bounced from the call. As noted above, this caused a jarring loud ring to be played when they called back in, and then somebody had to go to the end of the room to let them back into the call while everybody covered their ears. Having this happen a dozen times in a meeting is just not workable. Oovoo also suffered from the problem of playing back its own bloops. For example, if I sent an instant message to a remote party to mute their microphone, everybody would be jarred by the bleep of the instant message being played out the remote speaker and back into the remote microphone to us.

Oovoo's ability to have a user jump to higher quality is good, but was unusable because of how it rearranged the sender's screen.

Oovoo does not allow you to mute individual members of the call. This is odd as this feature is found in almost all other systems. At one point a participant took a phone call, and muted his speaker but did not mute his microphone. So we sat there listening to him make his long phone call and the only way to block him out was to mute the entire call. This was another deal-breaker bug in the view of many.

All this after judging Oovoo to be better than many of the other contenders, though we have yet to try Sightspeed. Skype now offers a multi-party video call, and I am eager to work with it, but at present it only works on Windows, and most of the people at our meetings are on Macs. My desktop is on Linux (which Skype is one of the few to support) but Skype Linux is more limited that either Skype Windows or Mac.