Thanks to the recent industry “pivot” towards livestreaming, I’ve seen a lot of forum posts, received a lot of phone calls, and talked to a lot of colleagues – mostly in house of worship audio – about how to achieve a successful livestream mix from front of house. Most of them boil down to three main issues, discussed here with my suggested approach for addressing each.
Problem: The FOH mix sounds tonally “thin” when broadcast to the livestream.
It’s very common for a typical live PA system to exhibit some amount of “spectral tilt” – that is, the response of the loudspeaker system itself is bass-heavy. On its own, this isn’t necessarily a bad thing, but the important takeaway here is that, when mixing through the house system for the people in the room, the PA response is making a tonal contribution to the sound of the mix. If you also route the board mix to other destinations (a computer, recording device, or a live broadcast), those destinations don’t get the benefit of the bass tilt that would be added by the PA in the room, and so sound “thin” and lacking in low-frequency energy.
When the mix will be routed to other places besides the PA, it’s helpful to tune the system to a more neutral target curve, reducing the bass tilt for an overall flatter response. This means the PA better preserves the tonal balance of the mix leaving the desk, and so the desired tonal balance can be achieved in the mixing console, with the PA simply reinforcing that mix in the space. The well-tonally-balanced mix should then translate better to livestreams and broadcasts.
Ideally, PA system EQ should be done in the system’s dedicated processing, but in a pinch can also be done on a console matrix. Routing the main left-right (LR) mix to a stereo matrix that in turn drives the PA allows the system response to be equalized without affecting the mix headed to other destinations.
An audio analyzer can be helpful in achieving very consistent tonal response from system to system, but this can also easily be done by ear – play a familiar commercial recording through the console, make sure there’s no EQ on the playback input or the main LR, and then adjust the LF EQ on the PA matrix – usually a combination of low-shelf and low-parametric cuts – until the playback sounds balanced and maybe even just a bit thin. Then put the PA matrix EQ out of your mind during the event, and simply mix to achieve a natural balance in the space. Cueing up the main LR mix in headphones should reveal a more tonally balanced mix overall, which will then translate better to livestreams and recordings.
Problem: The main LR mix is being sent to the stream, and it’s loud enough in the room, but it’s too quiet going into the computer.
This is a gain structure issue. A modern PA system, whether larger format or simply powered loudspeakers on sticks, has a considerable amount of voltage gain, which means that if it’s operated at “full tilt” or with the amplifiers set to “fully open,” it can achieve full output with an input signal far lower than a typical nominal +4 dBu line level signal.
Thus a relatively low signal level leaving the FOH console can drive the system to the desirable SPL in the space but is too low for the other destinations (audio interface, computer, streaming setup, etc). Relying on the voltage gain of the PA to achieve the necessary signal level leaves the other destinations short.
The solution is to set up the console gain structure so the mixes leaving the desk (to the PA, livestream computer, etc.) are achieving nominal line level (around +4 dBu, right around where most desks change from green to orange on the metering) with the peaks extending above that.
This may require going back through the inputs and turning up the preamp gains until the inputs are metering at a nominal input level (around “0 dB” on the meters for analog desks, or around “-18” for digital – again, around where the meter starts to change color). It allows a full proper line level to be achieved at the main mix bus, and the other mixes the signals are routed to (in-ear monitors, streaming feeds, etc.).
Then the inputs to the PA system can simply be turned down until the nominal signal level at the console produces the desired SPL in the space. This is most commonly done at the PA drive stage (DSP/system processor) but can simply be done by turning down the PA matrix fader (you are using a matrix to drive the PA, right?) by the appropriate amount, which is a quick solution I use all the time when I’m working in various venues without full access to the processing.
Problem: The balance between the various mix elements sounds great in the room, but the stream is lacking in some elements and has too much of others.
This is because the FOH mixer can take into account – and react to – acoustic energy from the instruments in the space. In a smaller room, the acoustic drum sound off the stage might mean that little to no drum reinforcement is necessary through the PA. Quieter sources like vocals and acoustic guitars will be heavily reinforced through the console.
When the board mix is consumed in isolation from the acoustic sound in the room, it sounds unbalanced. The band might be 20 dB higher than the pastor/speech mics in the stream, which might be perfectly appropriate in the space but is far from ideal for a livestream.
In larger scale productions, a discrete broadcast mix is done on a separate console by a separate operator. I sometimes see folks recommend a “broadcast-from-FOH” solution that entails double-patching all the console’s inputs and developing a completely separate mix on the same desk. This requires an enormous increase in the consumption of the console’s input resources as well as the operator’s attention and might be considered a somewhat inelegant solution.
I’ve had much better luck with a subgroup-to-matrix approach. Inputs are routed to a handful of subgroups – for example, Drums, Bass, Guitars, Keyboards, Vocals, and Pastor/Speech Mics. Not only are there some significant benefits to approaching a live mix using this subgroup-centric workflow, but it also makes achieving an appropriate balance for streams and recordings much simpler as well. The subgroups get routed to the main LR mix (and then to the PA matrix!) for the FOH mix in the space. The same subgroups can also be routed at different levels to a dedicated “Stream” or “Record” matrix.
We might start with all the band subgroups sent to the stream matrix at -18, with the Pastor/Speech group at unity. From there, adjust the band group sends up or down to achieve the necessary balance (typically drums up, vocals down). Making the group-to-matrix sends post-fader means they will naturally follow the mixing and balancing of the subgroups out front as the service progresses, but maintain their overall level offset necessary to balance the stream, meaning the operator’s focus is free to concentrate on what’s happening in the room, with less worrying about the stream.
Adjustments between different groups of instruments are made at the subgroup faders, and adjustments within a group (kick-snare balance or the balance between guitars) are made at the input faders. Some additional compression can then be added to the stream matrix if desired, but not much should be needed if the subgroup sends are reasonably balanced.