Sunday Service Streaming, trying to get audio from other capture device

[JacktheRipper88]

Hi folks,
I have a local church that I have been assisting with getting their church services streamed to Facebook updated and going.

I'm going to explain the hardware involved before describing the goal of what we're trying to do.

Streaming PC:
So on the PC that is running OBS we have 2 video capture devices in use:
1) A Magewell that is capturing the SDI feed from the pulpit camera as well as an analog audio feed from their audio soundboard ( collects microphones and CD player)
2) El Gato HDMI Video capture device that received video from the PC that controls the powerpoints and videos being played to the congregation on TVs

Powerpoint PC
Has 2 HDMI outs, one goes to the monitor sitting at the streaming desk and the other feeds into the El Gato for video & audio capture before feeding into an HDMI splitter that supplies the feed to 3 TVs in their sanctuary for the minister and congregation to view.

So as of this moment OBS has access to show the camera from the Magewell and encodes audio coming from the soundboard, works great. The El Gato can be accessed and the stream collects the audio input fine. But the issue we would like to overcome is that we would like to get audio from the soundboard while we are displaying the input from the El Gato. Basically we're aiming to make so that if we transisition the stream to show the powerpoint feed, the people watching the stream will be able to hear the audio coming from the soundboard in the event the minister (or guest presenters) want to show a powerpoint and talk over it at the same time.

At the moment I can't remember what the model is of the El Gato but it doesn't have an audio jack on it, its purely HDMI. I'm going to be there this afternoon (in about 30 mins-ish and will get model info then.

 

[AaronD]

HDMI audio extractor to the sound board. Then all of your audio comes from there.

---

Also for robustness, consider not relying on a passthrough on a machine that is otherwise invisible. If your streaming rig goes down for whatever reason, you don't want to kill your local displays.

And, you didn't say, but if you have multiple USB captures, make sure they're on different controllers. Not ports, but actual controllers. (internal hubs...) SD video is more than half of the total bandwidth of USB 2, and HD video is more than half of the total bandwidth of USB 3. So if you have multiples on the same controller, something has to give.

 

[khaver]

In OBS in the scene that just shows the El Gato PowerPoint video, place the Magewell video and audio source in the scene but place the video outside the canvas area so it's not shown in the scene. This will give you the soundboard audio from the Magewell source without the video.

 

[Lawrence_SoCal]

In OBS in the scene that just shows the El Gato PowerPoint video, place the Magewell video and audio source in the scene but place the video outside the canvas area so it's not shown in the scene. This will give you the soundboard audio from the Magewell source without the video.

I'm thinking there might be a better solution. Instead of routing the audio into the Magewell, which does help with audio sync, presumably, instead route Sanctuary Audio mixer straight into streaming PC, so it is its own Dedicated Audio Input, which you can use regardless of Video input. Just a thought

if following @khaver's approach, I'd seriously crop the video output so computer resources not used to render what isn't being shown (just in case hardware resource contention is a possibility... which it almost always is)

Or there might way, depending on OBS Studio version? of duplicating the Magewell input, and modifying one to basically be audio only? sorry not in front of OBS System to make sure that is possible, and if yes, how.

 

[AaronD]

...route Sanctuary Audio mixer straight into streaming PC, so it is its own Dedicated Audio Input, which you can use regardless of Video input.

That's pretty much what I said:

HDMI audio extractor to the sound board. Then all of your audio comes from there.

OBS is generally bad at audio. So try to not use it to process audio. Give it a single source that is already finished, and just pass that through unchanged. Nothing more.

Yes, it *can* take multiple sources and do *some* processing on them before mixing, and it sounds fine once you get it right...but it's clunky and hard to use. Much better to have something else do all of your audio work, and then OBS just glues it onto the picture and that's all.

