Checklist Items Under Test: Several, really, but for this story, 6.33: For conferencing mode, at the 65dB-SPL listening level, be able to demonstrate full duplex operation, with no reports of echo or “speech trails” (as detected from the far end).
Reasoning: Most of the time, when a system doesn’t work, the manufacturer’s device is not configured properly. People are quick to blame a box, yank it out and put in a new device from a different manufacturer…only to find that it exhibits the same issues. However, sometimes, it truly is a shortcoming of the manufacturer. Having the training and test equipment to know when the system is properly configured and still not performing can save a tremendous amount of time and energy.
The Story: AVR typically will continue to test a system until it has zero defects, and can be officially considered commissioned. It usually takes about two visits to get systems to this point but, if the integrator is not familiar with the AV9000 standard, it may take more.
There was one project, however, that was just dragging on and on. It seemed like every time we went to the site, the conferencing audio got worse. Duplex operation (“double talk,” where both ends talk at the same time and can hear each other naturally) was terrible, and the far end would sometimes hear themselves (line echo, not acoustic). Clearly, the system was not ready for primetime.
As much as we try to be objective all the time, sometimes that subjectivity rears its ugly head. After the fourth visit, after patiently explaining to the integrator how the site file should be configured, after watching over his shoulder while he fixed a few rooms, and still not getting acceptable audio in the other rooms, we began to think there was something wrong with the integrator. Then, he pulled the classic “we don’t know what we are doing” card: He blamed the manufacturer.
Not only did he blame the manufacturer, but he actually claimed that the manufacturer knew about the issues and told him that the device was supposed to work that way. Although it was true that the design called for a device from a manufacturer we don’t see too often, I have tested more than 1000 systems over the course of 12 years. I’ve never heard of a manufacturer claiming that its box is supposed to have poor audio. If mixers from 12 years ago provided great audio, this current mixer with the latest processing chips should, as well. Clearly, there was a configuration problem.
We reviewed the site file and gain structure with the integrator. There were some levels that could have been optimized a little better, but for the most part, it looked great. The references were all mixed properly with good levels. The signals were strong for all microphones and conferencing far ends. Audio was hitting the telephone hybrids and videoconferencing codecs at their sweet spots. However, the audio was still not acceptable. It was also telling that the acoustic echo was very effectively canceled. The issue was with the duplex operation and random line echo. The integrator suggested we talk to the engineer from the manufacturer he worked with…you know, the fabled “our box is supposed to work this way” guy. I was surprised he offered to call a fairy tale, but I wanted to hear what he had to say.
“Oh, yeah. It’s supposed to work that way.” What about the line echo? Isn’t that an impedance mismatch? “No, sometimes that happens.” What about the poor duplex, where one end will almost completely cancel the other? “Oh, we call that “see-saw duplex.” ” So, in your opinion, this device is configured exactly as it should be, and operating as intended by the manufacturer? “Oh, yes. We worked with them all last week.”
I was dumbfounded. I wasn’t expecting that at all. I was embarrassed that I thought ill of the integrator. I was impressed that the manufacturer was able to create a “feature” out of a shortcoming (“see-saw duplex?” Really?). I felt sorry for the integrator because he was faced with an impossible task: get something to sound great that wasn’t designed to work.
Because we were not hired to review the design, and only to verify the installation, things went much smoother after the call. We confirmed levels. We confirmed conferencing was possible, albeit with the dubious duplex operation. We confirmed that the systems were operating as best as the design allowed. We confirmed that the integrator did everything he could to best serve the client. We recommended to the client that the mixers be replaced with ones that offer better performance, but that this replacement would come at a cost. It turned out to not be worth it to them.
Here’s the kicker: 20% of the rooms, for whatever reason, worked great. Duplex operation functioned flawlessly and they never had line echo. There was no apparent reason for this (room size and finishes were similar, settings were identical, microphones and loudspeaker layouts were similar, firmware was the same, etc.). It would have bothered me less if none of the rooms worked, but the fact that some did stole a few hours of sleep from me.
I took a few things away from this experience. First, sometimes manufacturers just offer a poor-performing device. Second, if a model can work, that fact might be the exception, not the rule. Third, newer isn’t always better. Last, it is so important to stay objective in this complicated and demanding industry. If we (both AVR and the integrator) did not have the patience, experience, knowhow and testing equipment to confirm that it was indeed an issue with the mixer and not the configuration, I don’t think that job would have ever ended. Everyone is served well by having a testing process and the ability to put a number on things. Once that is possible, the issues can quickly be located, the resolutions can be reviewed with the client and the projects can be put snugly to bed.