What does speech intelligibility mean?

Speech intelligibility means that the listener can effortlessly recognize the spoken words and sentences. It is not just about whether the voice is loud enough, but whether it reaches the recipient in a clear form.

The most important factors are:

• level of direct sound,
• level of reverberation,
• background noise,
• uniformity of coverage,
• band of the system,
• method of microphone placement.

If the direct sound is clearly dominant over subsequent reflections and noise, speech is easier to understand. If reflections start to compete with the direct signal, syllables and word endings begin to overlap.

Therefore speech intelligibility is a problem of the entire system and room, not just the quality of the microphone. 

Why does louder not always mean clearer?

 Increasing the level helps only up to the point where it improves the signal-to-noise ratio. If we also amplify the energy hitting the walls, ceiling, and other surfaces, the amount of reflected sound increases.

In a room with good acoustics, increasing the level can actually improve listening comfort. In a reverberant hall, the same operation can make the voice even more blurred.

That’s why the statement “it’s too quiet, turn it up” does not always describe the actual problem. The listener may perceive a lack of clarity as a lack of level, while the real issue is too much reverberation.

At a conference, the goal should not be the highest possible SPL, but a sufficient and even level with as good as possible separation from reflections and noise.

How does reverberation affect speech?

 Reverberation causes the energy of earlier syllables to remain in the room at the moment the speaker utters the next ones. The longer the decay time of the sound, the more the speech starts to overlap with itself.

This is most noticeable in large, hard rooms:

• halls,
• high foyers,
• rooms with a lot of glass,
• venues without acoustic treatment.

Long reverberation particularly hinders understanding of fast speech and statements from people speaking quietly or unclearly.

In such conditions, it is very important to increase the proportion of direct sound. This can be achieved, among other things, by using more directional sources and positioning them closer to the audience.

The article describes the problem in more detail How to amplify a hall with significant reverberation?

How to ensure an even level throughout the room?

An even level means that participants sitting close to the stage and those further away receive comparable, usable SPL.

If one system set up at the stage has to cover a very long room, the front rows may be too loud, while the back rows are still too quiet.

In that case, a better solution may be to divide the system into several zones:

• main system,
• frontfill,
• additional side sources,
• delay line.

Each additional source should cover a specific part of the audience and be appropriately time-aligned.

Evenness is important not only for comfort. If the further part of the room requires continuous increases in the main PA level, the level in the front rows also rises, along with the amount of energy hitting the room.

How to choose the directionality of the speakers?

Directionality determines where the system sends energy. In a conference, it should be chosen so that as much sound as possible reaches the audience, and as little as possible hits the ceiling and side walls.

Too wide a pattern can cause:

• greater stimulation of the room,
• greater number of reflections,
• deterioration of intelligibility.

Too narrow may leave part of the audience outside proper coverage.

Therefore, the choice of directionality should always be based on:

• width of the room,
• height of installation,
• depth of the audience,
• location of the stage,
• architecture of the room.

A properly selected system does not have to be the largest or loudest. It should primarily control where the energy goes.

Central system or distributed sound?

A central system works well where one or two main speaker locations can evenly cover the entire room.

Distributed sound uses a greater number of smaller sources placed closer to the audience. This solution can be beneficial in long or reverberant spaces.

The advantage of a distributed system is the ability to:

• reduce the distance between the source and the listener,
• lower the level of each individual speaker,
• limit the energy sent far into the room,
• improve uniformity.

The disadvantage is the greater number of sources that need to be properly delayed and balanced.

There is no rule that a central system is better or simpler in every case. In a challenging room, several properly synchronized sources can provide significantly better clarity than one very powerful system at the stage.

How to choose and set up microphones?

The microphone should pick up as much of the speaker's voice and as little sound from the room as possible. The most important factor is its distance from the mouth.

The closer the capsule is, the better the signal-to-noise ratio.

Host

The host can use a handheld, headset, or lavalier microphone.

A handheld microphone provides good control over the distance if the user holds it correctly.

A headset keeps the capsule close to the mouth and provides a very stable level, which is a significant advantage for a moving host.

A lavalier is discreet but is positioned further from the mouth, which means it picks up more ambient sound and usually provides a smaller margin before feedback.