 

[Lawrence_SoCal]

That's pretty much what I said:

I know ;^)
And, Windows OS worst option (vs MacOS and Linux) for audio handling

That said, there is the question of 3 audio sources - Sound System and 2 PCs (OBS Studio and PPT)... I'm presuming cameras NOT picking up audio? if yes, then that makes 4 separate audio sources

I agree on doing as much audio processing in sound system as practical. The challenge is
1. if older, especially analog, system perfectly adequate for Sanctuary use, but {edit: doesn't may not} handle things like Compression that really helps for typical livestream audience.
2. synchronizing audio to video (not hard, but trips up users not prepared to address)
3. audio path
I'm a fan of keep it simple, when practical. As such, I prefer in principle, to have audio coming out of mixer straight into OBS Studio PC (vs Magewell capture card for video... BUT, that may actually work out fine (or even best) in this case... outside my area of expertise to be definitive at this point). And you don't necessarily need complexity of HDMI to audio extractor... might be able to simply take audio output of PowerPoint PC into sound board... presuming available input. Though having to deal with AC circuits and grounding, if sound system doesn't have available input, might just run TRRS cable from PPTx PC (output) to OBS Studio PC (input)??

What I set up, we had PowerPoint and OBS Studio on same PC... made it a little trickier, needing a little more training and focused operator, but absolutely made things simpler from a setup, Inputs, etc... just a single PC (dual monitor) to work with... but depends on workflow and other details/expectations. At this location, have screens inside worship space during service would be near heresy.. only special setup on portable screen for Pipe Organ concerts so folks downstairs could see organist & keyboard

4. the final consideration is if there is reason to separately Mute or adjust the various audio inputs, AND the physical mixer is NOT right beside OBS Studio PC (as was my case). If there is a 'sound engineer' manning the sound system, great, but that is often NOT the case. In that case, having separate inputs into OBS Studio might be necessary so OBS Studio operator can control the various inputs

joy right, answer is 'it depends' ;^)

 

[AaronD]

you don't necessarily need complexity of HDMI to audio extractor

I didn't think it was all that complex. HDMI passes through unchanged, including audio, and another copy of that audio squirts out the side. Done!

Doesn't do anything for a ground loop: that possibility remains the same either way. And just because you have a digital signal, does not mean you're immune to them. As soon as you convert to analog, whatever ground loop noise you might have gets in *there*.
(This includes a USB connection to an otherwise analog mixer, which is simply a standard USB soundcard inside the case. If the USB connection completes a ground loop, then it'll buzz like anything else. Do that with an all-digital console, and it may not because it *stays* digital.)

older, especially analog, system perfectly adequate for Sanctuary use, but doesn't handle things like Compression

I know what you mean, but... ;-)

Mostly inherited from various church upgrades, plus a custom cart. (woodshop out of frame to the left)
40-channel console, 16 channels of outboard compression/gating in the left rack (2x ACP88), video soundtrack on the two matrices and then through a 2-channel compressor at the top of the right rack (Alesis 3630), and then to a Behringer USB line-in. Below that is a crossover for subwoofers, post-everything. Center rack has a 2-channel FX unit and 24-track recorder/playback. Between the rack and the console are 40x custom highpasses to replace the console's 100Hz 18dB/oct yes/no with no other options, and two of them towards the right (FX returns on full channel strips) are using their option for a lowpass too.

Pure analog, except for the FX unit, the recorder that does no processing whatsoever, and the soundcard to OBS that also does no processing whatsoever.

The picture is when I was still building it. It's now temporarily replacing a Midas M32 while we change some things around, and they love it! (some have asked if the M32 is going to sound like that after we switch back...) Video sounds decent too...after a couple of services to listen back and tweak the matrix sends for next service.

Anyway, back on-topic: :-)

if sound system doesn't have available input

Then I'd say you need a bigger board. :-) Nothing wrong with having unused inputs, but swapping plugs mid-service, inaccessible submixers, and bypassing the board altogether, all bring their own problems and ballooned complexity, at least in part because they create so many exceptions to something that could have been a universal rule.

