I always knew that manufacturers provided their impressive, incredible device specifications under the most ideal conditions. So, while it may be possible to attain 110dB signal-to-noise ratio with 0.001% THD for a single mixer with perfect power, no headroom, a perfect source set to the optimal gain, with the settings optimized for this particular device specification measurement, no user would ever use it like that.
In reality, under nominal conditions with typical power, adequate headroom, installed cables and typical sources, most mixers can provide upwards of 60dB signal to noise with a less than 0.5% THD. That is perfectly fine for most conference room and classroom applications. However, I always thought that a technician should be able to get close to the performance specification numbers (110dB) while benching the device. So, since we don’t have a laboratory, we should be able to measure 90 or 100dB signal to noise with 0.5% THD. I’d be fine with that.
In one particular example, no matter what signal we passed through the mixer, we had THD of at least 5%. We tried several input voltages. We made sure we were nowhere near clipping in the GUI. We even took the processing out of the mix and virtually connected the input directly to the output. No dice.
We then called the manufacturer about it, thinking that we just didn’t understand the way the processing works. They assured us it was a failed device and sent us out an advanced replacement. Same results. So, we got another one from a different source. Same results.
Then, we took a closer look at the specifications. It turns out that this particular manufacturer wasn’t claiming that the mixer had the ability to pass a signal with a 110dB signal-to-noise ratio with 0.001% THD, but the input card for the mixer had the ability to pass a signal with 110dB. The output card had a similar spec. So, the device specifications weren’t describing the device as a whole, but the components that made up the device.
I used to poke fun at designers who just copied the performance specifications from individual devices and expected those numbers to be applied to an entire system. In today’s DSP world, manufacturers are doing the same thing. If we only know how well individual pieces of a device are supposed to work, how are we supposed to know if the device as a whole is working? As AV professionals we need meaningful, measurable specifications. If, as a manufacturer, you provide specifications that require board-level metering to verify performance, I would say those specifications are worthless (and would most likely void warranties). They look great, but they are meaningless to those of us in the field. And, as great as they look, you are opening Pandora’s box: since they can’t be verified, it automatically results in an RMA, wasting the manufacturer’s time, the integrator’s time, the designer’s time and the client’s time.
Give us specifications that actually apply to how the devices are used and, while we’re at it, specifications to confirm that the device has been configured properly.
What have you AV people seen out there?