Yea, the challenge of a range of devices.
Short answer - there is no mix that will sound good for all of them. So... the common suggestion I came across and went with is targeting the vast majority of audience, so common to be a smartphone or tablet.
I found that if you target sounding ok on smartphone or tablet, it will also be ok (not optimal, but listenable) on better speakers. A common suggestion is to use compression. Think audio optimization for radio broadcast
Realize that most streaming providers (YouTube, Facebook, etc) will re-encode the stream you send them (typically H.264 due to licensing complications) then re-encode to a more bandwidth-friendly distribution encoding (ie much more heavily compressed). This can and usually does mess with audio and video settings.
- So, I record locally so I have a higher-quality local file to work with (ex, share with family for wedding, baptism, funeral, etc)
- And I'd start with making sure local recording looks and sounds good.
- Then, different providers handle video differently, so optimization advice will depend on your streaming provider. As for audio, stream at highest bitrate provider allows.
- Then, adding audio compression will probably help for the stream audience. Beware overtaxing CPU on streaming PC though, so be mindful of hardware resource (CPU, GPU, Disk I/O, RAM, etc) utilization
- Ideally your mixer has the ability to separately adjust for in-house amplification vs AUX output. The setting will be similar but likely a little different, especially for talking voice mics)
What I did was put OBS PC in closet with mixer (Presonus analog AR12 USB). I then ran 50ft powered USB cable, plus a DisplayPort fiber optic cable (using MST) to drive 2 monitors along with keyboard and mouse in choir loft. I'm currently using the mixer's analog AUX out, but once I have time plan to switch to using the USB audio out on the mixer, so I can selectively mute certain mics as needed (pipe organ/choir mic for stream only audience (ie not amplified, but only sent to AUX output) in back of Sanctuary creating echo from main speakers). The challenge is that many (most) DAWs don't provide an audio output to the Windows OS audio system... meaning a physical or virtual lookback adapter, or something like the reastream VST plugin.
Then, in general a digital Usher can be useful during a service (think combined greeter and usher). Having that person also monitor audio and be able to provide direct feedback to livestream OBS operator (ex. text msg) on volume level was important when we first started livestreaming our services. And that usher can then also comment on livestream platform if there are technical issues, freeing OBS person to focus on fixing issue vs replying/commenting to livestream viewers.
Glad to help out further, if I can
