Sound for live event streaming - processing?

[TomHoward]

Hi

 

We've been doing some live event streaming recently as you might imagine, plays, concerts, church services, etc.

We generally will have a couple of rifle mics / ambients etc and the sound is okay for when the loud stuff is happening, but keeping an eye on the levels whilst also mixing for the room is getting a bit tricky.

 

With radio studios I've seen FM processors before for some multi-band compression and I think they might also do audio/loudness levelling?

 

My biggest problem is between the songs - we did a Nativity yesterday and the audio in the songs was fine, but the quiet lines between the songs are lost. I was using headphones in the room, and though I'd boosted it plenty, but looking at the meters at the time and the stream afterwards there is just a massive different in loudness.

 

The streaming hardware is Blackmagic ATEM so I could use some of the processing in there, or add an FM processor between desk and ATEM. We're going stereo line in to ATEMs and using no camera audio. We are streaming via OBS so I can also purchase / add any available VST plugins at that stage too if there's anything that might help there.

 

Any tips of how people are doing it gratefully received.

[timsabre]

For shows with spoken dialogue and music I would always have a compressor on the recorded audio. A basuc Behringer analog thing does the job. There's probably a plugin for OBS. Edited December 12, 2020 by timsabre

[p.k.roberts]

If you looking for a free VST to try out, then there's a multi-band 'broadcast' jobbie here;

 

https://vst4free.com/plugin/548/

 

but I would have thought that if you're trying to boost tiny voices in an ambient space (you mentioned a Nativity)then it's probably going to sound pretty grim once it's done it's job.

[bruce]

I’ve not used it for the purpose that you describe - but I do various things with community radio, and whether it’s in the studio or on an OB, you need a way to “tame” the audio - whether it’s someone who has turned the gain up to 11, or someone who whispers.

 

My “Swiss Army Knife” for this sort of thing is the Audessence PodBlaster/ALPS range. They’ve all got the same processing inside, the main differences as you go up the range are the external controls and interfaces. It’s got gating, limiting, levelling, compression, and it sounds fine.

 

I/O via analogue, or s/pdif

 

The entry level “Podblaster” is great, cos there’s no external controls at all - it’s managed via USB, so no opportunity for people to fiddle. I’ve found that it will cope with most things that I’ve thrown at it!

 

https://audessence.com/audio-levellers/

[GaryNattrass]

OBS has a compression plug in audio filter and it seems to do a pretty good job of keeping things under control.

[TomHoward]

Thanks for the pointers..

 

So, for example for the one we just did, the audio songs were pre-recorded in a better space and played back into the stream due to the excessive "ambience" of a sports hall, and played back into the space for a little bit of 'live' singing to be added.

I really could have done with some mics upstage and closer to some of the voices for the spoken parts - it's certainly true they wouldn't sound great once boosted.

 

When using a straight compressor to provide some levelling, are you clamping it down and using a big make-up gain for those quiet bits? Or relying on one with some auto-gain for the make-up?

What kind of time settings do you use for levelling audio rather than limiting it as such - particularly with regards to the make-up boost needed?

 

For some concerts, we'd have a 'proper' X32 running into the ATEM, so I can do some processing at the output. For the smaller jobs (eg Nativity in a sports hall) I'm using a EPM12.

I think I need some decent meters maybe on the output of the desk to confirm average level and try ride it similar... in my headphones at the time it sounded okay, but watching it back I've got to ride my laptop volume level up and down quite a lot to hear the quiet sections.

[dosxuk]

I've found a proper loudness meter to be invaluable for getting proper streaming levels right. I've now got my own one of these - https://m.thomann.de/gb/tc_electronic_clarity_m_stereo.htm - coupled with a blackmagic SDI-HDMI converter being used to output AES to give me proper metering from the any SDI feed.

 

On the mixing side, it's all down to compression strategies. It feels wrong coming from a live sound perspective to stick so much gain and high ratio compression on, but it's what I feel I need to keep things in the right region on the stream. (I'm also hoping for some tips and tricks on this thread as I'm never entirely happy!)

