When it comes to a church service, or any spoken word event, vocal intelligibility is king.
Because it’s all a surprise. Everyone in the audience is hearing the words for the first time.
Imagine if we tried to have a conversation while standing on the opposite ends of a giant cathedral. It would be impossible. On the other hand, if I sang a song, you could still enjoy it and if you knew the song, you could sing along.
[This is a guest post by my friend Nathan Lively]
While music is all about tonal balance, speech reproduction requires a higher signal-to-noise ratio, especially for consonants.
Vocal intelligibility (aka speech intelligibility) describes how much information is communicated clearly to the audience. In the past, this would be tested with an audience transcribing speech. The result could be scored as a percent of correctly understood words.
How do we score high on a vocal intelligibility test?
By delivering the waveform from talker to listener with as little change as possible. This requires optimization of every link in the signal chain, but today we will be looking at speaker placement only.
To improve the vocal intelligibility of your church service, avoid these three speaker placement mistakes:
- No line of sight.
- Speakers are too low.
- Speakers are placed far from the center of their coverage areas.
#1 – No Line of Sight
If anyone in the audience cannot see a speaker, they won’t hear it.
- Wait, can’t sound waves bend around objects?
Yes, but only if they are big enough. Consonant sounds like “S” will only diffract around objects smaller than 3 inches. The larger the object, the more high frequency reduction.
The distance of the speaker from the obstruction and the listener also matters. A 3-inch piece of pipe 20 feet from the speaker while you listen at 40 feet might not be noticed, but 2 inches away from the speaker while you listen at 3 feet definitely will. The closer the obstruction is to the source, the bigger the problem. This is the danger of hanging speakers inside of a scaffold or soffit that may have nearby obstructions. I have also seen speakers trying to play through truss or a lighting instrument.
If you can’t fix the placement, you might at least be able to flip the speaker so that the high-frequency driver is unobstructed.
Need proof? Get your audio analyzer out. These are things you can measure. Hold an object like your phone 2 inches in front of a speaker and take a measurement. Then try moving the object 1 foot and 2 feet away from the speaker and compare the frequency response and coherence.
- But Nathan, human heads are wider than 3 inches! Aren’t those blocking line of sight to the speaker?
Yes, they are, which brings us to the second speaker placement mistake…
#2 – Speakers are too low
How high should your speakers be? Probably higher than you think. This won’t help your sonic image, but it will reduce the level contrast from front to back.
Speech intelligibility requires a signal-to-noise ratio of at least 25dB around an octave centered at 2kHz. What does that mean? In practice, you turn it up until it’s loud enough to be understood above ambient noise, but not painful.
The FOH mixer does by instinct, but for them to be able to do that well for the entire audience, the entire audience needs to hear what the FOH mixer is hearing. To improve level distribution we look at reducing the front-to-back distance ratio.
Take a look at this speaker 5 feet high on a stand. The front-to-back (F2B) ratio is 9.4ft to 42ft, which can be reduced to 1:4.5.
That’s a difference of 13dB from the first row to the last. Would you mix two vocals in a duet at 0dB and -13dB? I don’t think so. And that doesn’t even count all of the loss from trying to play through human heads. Here’s a prediction at 4kHz. Note the colors show the drop in sound pressure level amplitude.
Once we get the speaker 3 feet higher into the air we have an improved F2B ratio of 1:4.1.
That’s a F2B difference of 12.3dB. This doesn’t seem much better at first, but now we’re not playing through human heads. Everyone in the audience can see the loudspeaker.
This, combined with some room gain, and we might be able to make it to the last row in under 6dB. Let me show you what I mean. I’ll turn on the wall reflections in this prediction. Now, instead of loosing 12.3dB, we’re only losing 7dB, which is a more realistic estimation.
How do we know if it’s good enough? It would help to have a goal. The goal that we use in Pro Audio Workshop: Seeing Sound is to keep the F2B ratio below 2, if possible, but definitely below 4. This is the extreme limit.
