Scroll Top

ESCO Case Study – Speaker Tracking

As we know, speaker tracking utilizes sensors that work out various algorithms and processes to focus on the speaker of a scene. It had gain recognition in recent years due to applications such as automatic PTZ (pan, tilt, and zoom) during a conference call. In this case, instead of relying on sensors, we used several elements and DSP as the catalysts to enable the tracking. It is not as seamless as the traditional speaker tracking created by the manufacturer, but this allowed us to integrate third-party equipment to create a solution replicating the function. Join Leslie Teo, Assistant Manager, Engineering as he brings us through the set-up of one of ESCO’s recent installations.

In this system, we are looking at a control-centric room system. It consists of various elements, a control processor as the brain of the system, 16 table microphones with designated spots to ensure participants in the room are audible and intelligible throughout the conference, a DSP (Digital Signal Processor) to process and elevate the sound quality being transmitted to both the far end and near end, three 4K resolution cameras to capture participants who are speaking in the room and a camera controller to facilitate the changes in the cameras by recalling presets.

Combining all the above together made speaker tracking possible in this solution. To start off, the microphones will have to process the audio captured in the surrounding, with multiple participants speaking, the DSP processing the audio will prioritize the speaker by simply tracking the speaker who is the loudest in the room. Through this process, the prioritized microphone will generate a control signal, and the cameras with the relevant signals will recall the preset for the camera to focus on the speaking participant. As mentioned previously, the table microphones are in designated spots to ensure participants speaking will be “tracked” when speaking. The presets of the cameras are made to focus on these designated spots to replicate the speaker tracking function.

Formulating these logics and functions isn’t a walk in the park. It took numerous tests in both the pre-installation and post-installation phases to get it working. We had to factor in the ambiance noise generated by the A/C to create a proper gain structure to ensure the dynamics we are using in the DSP work as intended. Minute details in the components used in the design greatly impacted the outcome of the functionality. Every DSP differs from each manufacturer, the functionality of the components used in the DSP is also different from one another. Finding the right components to use in the design was quite a feat. Plenty of trials and errors went into selecting the components to get the right function.

Thankfully, our engineers are always up to date with their proficiency with the solutions that we deploy. They worked together to achieve the results both client and ESCO are happy with. This is a milestone for us at ESCO. I believe this isn’t a small feat to integrate various equipment from various manufacturers to achieve the result we currently have.