A pastor steps onto the platform while the band is still playing. A host closes the service over a final musical tag. Vocal mics sit in front of a live drum kit. A headset mic picks up PA spill, room reflections, crowd noise or the scrape of clothing. The engineer needs more level, but the room is already close to feedback. The old answer is usually some combination of mic choice, EQ, gating, careful mixing, stage discipline and compromise.
L-Acoustics Source Intelligence is aimed directly at that problem. Introduced as a licensed application running on the L-ISA Processor II core, Source Intelligence uses machine learning to identify the voice in a microphone signal and reduce what is not voice in real time. According to Ryan John, Director of Product Management, Software at L-Acoustics, the system can remove up to 40 dB of non-voice content from a vocal or speech microphone.
For churches, the number matters less than the result. Source Intelligence is not simply a feedback tool. It is designed to address the unwanted sound surrounding the voice: stage bleed, PA spill, room noise, crowd noise, drum wash, handling noise and reflections that make speech less intelligible and vocals harder to place in a mix.
John, who spent nearly three decades as a touring front-of-house engineer, says the idea grew from a familiar live-sound problem. Put a singer, speaker or host in front of a PA or near a loud crowd, and the microphone is no longer capturing only the person speaking or singing. It is capturing the environment around that person.
Source Intelligence gives church audio teams a new way to separate the message from the noise.
Many existing tools approach the problem by opening and closing the microphone, much like an engineer riding a fader: bring the mic up when someone speaks or sings, then pull it back down when they stop. John says Source Intelligence is different because it reduces non-voice content even while the voice is active.
“We spent the time to try and figure out how to build a tool set that removes the noise when the voice is open,” John says. “That is what’s notably different. So when the singer is singing, the noise is still gone.”
That distinction matters in churches because many of the hardest production moments do not happen in silence. They happen when someone is actively speaking or singing while the band, room or PA is still contributing sound around the microphone.
At North Point Community Church in Atlanta, Source Intelligence came into focus as part of a larger L-Acoustics L Series and L-ISA deployment. Clark, the Atlanta-based design and integration firm, worked on the project. JB Bowling, COO and Principal at Clark, says the system already delivered musically, but the spoken-word side still needed a little more margin.
“With the spoken word, we just wanted to be able to get a few extra dB out of the speech bus,” Bowling says. “We really felt like if we could just push it just a little bit more, Source Intelligence helped us get that extra dB or two that we were looking for.”
That extra dB or two may sound modest, but in a live room it can be the difference between a pastor sitting comfortably in the system and an engineer feeling like the mic is living on the edge. Speech microphones in churches are often exposed to several problems at once: reflected energy, ambient noise, PA spill and the feedback risk of putting a live microphone under or near loudspeakers.
Luke Roetman, AV Systems Applications Engineer with North Point Ministries, says Source Intelligence has given the team more usable level without changing the basic character of the voice.
“When [Pastor Andy Stanley] is on stage, it gives us several more dB of gain before feedback of just clear audio with no weird artifacts or anything like that,” Roetman says. “That is truly what has been impressive about it.”
For North Point, the benefit is not limited to the sermon. Roetman describes a common church production scenario: a pastor or host returns to the stage at the end of a service while the band is still playing. In the past, the production team might have asked the drummer to stop playing or simplify the part so the headset mic would not fill with cymbal, snare and stage wash.
With Source Intelligence, that conversation changes.
“It is pretty impressive what you can do when you’ve got a closing song and you’ve got Andy coming back out, or even one of our hosts come back out on stage with the drum kit, and they’re still playing,” Roetman says.
That is more than a convenience. In many modern worship environments, transitions are tightly arranged. The band may be carrying energy into or out of a moment while a pastor, worship leader or host needs to speak clearly over the top. Every time the audio team has to ask the band to stop or change a part for the sake of a microphone, production needs begin shaping the worship flow.
Roetman says Source Intelligence gives the team more freedom in those moments.
Source Intelligence is not simply a feedback tool. It is designed to address the unwanted sound surrounding the voice.
“I would literally say, drums can’t keep playing,” Roetman says. “I need keys only in this moment. I don’t need to ask him to do that anymore.”