Panel

A discussion panel is more demanding because multiple microphones can be open at the same time.

Each additional open channel increases the amount of collected reverberation and noise. Therefore, the following are important:

• proper microphone selection,
• channel management,
• levels,
• possible use of an automixer.

For panels, lavalier, handheld, gooseneck microphones, and discussion systems work well.

Questions from the audience

For Q&A, handheld microphones are most often used, passed around by the room staff.

This is a simple and effective solution because the microphone can be close to the mouth of the person asking the question.

An alternative is microphones on stands in the aisles or a discussion system. The worst option is trying to "catch" a question from a distance using a stage microphone.

Gain before feedback – why is it so important?

Gain before feedback means the maximum usable level of the system before feedback occurs.

Feedback occurs when sound from the speaker reaches the microphone, is amplified again, and returns to the speaker, creating an increasing loop.

The margins before feedback are influenced by:

• the distance of the microphone from the mouth,
• the position of the microphone relative to the PA,
• the directionality of the microphone,
• the directionality of the speakers,
• the number of open channels,
• the acoustics of the room.

Therefore, placing a more directional microphone close to the mouth can improve the situation more than trying to fix everything with EQ correction.

In a conference, gain before feedback is particularly important, because microphones are active for most of the event, and the host and panelists may change their positions.

How does the operator improve intelligibility?

The operator can improve clarity using filters, equalization, automixing, and proper level management. However, these tools work best when the system has been properly designed.

Filters

High-pass filters help reduce unnecessary energy in the low range, such as rumble, mechanical noise, or vibrations.

Speech does not always need the full range of the lowest frequencies, so limiting them can improve the clarity of the mix.

EQ

An equalizer is used to correct the signal's characteristics.

It can help reduce frequencies that are excessively exposed by the microphone or room, and improve the tonal clarity of the voice.

However, it should not serve as an attempt to fix poorly set speakers or very long reverberation.

Automixers

An automixer manages the levels of several microphones depending on who is currently speaking.

In panels and discussions, it can limit the number of channels operating at full gain, which reduces summed noise and reverberation.

An automixer does not replace the sound engineer, but can significantly facilitate the management of a large number of microphones.

Level management

The sound engineer responds to differences between speakers, the way of holding the microphone, and changes in the program.

One person may speak very quietly, while another may almost shout. Automation can help to some extent, but control of levels and dynamics is still needed.

What to do in a challenging acoustically conference room?

In a difficult room, it is essential to limit the amount of energy exciting the space.

The following may be helpful:

• more directional speakers,
• a shorter distance between the source and the listener,
• a distributed system,
• proper microphone placement,
• limiting the number of open channels,
• acoustic adaptation, if possible.

It is a mistake to try to solve the problem solely by increasing the PA volume.

If the event takes place in a demanding venue, it is worth analyzing the acoustics of the space and the choice of sound system beforehand., instead of designing the system only during installation.

For complex events technical support for the conference should include both the audio design and the way of conducting panels, microphones, multimedia, and possible streaming.

How to check the system before the participants enter?

The system should be checked in conditions as close as possible to the ones in which the conference will take place.

The test should include:

• all microphones,
• main speaker locations,
• the panel,
• the lectern,
• Q&A,
• the farthest seats in the room,
• side sections,
• possible delayed systems,
• routing for streaming and recording.

It is worth walking around the room and checking the system at several points, and not just evaluating it from FOH.

It is also necessary to conduct a test with microphones in real positions. A microphone test held by a technician 2 cm from the mouth does not reflect the situation in which during the event a lavalier will be several dozen centimeters from the source.

Summary

Good speech intelligibility at the conference requires primarily an appropriate relationship between direct sound, reverberation, and background noise. A high SPL alone does not guarantee clarity, and in a reverberant room it can even worsen the situation.

The most important factors are directionality and placement of the system, even coverage of the audience, proper microphone selection, and appropriate gain before feedback. In long or acoustically challenging rooms, a distributed system may be a better solution than one very loud PA.

The operator can additionally improve readability using filters, EQ, auto-mixing, and level control, but processing cannot replace a proper design. Therefore, the conference system should be checked in the entire room before the participants enter.

Frequently Asked Questions