[Dave m]

The ATEM software is actually bloody good.I tend to think the issue is that you can't hear what you are doing?

You can run another computer on the network and monitor the sound in a better environment - that would help a lot at minimal cost

[J Pearce]

We've just streamed 3 theatre shows. Ours had designed sound from a sound designer, so perhaps more controlled than your shows/events? We had a 2:1 compression on each mic, set to start cutting in on anything above natural speech level, 1.1:1 compression across all the mics with the threshold right down at -60dB, and then 2 stage heavy compression and limiting across the whole audio bus (once the designed sound had joined the mics) - something like 6:1 at -10dB and a limiter at -1dB. Everything except the limiter was soft knee and slow release to avoid any pumping artefacts.

 

We were mixing to a dynamic range of around 20dB, that was covering everything from atmos/underscore up to some rather loud battle effects and angsty screaming - what would probably normally be more like 40dB of dynamic range acoustically in a venue.

[TomHoward]

Ours are certainly not sound designed, they are “classical” / varied concerts, nativities, etc. Usually we’ve had lectern mics for readings or close mic’d performers but this one was ambiently mic’d and it was very difficult.

I think the loudness meter would be a big help. A slight complication is at the moment I am monitoring from the desk before it goes into the Atem. The atem does have some possible dynamic processing.

 

I think the problem is as I am listening in the room wearing headphones I boost the mics and think “I can hear those lines, that’s fine”, but the listening environment on a laptop at the viewing end is much less forgiving. I think the level meters mentioned might be the way - I wonder if there is a VST one to save cost. There are meters in OBS but they are momentary so a quick glance doesn’t really give the full picture.

[J Pearce]

We had our sound designer working remote monitoring exactly what the stream heard.

It’s going to be tough doing that in the same room as the performers, but in ear monitors might get you much closer?

 

FWIW, our main bus compression and limiting was done in our ATEM Mini, the rest in the desk. The dynamics in the ATEM Mini seem to be fairly transparent and ideal for ‘mastering’ compression.

[TomHoward]

Thanks.

 

I’m thinking about doing something convoluted like using the two line inputs of the mini, and maybe routing some channels to one and some channels to the other, and compressing the life out of the speech / ambient input and leaving the musical ones with a bit more life. Combined with a good metering VST if I can find one. It seems to be me underestimating just how much gain increase I need between the songs / loud sections.

 

My monitoring is DT150s, I think the problem is my listening environment is a bit too optimum and I can hear it clearly when it’s still too low. Sounds daft but an expander on the headphones maybe might give me a cue when I’ve hit an acceptable level.

[timsabre]

I think people are used to tv sound being very compressed so they do not expect much variation in volume. Having good metering to keep a consistent level will probably help a lot.

[J Pearce]

When we stream our classical concerts we do it fairly uncompressed, because our audience expects a hi-res outcome. Few of our streaming audience for music concerts are watching on device speakers, they’re on headphones or hifis.

 

I wonder if you actually need to approach it from the other direction, and drop the louder stuff as much as you raise the quiet stuff, rather than raising the quiet moreso? A safety compressor and limiter to catch anything that slips past you is wise for avoiding clipping.

[TomHoward]

These certainly aren't classic concerts in the proper sense, this is all school / college level and the audience are parents / family. I suspect the devices are mainly laptops, phones and TVs at best.

The concerts are often a mix of live solo instrumentalists, trios, readings, and pre-recorded videos of much the same content.

The church services also have musicians and choir but the speech in those is from a lectern with a mic split so it's a stronger level. It's basically the content from PZM / rifle mics that I'm struggling to judge getting audible level on a stream when mixed with this other content.

 

I've realised what I need to do is download a copy of the audio off the stream I wasn't happy with and see what works to process it. I might try a VST loudness meter and see how it reads across the recorded content or there's a Waves playlist rider plugin that could ride level, or I'll try doing it with compression also.