Like many of you, I live on conference calls. Meet me on the bridge is a standard part of our vocabularly, with calls on the calendar to avoid a call you're expecting being missed by a call you didn't expect.
Yesterday on a conference call using Citrix's GoToMeeting with James Tagg, CTO and Founder of Truphone, as well as a few others. James and I were calling in via the GoToMeeting client on our Mac Book Pros. One of my team was calling in via a mobile phone on AT&T, another person on the Truphone team was calling in via a SIP based PBX, and others who were doing more listening than talking were hearing clearly via GTM's clients on Macs or iPads.
The call quality ranged from crystal clear for those using the GoToMeeting clients, to muddy, depending who was doing the talking, and who was doing the listening. And this was on a conferencing bridge that actually mixes the audio (and video) to provide a combined stream back to all participants (Note Citrix is a former client, and they incorporated technology from former client Hi-Def Conferncing which they acquired.)
James and I heard each other like we were across the table from one another. But my colleague, on an iPhone over AT&T sounded very tinny, and almost mechanical to James, but not to me. One of James' colleagues calling via an IP phone, but like my colleague, was dialing in via a local dialup POP number provided by GoToMeeting in the USA and UK. James' comment about his colleague was even though he was one floor down, sounded "unintelligable." To me the other Truphone person was slightly audible, but not as clear to me as my colleague on the iPhone who to James wasn't.
Here's the problem. I call it ins and outs.
Ins and outs are the way calls connect to the conferencing bridge. It includes the number of hops and number of network handoffs. It's the routing and it's the conversion of codec mismatches. It's the Session Border Controllers from multiple manufacturers and most of all, it's the conversion from IP to TDM, back to IP, vs. the all IP calling that James and I were experiencing that was crystal clear and coming through in beautiful Hi-Def audio.
I'll compare this experience to calls on bootstrapped startup Voxeet, where the audio, over 3G, 4G or WiFi is amazing. No calls ever touch the PSTN. None are dependent on a SIP based PBX that has to dial out over the POTS line. Pal Craig Walker's company, UberConference, uses WebRTC inside Chrome. There, if everyone uses WebRTC and calls from the browser, the experience is much like using CounterPath's Bria and either ZipDx or TurboBridge, if everyone calls in via a SIP URI and has the g.722 wideband audio codec installed.
So the issues-the SIP providers don't peer with Citrix's Go To Meeting, so an IP based call that should stay IP, has to go in and out a few times. In the case of AT&T Mobile call, the dial in POP was not in Georgia, but in California. That caused latency in addition to going from and to and back between IP and TDM.
The problem isn't new, but it's getting worse and worse each month. It's a subject I've discussed in the past with Walker, as well as founders of Voxeet, TurboBridge and even with Alec Saunders, back in the day when he was leading iotum which operates Calliflower.
It seems I'm not alone in sensing the problem. Leading headset manufacturer, Plantronics is conducting a survey about conference call quality. Sadly, this is not an issue that is simply related to the input, but what's all going on in the middle. And that's something, the headset, or input device can't solve.