Taking on projects in retirement is a good thing. Taking on projects that expand your horizons is a great thing. Bonus points for projects that provide service for a community.
I have to admit, this one was forced upon me. And before retirement.
Remember Covid?
When the world went into shutdown, our Church had two options: stop meeting until the lockdowns were over or meet virtually. After much debate, we decided to explore the brave new world of virtual meetings. If you will recall, in 2000, the video conferencing landscape was very much aimed at the corporate world. Meaning that the primary focus was speech and video and not music. The primary constraint on the Church was budgetary: We had no budget!
We experimented with all the available platforms and finally landed on using Zoom. They were ahead of their time and, for our needs, the lowest tier (read: cheapest) would suit our needs.
In the beginning the congregation would meet from their individual homes with their individual setups.
Lesson 1 — never make assumptions about your constituents
Our average congregant was barely familiar with their computers/phones/tablets, let alone video conferencing. There was much consternation. I put together some basic presentations and held training sessions. The fun training sessions were the happy hour sessions. They started out as training sessions but evolved into much needed community gathering times.
It was an interesting time of re-imagining worship in a virtual world. Eventually, we found a comfortable format.
Lesson 2 — flexibility
We started out with the standard “family squares” format on Zoom. That was fine, but it was distracting. So we learned about the “spotlight” feature in Zoom and used it to focus on the current speaker.
We learned that Zoom considered all music as noise and filtered it out. Eventually they introduced the ability to remove “sound enhancements” and we trained our musicians on turning that off. In the fullness of time, Zoom introduced the “Original Sound for Musicians” option, and we learned how to use that.
Lockdowns were eventually lifted. But return to Church was delayed. The North Texas IcePocalypse of February 2021 caused pipes to freeze and burst, resulting in major damage to the building (fortunately the sanctuary and organ were spared!) It took until October 2021 to repair all the damage. There was much discussion as to whether we continue offering Zoom as a worship option. We decided that it would be beneficial for our home-bound congregants. Which meant we now had to go into Streaming mode, rather than just video conferencing.
Lesson 3 — streaming and recording
Our prime directive through this whole thing was to do it all with little or no budget. Which meant making use of open source or free software, using the existing audio equipment and relying on donated cameras.
Our setup
flowchart TD
%% --- AUDIO SYSTEM ---
subgraph Audio_System[🎚️ Audio System]
L[Lavalier Microphones]:::audio
P[Pulpit Microphones]:::audio
A[Phoenix Microphone Receiver]:::audio
M[Directional Microphones]:::audio
B[QSC TouchMix Mixer]:::audio
C[QSC Amplifier]:::audio
D[Speakers]:::audio
end
%% --- VIDEO SYSTEM ---
subgraph Video_System[🎥 Video System]
E[Camera 1]:::video
F[Camera 2]:::video
G[OBS]:::video
H[Zoom]:::video
I((Records Locally)):::note
end
%% --- AUDIO CONNECTIONS ---
L --> A
P --> A
A --> B
M --> B
B --> C
C --> D
%% --- MIXER TO OBS ---
B --> G
%% --- VIDEO CONNECTIONS ---
E --> G
F --> G
%% --- STREAMING CONNECTION ---
G --> H
%% --- LOCAL RECORDING COMMENT ---
G -.-> I
%% --- STYLES ---
classDef audio fill:#f9f5d7,stroke:#b58b00,stroke-width:2px,color:#000;
classDef video fill:#d7e8f9,stroke:#005c99,stroke-width:2px,color:#000;
classDef note fill:#e6ffe6,stroke:#009900,stroke-width:1px,color:#000,stroke-dasharray: 5 5;
We started with phone cameras, and the existing sound setup. Eventually, we added the Phoenix microphone system and the directional microphones as budget allowed and a pair of Fomako PTZ cameras were donated.
It was a gradual stepwise improvement of the setup.
OBS
Open Broadcaster Software®️ (OBS) is the most popular open source software for video recording and live-streaming. It is a platform for building broadcastable content from video and audio streams.
There are dozens of YouTube videos on setting up and using OBS. For someone coming into this with no prior experience (other than software development 😁) it was daunting.
OBS runs on a laptop which the “Worship Tech” uses to control the whole production.
Some basics I grokked over time:
- “Scenes” are containers that comprise audio, video, and graphics sources.
- We used scenes to change the camera point of view. We have one camera pointing at the pulpit at the front of the church, and another camera pointing toward the door at the back of the church. (There’s lots of movement in our Liturgy.)
- “Sources” define the components feeding into OBS. This is where we tell OBS what to connect to, how to connect to it, where in the scene to show it (in the case of cameras or graphics) and what audio to play in the scene.
- Finding the right balance of video/audio quality while keeping recorded file sizes reasonable took a lot of experimentation. We settled on a video recording Bitrate of 2500 Kbps and an Audio Bitrate of 160 using H.264 encoding for the video and AAC encoding for the audio. We used the Matroska Video (.mkv) recording format as it made it easier to edit the videos later (because it keeps video and audio tracks separated.)
