It's About Time… The Effective Haas Effect

Theory and Practice – By Dale Shirk

To implement the Haas Effect, understanding the theory is a must.  Dale Shirk explains the theory and ways to implement it.

The Haas effect, or Precedence effect has been well known for many years. It is a psycho-acoustic effect related to the ear’s localization of sound. Simply stated, when two identical sounds from different sources arrive at the ear with only a small time difference, the ear will localize the sound as coming from the earlier source only. This is a quite natural everyday occurrence. The direct path from a sound source to the ear is the shortest and always arrives first. Later a reflected copy of the same sound may arrive at the ear after having traveled farther and reflecting off the floor, wall or ceiling. Our ear correctly identifies the first arrival as the direction of the sound source. What is somewhat surprising however is that the later arrival can actually be higher in level than the original arrival, and our ears will still insist that the sound is coming from the first source.

There are of course limits to both the time and level over which this effect works. Take a moment to look at the classic Haas Effect chart. The effect doesn’t really come to full force until the time difference reaches 5 milliseconds. Shorter times than that tend to produce the sensation of a single source, often of nebulous location, but with obvious comb filter frequency characteristics. From 5 to about 25 milliseconds of time difference the Haas effect is quite strong, allowing the later arrival to be higher level than the first arrival. Longer time differences tend to be heard as a thickening, or doubling sound, until at about 50 milliseconds you hear it as two separate events.

The classic Haas chart also shows the later arrival being up to 10 dB greater than the first arrival. It should be carefully noted that this is for equal perception of localization. In many cases we are interested in keeping the later arrival “hidden”, so it should stay less than 10 dB above the first arrival.

The most common deliberate application of this effect is when using fill speakers, say to cover an area under a balcony or over a balcony, or even just using multiple delayed speakers in a long room. To determine the proper delay for these speakers we need to examine how the time offset varies with floor position. The place where the second speaker will arrive earliest, relative to the first speaker, is where you are looking through the second speaker to see the first. There may not be a seat at that location, but the closest you can get to it is where you start for setting your delay. Since the early limit for Haas effect is about 5 milliseconds, that’s what you start with at that seat.

Every seat in front of or to the side of there, will have the second speaker arriving later than the first by an increasing amount. As you move forward the offsets will eventually go way beyond 25 ms but not be a problem because the level from the second speaker arriving late, will also be lower than the level than the first speaker due to it’s directivity. In essence it becomes just another room reflection. For the sake of intelligibility, the later arrival should be lower in level than the first arrival when the time difference is 30 milliseconds or more. If the time offset becomes too great before the level differences become great enough, then some compromise will need to be made at the rear, minimum offset location. This is best looked at carefully at the design stage using modeling, noting the time differences and level differences in each area.

It’s important that the first speaker provide a strong enough early arrival, within 10 dB, to all areas where you want the Haas effect to localize to it. It’s also important that the sound quality and frequency balance delivered by the front speaker be even and consistent to all seats, even though the level is falling off. Our ears judge sound quality by the first arrival. You can’t fix it by compensating with the second arrival.

At times attempts have been made to use delay and the Haas effect to make the whole PA disappear by delaying it behind the natural sound of the talker. Usually this meets with little to no success, due to the fact that the natural sound ends up being more than 10 dB below the sound from the PA. However in small rooms with strong enough acoustic sources – in other words, the PA is barely needed – It can be done quite well. For it to be successful the acoustic gain of the PA is under 10 dB.

Attempts have also been made to delay the PA to match the backline, instrument amps and drums. Again the 10 dB acoustic gain limit is usually violated. For much of the audience the first arrival will be the spill off the back of the floor wedges. As I mentioned above, we subjectively judge the sound quality based primarily on the first arrival. That should not be the mud off the backs of the wedges. Keep the main PA arriving first for as much of the audience as possible. Front row seats may benefit from lip fills arriving first to pull localization down from the overhead speakers.

The Haas effect can also hide reflections which are detrimental to intelligibility. One church I measured had a reflection from the balcony face to a group of seats in the front corner. At this location the reflection was 7 dB stronger than the direct sound and was 35 milliseconds late. The reflection was stronger because the balcony face was on axis of the speaker, while the seats received less direct sound because they were off axis. This reflection was only audible as such when you cupped your hands around your ears and turned towards the balcony face. Yet it completely destroyed intelligibility there. The solution was to treat the balcony face.

Get beyond the simplistic prescription, “Just add 20 milliseconds.” By carefully considering all the implications of the Haas effect, including both time and level, you can accurately predict when it will and won’t work, and use it to it’s fullest benefit. ds