The same concept applies to sung vocals. Opening several vocal mics in front of a live drum kit has a cumulative effect. Each mic may capture some drum bleed, and together those mics can change the character of the mix.
“You take four vocal mics that are in front of a drum kit at full volume, and all of a sudden you clean up the drums in all four,” Roetman says. “What that does to the mix is wild. It really, really cleans it up and doesn’t change the tonality of the voice at all.”
That transparency is a recurring theme. Bowling says the Clark and North Point teams tested the system by switching it on and off to see if listeners could identify a tonal change.
“It’s truly imperceptible until it starts, and then it builds its algorithm and starts cutting out all the noise,” Bowling says. “It’s so much more than feedback.”
Joshua Maichele, Global Application Lead, House of Worship at L-Acoustics, says that transparency is central to the product’s value. Many tools that try to control bleed, feedback or noise can affect the consonants and detail that make speech intelligible. Source Intelligence is intended to preserve the quality of the voice while reducing the unwanted material around it.
For churches, that means the technology is not limited to one worship style. A liturgical church with speech-heavy services, a contemporary church with loud stages and live drums, a multisite church feeding other campuses, and a broadcast-focused church trying to clean up PA bleed can all run into versions of the same issue. The voice matters most, and the microphone rarely captures only the voice.
That also makes the broadcast implications important. North Point currently uses a front-of-house mix for its online feed rather than a separate broadcast console, but Source Intelligence can still affect what goes downstream. Maichele says the technology can be equally powerful in broadcast because it gives engineers a cleaner source before they decide how much room to add back in.
In a traditional broadcast mix, PA bleed or room ambience in a pastor’s mic can fight against dedicated room microphones. A cleaner vocal source gives the engineer more control. Instead of being stuck with whatever ambience is leaking into the headset or lavalier, the broadcast engineer can choose how much room to include.
“Now we have a very, very clean vocal,” Maichele says. “And now as an engineer, we get to make a choice of how much of the room do we want to add.”
That choice can matter in multisite environments as well. A completely cleaned-up vocal may be ideal for some applications, but a campus feed may still need to feel like it came from a live room. Because Source Intelligence can remove up to 40 dB of non-voice content, teams can choose less reduction when the application calls for it.
The technology does involve latency, but L-Acoustics says total input-to-output latency with Source Intelligence is under 8.5 milliseconds, including conversion and processing. Roetman says that has not been noticeable on stage at North Point, even with a pastor who is sensitive to changes in sound.
The most important input in church audio is often the hardest one to protect.
It is also important to understand what Source Intelligence is and is not. It runs on the L-ISA Processor II core, but it is not limited to immersive L-ISA deployments. Maichele describes the processor as a platform that can run different licensed applications. Source Intelligence is one of those applications, and it can be deployed with different types of L-Acoustics systems, including mono, stereo, distributed mono or L-ISA configurations.
That hardware approach matters for churches considering show-critical audio processing. Maichele says Source Intelligence is running on a dedicated, closed L-Acoustics platform rather than a general-purpose computer connected to the internet and vulnerable to updates or software conflicts.
North Point’s team says Source Intelligence has provided additional gain before feedback on Pastor Andy Stanley’s headset mic without introducing distracting artifacts.
John says reliability was central to the product’s development. Source Intelligence was tested for more than a year and a half in the field, including on major tours, before L-Acoustics brought it forward as a product.
“For most scenarios, the voice is going to be the most important element in an event,” John says. “It will be the message, it will be the emotion, it will be the music, whatever it might be.”
That may be the most church-relevant way to understand Source Intelligence. The technology is sophisticated, but the goal is simple: protect the source that carries the message.
For church production teams, intelligibility is not just an audio preference. It is the difference between a message being heard clearly or fighting the room. It affects the sermon, the welcome, the prayer, the worship vocal, the online feed and the person listening from the back of the room or on a phone.
Source Intelligence will not replace good system design, thoughtful mic technique, appropriate stage volume, trained operators or acoustic judgment. But the experience at North Point suggests it can give church audio teams a new kind of margin in moments where margin is often hard to find.
And in live church production, a few more dB of clean, natural voice can be a very big deal.