If space is a problem (shortsighted design, or never intended for the present use in the first place), there are compact digital mixers that pack a LOT of channels and processing into a small space...with the tradeoff of only seeing a tiny little part of it at any given time. Unlike analog, which puts everything in front of you all the time, a paged digital board requires you to already know what else it does and what page it's on.

PowerPoint and OBS Studio on same PC

I can see that, for screen capture and looped-back audio inside the same box, for perfect reproduction with no external wiring. But I still see a big operational benefit to having a dedicated machine for each task, even if the same operator runs both/all of them. Conceptual boundaries like borders on a map, simultaneous cues without mouse contention, fault tolerance, etc.

And like the HDMI audio extractor, I didn't think a physical wire from GPU out to Capture in was all that complex. Even if there's a multi-in/multi-out matrix switch/splitter in between.

Conceptually, "this machine feeds that display." "Okay, got it." "Now tap off of that signal, and also feed it to this machine to put in the corner of the frame, and here's the wiring that does exactly that, verbatim..."

If there is a 'sound engineer' manning the sound system, great, but that is often NOT the case. In that case, having separate inputs into OBS Studio might be necessary so OBS Studio operator can control the various inputs

I'd call that a strong case for a remote-controllable digital mixer of some kind. FoH Engineer or not, the Broadcast Engineer can still remote control that separate mix in the console, and so everything stays routed through there.

I once had the idea to mix the video sound on the main bus, and FoH on some post-fade aux'es, with the reasoning of:

• It's actually FoH that needs the "weird" mix because of stage spill and acoustics, while the video is actually the "sane" one.
• The video is pretty much always stereo full-range at that point (no center or sub), which matches most main busses, while FoH is, again, a weird combination of things that can easily be accommodated by a series of aux'es as needed.

The crew rejected that idea almost immediately.

I guess the common thread here is, "*operational* simplicity, at the eager expense of some additional boxes and wiring if necessary." All the controls for each concept are in one place, and each concept has its own separate controls. But hard-reject something that promises to do EVERYTHING all on its own, because it very likely won't (designed for someone *else's* "everything", not yours), or will fail at some point as an unfixable black box. Make it obvious, or easily intuitive once shown, how it has to be, so that troubleshooting becomes easy too...for someone who doesn't demand a brittle list of specific mouse movements and reject any semblance of actual understanding.

 

[Lawrence_SoCal]

I'd call that a strong case for a remote-controllable digital mixer of some kind. FoH Engineer or not, the Broadcast Engineer can still remote control that separate mix in the console, and so everything stays routed through there.

Certainly.... BUT, many of the HoW I assisted had little to no money for livestreaming, and a VERY limited existing setup (no PTZ even). So an older simple 4-8 channel analog mixer was not uncommon (... though to be fair, I was not working with evangelical denominations, setup with a band, for the most part.) Some (small) places didn't even have a sound system at all (not necesary).

The setup I started with had a older 8-channel mixer in a closet, that was usually closed during service (set once, done, adjust only on rare occasions) as FOH setup did NOT need adjusting. I did finally buy a replacement digital mixer for that location, which could be remotely controlled, but many of the places I consulted with could not afford such an upgrade. Many were operating on starting livestream just before service started, on a fixed position camera, an no livestream operator at all. My setup required a single person to handle OBS Studio/livestream including Digital Usher, PTZ camera controls, and audio. That took some real training and a specific personality type to handle, with benefit of limited volunteer staff required to run system, with a downside of really limited pool of people who could actually operate such a setup. Depends on one's budget, pool of talent and willingness of volunteers.
Just saying, the operating realm varies, and with it the appropriate solutions. There are plenty of ideal solutions that aren't practical for lots of HoW.

I was fortunate in that donations covered cost of initial setup, which was then VERY well received, and more donations came in to handle some upgrades (like extra (used) PTZ cameras and the new audio mixer)