Multimedia production is a double-edged sword.
Audio feedback, during a church service, is an excellent example of how the medium can hijack a program.
On a grand scale, media is designed to be mass-consumed. The approach is always to maximize potential viewership.
As churches and ministry organizations, the goal is to promote and connect, using media to impact people’s lives for the better.
Inversely, from a production standpoint, we never want the media to become the focus, meaning we need to do everything possible to keep the medium transparent, with the hope that the content always remains front and center.
Church production is a choreography of audio, video, lighting, fixturing, and graphic presentation (to name a few aspects).
In a best-case scenario, the viewer of each element of production, whether in-person or participating via broadcast, should always contribute to the prime directive (please pardon the Star Trek reference), ensuring that the goal and purpose of a worship service or event is fulfilled.
Audio feedback, during a church service, is an excellent example of how the medium can hijack a program. We’ve all experienced this: a pastor is speaking, or a worship service is progressing beautifully, when we’re attacked by a low frequency feedback loop. The focus of those there or those listening is immediately no longer on the content, but on the medium.
The same principle applies to video streaming. When viewers connect to a stream and develop a connection with a church, it is exceedingly counterproductive when perceivably minimal issues remove the focus from the content.
These are the issues we need to work hard to prevent.
Video streaming could alternatively be described more broadly as audio/video streaming. It’s rarely termed to represent both. In fact, 99 percent of the time, “video” is the reference term. This is understandable, because video transport requires the most resources, at all levels, and is obviously the most visible, but we absolutely cannot forget the importance of audio as part of the package.
I recently set up an elaborate, “high end” video encoding and distribution workflow for a large, multicampus church.
Immediately upon going live that first time, I must admit, I was blown away with the quality of the image. It wasn’t just a subtle improvement from the previous system, the image quality was glorious!
However, the image quality alone didn’t complete the project.
Working with this group over time, I quickly became aware of a shortcoming in their audio production for streaming. The volume was consistently too low!
While we were presenting a stunning visual image, the people struggled to hear the content, which was exceedingly invalidating.
This is what I’ve learned:
While it’s nice to have the best image in the business, without a proper audio mix, we’re wasting our time.
Mixing audio for video streaming is an exercise in the discipline of understanding how our viewers consume the stream and making sure we are producing a quality audio signal that the consumer can actually enjoy. It sounds simple, but it isn’t.
In most workflows, we have control of one or two settings, while as many as three adjustment points exist on the viewer side. We can control the level coming out of the sound system. We can control the audio input to the encoder.
At the same time, we are unable to control the volume level on the media player, in the viewer’s browser. We aloso cannot control the volume level on the viewer’s computer or device.
The same applies for the volume on a set of desktop speakers.
The onus is on us to produce a quality audio stream and sound level, mastering the points that we can control.
The first stop along the audio journey is the microphone, feeding into a mixing console, which is typically sending line level video to the encoder.
More elaborate workflows may include a dedicated split and audio mixer for live recording and streaming, which is optimal.
Whatever the case, it is critically important that the “mix” sent to the live encoder includes all the visual elements in the service (especially the person speaking), and that you are providing enough “level” for the encoder to capture and to be sent out to the viewers.
Let’s resolve to obsessively monitor audio levels at the encoder. Most encoders, both software-and hardware-based, provide audio metering and adjustment settings (see example).
In preparation for each broadcast, include ample rehearsal time to check audio levels, from the initial source, making sure the signal isn’t peaking, which causes unpleasant distortion, or sinking too low, making it difficult or impossible to hear.
As a best encoding practice, Netflix recommends: “Maintain +18 db (-2 dbfs) maximum level (true peak) over reference of -20 dbfs, achieved by peak limiting and not lowering the mix level.” This is an excellent reference for streaming, meaning the audio that is passing through your encoding workflow should not exceed -2 db, and should remain above -18 db (not including actual silence in the room). To achieve this, I would set my encoder at about 75 percent, working with the audio engineer in adjusting the level from the mixing console to get close to this range.
Example from Telestream Wirecast:
Maximum Level -2 Db (left) and Minimum Level -2 Db (right)
When the audio level is set in the encoder, [obsessively] monitor this meter throughout the broadcast. Make it a point to adjust volume based on the content with the goal to stay within range for the entire program. You may want to ask the audio engineer to reduce or increase level from his console, at any point in the broadcast.
It’s also a good idea to make adjustments on the encoder – just be sure you’re paying attention to this detail. Do not ignore the audio!
Another frustrating issue is audio-video sync. Video requires more time to process than audio, and in some cases, it can arrive at your encoder a few milliseconds sooner that the video. This sort of latency can quickly turn your pastor into the form of a ‘70s B-lot karate film star, causing an immediate distraction!
Fortunately, some encoders offer adjustment for audio delay (i.e., OBS), which should be considered when the audio is not aligned with the video in the resulting stream or recording. Another solution is to use a simple audio delay component between the audio source and the encoder, like the AD100-M from Datavideo.
When we have completed our job, making sure audio level is in range and on time, we gain the ability to support our viewers in a credible capacity.
When the audio level drops, users will manually increase their volume. Because our church live streams are typically interactive, murmurings and complaints will always filter to your attention when audio problems arise. Viewers will be confused with the level in their player, on their computer, or even on their speakers.
When we are confident that our audio level is in spec, we will receive fewer complaints from our viewers and we will stand credible, responding with a confident “set your player, computer, and speakers to 50 percent”, which will most likely resolve the issue.
Mastering the audio mix will become routine over time.
Let’s get good at this, keeping eyes on the content, not on the medium.
The best feedback is no feedback (pun intended).