Height Channels in Surround Systems

In multi-channel audio content. Room acoustics might conflict or maybe assist the target space in some cases.

There are inherent limitations of some of the existing surround layouts; such as elevation and directivity, which are crucial elements in sound reproduction in a multi-channel format.

Screen Shot 2019-06-11 at 9.52.16 PM.png
surround_model2_large.jpg


I’ve conducted a small scale experiment to investigate the importance of surround-height channels and its effect on the listening experience for an audience.

I started by recruiting 10 untrained listeners and I asked them to compare seven 7 clips played back in 3 different formats. (Atmos-DTS:X-NHK 22.2)

Results have indicated that surround-height loudspeakers have a greater influence on perceived sound quality and localization of sound sources for untrained listeners.


The theory behind heigh perception and elevation

It’s not only important in spatial awareness, which enables the listener to pinpoint sound source of sound components and improves accurate localization. One of the defining elements of an audio system is how well it can reproduce spatial impressions. 

Conventional surround sound formats add the height dimension to the existing width and depth dimensions.  Naturally you’d expect the added height channels to enhance perceived spatial impression; However, there is the question of how effective each format is for the target listener to perceive elevation, height and how fundamentally important are the height channels.

Screen Shot 2019-06-11 at 10.16.51 PM.png

There are two ways to map signals to height channels and represent sounds in the elevation plane; either by routing discrete sources to height channels or by creating a vertical phantom image between main and height loudspeakers; however, there is always the question of how effective each method is.

images courtesy of microsoft

images courtesy of microsoft

When it comes to spatial awareness in a surround system auditory height perception is paramount as it enables listeners to locate elevation of objects and highly improves localization. 

We localize sounds in the height plane differently from the median plane; we are reliant on a combination of the time and level differences between a given source arriving at each of our ears as well as on the directional filtering of the sound source by the pinnae, in the median plane however; binaural cues are not present considering sounds arrive at each ear simultaneously; therefore median plane localization is solely based on spectral cues.

There are a number solid research confirming how localization in the median and height plane is governed and several papers that demonstrate that the perceptual range of pitch-height depends on the physical height of the loudspeaker that presents the signal; However height speakers in surround sound and how to vital it is to the perception of sound in the vertical plane remains under-researched.

A conventional surround sound system in a cinema reproduces a sound field that is particularly related to the visual image and gives priority to front sound reproduction; key issues in such an approach however-are presence and reality - presence encompasses sound image localization- envelopment -depth and spatial impression.


The experiment

For that we recruited 10 untrained listeners, stimuli were short clips with height/elevation information- each one lasted no more than 10 seconds.

The clips were mixed for 3 different surround sound formats; each clip was mixed and adjusted to provide the same loudness level and same reverb for three different reproduction systems “NHK 22.2”, “DTS: X”, “Atmos”.
Participants were given a questionnaire that lists seven attributes of each stimulus and they were asked to rate each attribute on a scale from one to five.

After familiarizing themselves with the testing procedure and stimuli, participants rated the seven clips in three surround formats (NHK 22.2; DTS: X, Dolby Atmos) for seven salient attributes: perceived depth, spatial clarity, impression being enveloped, believability, harshness; fullness and spectral balance of the sound source and its distance. These attributes were chosen from previous literature on spatial audio and its perceptual evaluation.

Each participant’s ratings were standardized (centered and scaled using the mean and standard deviation) along five ratings for each attribute. Below are the figures illustrating descriptive statistics of scores from seven attributes.  Four stimuli were denoted with characters and three labels (NHK 22.2; DTS: X and Dolby Atmos) indicating reproduction formats.

Each stimulus is indicated with one of four characters (S.1, S.2, S.3, and S.4) and one of three labels (NHK 22.2, DTS: X and Dolby Atmos). For instance, any stimulus ending with a label Atmos is a Dolby Atmos format that reproduced the effect.

It appears that the rated attributes differentiate two multichannel reproduction formats (NHK22.2-and DTS: X). Depth explains the unique characteristics of the DTS: X format reproduction method. In other words, when a sound effect was reproduced through in the DTS: X channel system, listeners perceived the depth benefits of a reproduced sound in that speaker configuration.  The unique features of NHK 22.2 and DTS: X formats changed listeners’ impression of spectral/spatial integration and localization of sound components.

For those interested in the literature behind these articles; I’ve listed some cool references below.

1. The First Step Toward 3D Audio: DTS Neo: X, URL:

2. Dolby, Authentic Cinema Sound by Dolby Atmos, URL:

3. S. Ferguson and D. Cabrera. “Vertical Localization of Sound from Multiway Loudspeakers,” J. Audio Eng. Soc., 53(3), 163–173. 2005

4. D. Cabrera and S. Tilley. “Vertical Localization and Image Size Effects in Loudspeaker Reproduction,” Audio Engineering Society 24th International Conference: Multichannel Audio, the New Reality. 2003.

5. C. C. Pratt. “The Spatial Character of High and Low Tones,” J. Exp. Psy, vol. 13(3), pp. 278–285. 1930.

6. C. C. Pratt. “The Spatial Character of High and Low Tones,” J. Exp. Psy., vol. 13(3), pp. 278–285. 1930

7. M. Morimoto, M. Yairi, K. Iida, and M. Itoh, “The Role of Low-Frequency Components in Median Plane Localization,” Acoust. Sci. Technol., vol. 24, pp. 76–82 (2003).

8. J. Blauert, Spatial Hearing, rev. ed. (MIT Press, Cambridge, MA, 1997).

9. S. K. Roffler and R. A. Buttler, “Factors that influence the localization of sound in the vertical plane,” J. Acoust. Soc. Am., vol. 43 (6), pp. 1255- 1259 (1968).

10. Hamasaki, K. H. (2004). Advanced multichannel audio systems with a superior impression of presence and reality. 116th AES Convention.

11. Kimio Hamasaki 1, 2, Koichiro Hiyama 1, and Reiko Okumura 1 (2005). The 22.2 Multichannel Sound System and Its Applications. 118th AES Convention.

12. H. Lee, C.Gribben, R.Wallis. (2014) Psychoacoustic Considerations in Surround Sound with Height RTonmeistertagung - VDT International Convention.


Ahmed GelbyComment