One of the main features of the Stingray is that most of its microphone inputs are automatically mixed. The user is not required to adjust and set each microphone’s levels, changing how they contribute to the unit’s final output. This is typically a very tedious task that requires a lot of audio experience. Furthermore, the desired behavior of each microphone might change if it’s moved, or even if the acoustic surroundings of the room have been altered (new furniture, installation of absorbing materials, number and location of participants, etc.).

The advantage of eliminating this task is clear. It makes the installation easier and quicker. The question is whether the automatic mixing is good enough to provide a good overall performance, as we would expect from a room set up and tuned by an acoustic engineer.

To describe how the Stingray performs the automatic mixing, we use the term “Smart Mixing,” or “Distributed Array Technology.” These are terms we don’t use lightly. We believe that our mixing algorithm is smart and the result is as good as, if not better than, manually setting the parameters of the microphone. Obviously, the details of the technology are confidential, but we will share some of the core principals in this blog.

The basic idea is that our algorithm examines each of the microphone’s pickup quality. Based on this value, each microphone gets a “score.” We then sum up all of the microphone inputs proportionally based on their weighted score. In order to provide smooth and seamless switching, the score value of each microphone is filtered and smoothed, creating a fade-in / fade-out effect.

There are two challenges with this approach:

(1) Determining a microphone’s “score”

(2) Sharing this information along a daisy chained Stingray system (when using more than one unit)

To address the scoring issue, we apply three unique technologies to determine an individual microphone’s score. First, we analyze the spectral content of the pick-up signal to determine if it’s the result of speech, wide band noise, narrow band noise, or transient noise (door slam). Once we determine that the signal is speech, we use a technology that we have developed that measures the level of reverberation in this signal (we call it “Signal to Reverb Ratio”). Finally, the Signal to Reverb Ratio is combined with the signal RMS level (basically amplitude) to assign it a score. This approach is significantly different from just using the signal level, which is what most, if not all, other solutions offer. We found that in some cases, a stronger signal is the result of a higher level of reverb and is not necessarily a better quality signal. In practice we want to decrease the level of reverb added to the output and choose “clean” audio vs. “strong” audio that’s low in quality.

When adding into this equation the ability to daisy chain multiple units, the scoring system becomes more complicated. Instead of having just a single unit calculating its own microphone input scores, we now have multiple units doing this in parallel. However, this system will end up having only one output, so we must design a system that allows all the units in the chain to communicate and share their different microphone input scores. That’s exactly what we did. The different scores are calculated locally at each unit and then fed up the chain to the Primary unit. The Primary unit receives the scores from all the participating microphones and sends back, down the chain, filter commands to the individual Stingrays and the individual microphone inputs within each Stingray. These parameters determine the level of participation of each of the individual microphones. The individual Stingray will run these filters on its own inputs, add it to the filtered input received from the unit below it in the chain, and send the final result to the unit above it.

Sound complicated? The good news is that you don’t have to worry about the details, since it’s all done automatically. When you daisy chain multiple Stingray units, the devices handle everything for you. We’ve been implementing and perfecting elements of these technologies in many of our other products and are happy to say that they’ve been proven to provide reliable, accurate, and seamless results for years!