Let’s see if we can do even better by moving the speaker up to a height of 20 feet. Now that we have reduced the F2B ratio to 1:2.3, we should definitely be able to keep the loss below 6dB.
A prediction at 4kHz reveals that we do.
Remember, the only silver bullet in sound system design is headphones and e-drums. Everything else is a compromise. While improving level variation we have raised the potential for excessive reverb. #tradeoffs #triage #tanstaafl
If you’d like to play with this design yourself, please download the MAPP XT file. If you’ve never used MAPP XT before, it’s a free 2D design app from Meyer Sound.
#3 – Far from center
One of the most common causes of problems with speech intelligibility is excessive reverb. We can reduce reverb by placing our speakers to focus on the audience and avoid the walls.
Speakers want to be near the center of their coverage areas. The farther you move them away, the bigger the compromise.
I say coverage areas instead of audience, because many times the audience will be sub-divided and each speaker will be assigned sole custody of a division, depending on audience shape vs speaker coverage shape.
The mistake here is in thinking that pushing your speakers out to the walls will improve gain-before-feedback and speech intelligibility. In practice, you won’t cover the audience as evenly and you’ll have more coverage overlap onto the walls and stage. The uneven coverage will diminish gain-before-feedback and the extra reverb and reflections will reduce intelligibility.
Here is a show that I worked on recently at a converted train station in Minneapolis with lots of glass and brick on the walls. Often, these rooms are designed with the speakers pushed out to the sides.
There are a couple of things I want to draw your attention to in this prediction at 5kHz. First, notice the dark blue color at the back of the audience on the right? That same color is covering my tiny stage. That means that I will have the same level at 5kHz at the back of the room as I will have on stage. Potential for microphone feedback? I think so.
Second, notice which colors are touching the audience that are also touching the walls. All of them! Remember the glass and brick? Get ready for lots of reflections.
Here is the same room, but with the speakers moved to the center of their coverage areas.
How did I decide where the center of their coverage areas were? First I compared the forward aspect ratio (FAR) of the audience with the FAR of the speaker and found that my speakers were too wide. The audience needs a FAR of 1.9, but my speaker has a FAR of 1.31. For more on forward aspect ratio, see my article One Simple Tool to Find the Right Size Speaker for Any Space.
Since my speakers are too wide, I decided not attempt any sub-division of the audience plane. Each speaker will cover the entire audience just fine by itself.
Second, I drew diagonal lines through the audience to find the center. This is an important step since the audience is not always symmetrical with the room or the stage.
Look at our stage again in this prediction. We can see some coverage at downstage center, but not nearly as bad as before. Next, notice that the light blue color at the back of the audience is also just touching the walls. This means that we will be hitting the walls with the same level as the back of the audience. It’s a 15dB improvement compared to the speakers pushed all the way out.
Now let’s look at what really happened on show day because we never get the positions that we want, do we? Speakers next to the stage would have blocked sight lines to the screens so I had to move them out, but I didn’t move them all the way out. I stayed as close in as possible.
This is a compromise so let’s look at the results. The light blue color at the back of the audience barely touches the stage. That’s good. Light blue is also just starting to cover the walls. We would prefer if it were only covering the audience, but it’s still better than the original example.
If you’d like to play with this design yourself you can download it here.
- If anyone in the audience cannot see a speaker, they won’t hear it.
- Get your speakers up higher for less level change from front to back of the audience.
- Place speakers near the center of their coverage areas.
Next Steps (from Chris)
Do you want to quickly and effectively maximize sound system tuning in any room? If so, check out:
You might have heard my podcast interview with him several months ago – here it is if you’re interested.
It’s over 58 video lessons with regular live office hour meetings. This means you can learn about sound system tuning AND tackle any problems specific to your room. Last year, he held this amazing workshop for a limited number of people and TODAY he’s opened it up again – but registration is only open until January 22, 2019.
Thought? Questions? Comments?