New Reference Listening Room for Two-Channel and Multi-Channel - - PDF document

(Design - Measurement - Modelling) Éva Arató-Borsi, Tamás Póth

Hungarian Radio, Technical Department

Andor T. Fürjes

Technical University of Budapest Hungarian Section of AES

104th AES Convention, May 1998

New Reference Listening Room for Two-Channel and Multi-Channel Stereophony

104TH AES CONVENTION, MAY 1998

Contents

2

• 1. Introduction
• 2. Design Considerations
• 3. The Measured Sound Field Parameters
• 4. Listening Test
• 5. Modelling
• 6. Conclusion

104TH AES CONVENTION, MAY 1998

• 1. Introduction

3

The need…

to test different techniques by listening tests all over the world needs to improve the listening conditions

The demand…

to achieve a special environment for listening conditions

The design

• f a new reference listening room in the Hungarian

Radio

The aim

to achieve the requirements given by recommendations.

2-channel 5-channel

104TH AES CONVENTION, MAY 1998

• 2. Design Considerations

4

Basis of the design considerations

EBU recommendation: the last version EBU Tech 3276 "Listening conditions for the assessment of sound programme material: monophonic and two-channel stereophonic"

The room geometry

the listening room was designed into an existing room Floor plan and the 3- dimensional view

104TH AES CONVENTION, MAY 1998

5

• 3. The Measured Sound Field Parameters

For practical reasons the room is used for talk studio too compromises

The room dimensions: All dimensions suit the recommendation:

proportions of the room floor area volume

Since these calculations assume perfectly rigid walls and rectangular shape… …a finite element model was created to investigate the low frequency behaviour

• f the room (SYSNOISE).

length 7.8 m width 7.05 m height 4.85 m area: 55 m2 volume: 267 m3

104TH AES CONVENTION, MAY 1998

6

• 3. The Measured Sound Field Parameters

The choice of the acoustical elements

• n the basis of the calculations placements of the elements

Wide-band absorbers Diffusers Low-frequency absorbers

104TH AES CONVENTION, MAY 1998

• 3. The Measured Sound Field Parameters

7

The properties of the sound field produced by the loudspeakers in the listening area determines the quality of the listening environment.

The sound field parameters specified by the EBU:

A Direct sound

B Early reflections C Reverberant field D Operational room response curve E Background noise

The measurements…

with the MLSSA analyser of DRA Lab.

104TH AES CONVENTION, MAY 1998

8

• 3. The Measured Sound Field Parameters

A Direct sound

Loudspeakers:

FL,FC, FR - Genelec 1038A SL,SR - Genelec 1032A

…suit the specifications B Early reflections For the two - channel stereophonic: Acceptable level of the early reflections: -10dB after the direct sound up to 15-20msec. ETF (Waterfall) Energy - Time - Frequency response at the reference listening point.

Front - left Front - center Front - right

104TH AES CONVENTION, MAY 1998

9

• 3. The Measured Sound Field Parameters

The results of the subjective tests show… …the influence of the early reflections around the surround loudspeakers can be disturbing. Acceptable level of the reflections?… …need for further investigations. The design objective: to avoid the strong early reflections for every channel. ETF response at the reference listening point - from surround speakers.

Surround - left Surround - right

104TH AES CONVENTION, MAY 1998

10

• 3. The Measured Sound Field Parameters

C Reverberant field The recommended nominal value 0.2 sec < Tm < 0.4 sec

THE MEASURED REVERBERATION TIME 0,1 0,2 0,3 0,4 0,5 0,6 10 12 5 16 20 25 31 5 40 50 63 80 1k 1,2 5k 1,6 k 2k 2,5 k 3,1 5k 4k 5k 6,3 k 8k f[Hz] T60[s] T 60 [s] EBU references

The measured reverberation time

100 125 160 200 250 315 400 500 630 800 1k 1.25k 1.6k 2k 2.5k3.15k 4k 5k 6.3k 8k

104TH AES CONVENTION, MAY 1998

11

• 3. The Measured Sound Field Parameters

D Operational room response curve Definition: the sound pressure level produced by the loudspeakers at the reference point.