- Video/Audio synchronization required us to introduce delays in the audio. The video frames were coming in roughly 500 ms ahead of the audio frames. The challenge there was that it was a moving target for a while. I suspect that this was due to some unrelated network issues that have since been addressed. Aside: in case it’s not obvious, there are a lot of moving parts in this endeavor!
Audio
There are also dozens of YouTube videos covering all aspects of audio recording. The first thing to understand was the concept of Gain Staging.
Long story short:
- Audio signal goes from a source to an output, via pre-amps, amps, compressors etc.
- Audio signal can be affected at every stage.
- Between the noise floors (what is generated when the source is not picking up anything at all) and the headroom (the audio level beyond which the sound is distorted) is the sweet spot of getting decent audio out.
- The trick with our limited budget microphones was to find a gain level high enough to pick up spoken word and music. We found that was quite a high gain.
- To compensate for that, we had to minimize the gains on every other step of the signal chain.
- The result was a balance of decent sound in the space (the sanctuary), and decent sound on Zoom.
- An interesting challenge was finding the right filters to filter out the ambient noise of the HVAC in the building and the noisy fans on the amplifier, without losing music fidelity. We found the Reaper Plugin, which hasn’t been updated in years but works for our use case. Aside: there’s a whole discussion about using the latest and greatest software vs. software that just works that could go here. We landed on: if it does the job, use it and don’t worry about it. Also: the price was right.
We ended up preferring audio levels that worked for the physical space and the subsequent recordings and not worrying to much about the Zoom audio quality since Zoom does weird things to audio anyway (see discussion below.)
Video
When we went to the Fomako PTZ cameras, video setup became much easier. We pretty much took the factory presets for output parameters, then used Network Device Interface (NDI) to connect the cameras to OBS. Then we just had to set up the presets for the cameras and then train the worship techs on scene and preset switching.
Video Editing
We decided that posting the service recordings on YouTube would be a good idea, both as an outreach tool and to allow members that couldn’t participate in the service live, as it were, to see the service in their own time.
Initially we just uploaded the raw recording up to YouTube after converting it to MP4. As time went by, we decided that a bit of editing would not go amiss.
Mostly this consisted of getting the audio levels to be consistent throughout the recording, cutting out long silences. Periodically, we’d also have to “fix in post” where a camera transition was missed, or where audio dropped out… Or any of a number of minor glitches that naturally occur in amateur recordings.
I tried most Open Source editing tools. All worked to some extent or another but were not particularly user friendly. Or at least: not total beginner user friendly. I discovered Black Magic Design’s free DaVinci Resolve. It too had a pretty steep learning curve, but there were significantly more/better training resources available.
Things I learned:
- Video editing requires a pretty robust machine and graphic card. It is a heavily memory intensive process. Of course, if one has patience (which I don’t) a consumer-grade PC would work too. But it would be a painful process.
- If audio and video are not synced perfectly, separating the audio and video tracks then shifting one or the other by however many frames works just fine. As we got better at getting the OBS time delay settings right, this became less of an issue.
- Chopping the silences (eg: waiting for readers to reach the lectern, or long breaks in the liturgy) is a matter making sure audio and video are linked again (if they were seperated in the synch step) and cutting the beginning and end of the silences and deleting them. Aside: the Studio (paid) version of DaVinci Resolve now has AI/automated ways of doing this. But we still don’t have the budget for that.
- Things that are obvious to professional Video Editors that weren’t so obvious to me:
- Your video editing all happens on the Timeline.
- To get a clip, or a graphic, or an audio track into the recording, you first add it to the media pool, then add it to the Timeline.
- You can have multiple video tracks on your timeline. There are many reasons to this. When we have to overlay graphics or text on a clip, we add a video track above the main track. (There are ways to do it directly on the main video track, but I found this gets confusing quickly.)
- The top-most video track plays over tracks below it. The top-most audio track plays over tracks below it. Of course you can adjust transparency or audio levels to get interesting effects.
- Keyframes are markers to tell the editor where to start or stop something. They can be used for a number of purposes. For example, you can do dynamic zooming by selecting a point to start the zooming and adding a keyframe. Then select a point to end the zooming. At the end point you set the frame to be the part of the video which you want to fill the screen. Black magic (pun intended) happens and the software determines a gradual zooming into the new frame.
- There are many ways to accomplish the same thing, learning all of them is a matter of either reading all the manuals and watching all the tutorials (HaHaHA!) or discovering them by accident. For example, I accidentally discovered that there is a builtin dynamic zoom feature
- It’s ok to experiment. Most editing tools have a pretty robust undo feature.
- It’s ok to not chase perfection. I’ve come to terms with the fact that I’m not putting together cinematic masterpieces, they are recordings of services who’s primary purpose is for our viewers to experience a version of the service, they are not expecting a big production!
If you’re interested in seeing our progress, take a look at our YouTube worship playlist to compare and contrast early recordings to more recent recordings.