• 10
• 5

5 10 15 20 25 40 63 100 160 250 400 630 1k 1,6k 2,5k 4k 6,3k 10k 16k

f [Hz] A [dB]

FL FR FC

• 10
• 5

5 10 15 20 20 25 31, 5 40 50 63 80 100 125 160 200 250 315 400 500 630 800 1k 1,2 5k 1,6 k 2k 2,5 k 3,1 5k 4k 5k 6,3 k 8k 10k 12, 5k 16k 20k

f [Hz] A [dB]

SL SR

Measured operational room response curves

Surround -

Left (SL) Right (SR)

Front -

Left (FL) Center (FC) Right (FR) 20 31.5 50 80 125 200 315 500 800 1.25k 2k 3.15k 5k 8k 12.5k 20k

104TH AES CONVENTION, MAY 1998

• 4. Listening Tests (5-channel)

12

4 m 3.4 m 3.4 m 3.4 m 3.3 m 3.3 m

Front Surround

Left Center Right Left Right 4.6 m

Position of speakers and listeners

Test procedures and equipment

Yamaha DSP-E492 commercial Dolby Pro Logic

decoder with Dolby Surround encoded music/effects

Tascam DA88 8-channel digital recorder with discrete 5-

channel recordings

104TH AES CONVENTION, MAY 1998

• 4. Listening Tests (5-channel)

13

4 m 3.4 m 3.4 m 3.4 m 3.3 m 3.3 m

Front Surround

Left Center Right Left Right 4.6 m

Position of speakers and listeners

Subjective results

Dolby Surround: better if surround speakers are turned to face each other Discrete 5-channel: better if surround speakers are turned to listeners Surround speakers shall sound as diffuse as possible Spatial impression is less position dependent

104TH AES CONVENTION, MAY 1998

• 4. Listening Tests (5-channel)

14

Subjective and objective parameters

Impulse Response Processing Objective parameters

modified M-factor: Centre time:

( ) ( )

M k t k t = − 10 10

10 2 1 10 2 2

log log

( )

t t p t dt p

s =

⋅

∞

∫

2 2 max

???

( ) ( ) ( )

k t p t dt p t dt

t t 2 2 2

=

∫ ∫

∞

104TH AES CONVENTION, MAY 1998

• 5. Modelling

15

Why computer aided modelling?

prediction in design phase, shorter design period, low costs; modelling posteriorly helps improvement of modelling procedure; experiments with “virtual” set-ups.

Modelling with computers - review

numerical solutions (FDM, FEM, BEM) geometrical room acoustics (Image-source models, Ray- tracing, Cone-tracing, Beam-tracing)

104TH AES CONVENTION, MAY 1998

• 5. Modelling

16

Modelling technical rooms - problems

relatively small sizes (architecture and furniture together, near-field modelling of sources and reflections, limits of geometrical acoustics) special surfaces and materials (diffuser, absorber, etc.) special requirements (what parameters, measuring surface material and sound source properties - no high resolution data available)

Low Frequency Modelling (SYSNOISE)

acoustic modes positioning of low-frequency sources effect of the “rigid” wall

104TH AES CONVENTION, MAY 1998

• 5. Modelling

17

S

S S’ ϕ’ ϕ ϕ ∆t⋅c

High Frequency Model

why developing a new software…

Basic Assumptions

beam-tracing beam surfaces are exact pieces

• f the wave front

boundaries may be curved - beam distortion diffuse reflection means the change of solid angle

• f the beam

“sampling” in detection of reflections

104TH AES CONVENTION, MAY 1998

• 5. Modelling

18

Results of Modelling

predicted measured frequency response energy-decay

Improvements

correct diffuse reflection model correct diffraction model - sampling a very coarse approach introduction of phase near-field model of sources and reflections precise data about sources and materials

104TH AES CONVENTION, MAY 1998

• 6. Conclusion

19

Requirements Measurements Subjective Test Modelling Next…

104TH AES CONVENTION, MAY 1998

17

• 5. Modelling

S

S S’ ϕ’ ϕ ϕ

High Frequency Model

why developing a new software…

Basic Assumptions

beam-tracing beam surfaces are exact pieces

• f the wave front

boundaries may be curved - beam distortion diffuse reflection means the change of solid angle

• f the beam

104TH AES CONVENTION, MAY 1998

18

• 5. Modelling

Results of Modelling

predicted measured frequency response energy-decay

Improvements

correct diffuse reflection model correct diffraction model introduction of phase near-field model of sources and reflections precise data about sources and materials