MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high...

23
Olaf Korte Fraunhofer IIS Broadcast Applications Audio & Multimedia Realtime Systems www.iis.fraunhofer.de [email protected] phone: +49-(0) 9131 / 776-6330 fax: +49-(0) 9131 / 776-398 Fraunhofer Institut Integrierte Schaltungen MPEG Spatial Audio Coding Multichannel Audio for Broadcasting

Transcript of MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high...

Page 1: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

Olaf Korte

Fraunhofer IISBroadcast ApplicationsAudio & Multimedia Realtime Systems

[email protected]: +49-(0) 9131 / 776-6330fax: +49-(0) 9131 / 776-398

Fraunhofer Institut Integrierte Schaltungen

MPEG Spatial Audio CodingMultichannel Audio for Broadcasting

Page 2: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

2Olaf Korte © by Fraunhofer IIS, 9st May 2005

Overview

– Motivation and Broadcast Requirements

– Spatial Audio Coding

– MPEG Standardization Process

– Current Status (MPEG RM0)

– HD Radio and todays Demo setup

– Conclusion and Outlook

Page 3: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

3Olaf Korte © by Fraunhofer IIS, 9st May 2005

Motivation

– Killer Application for Digital Radio

– Competitive Media support Surround Sound

– Broadcasters start with Surround Sound

(DVB-S, XM, Sirius)

– Car Industry supports Surround Sound

Page 4: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

4Olaf Korte © by Fraunhofer IIS, 9st May 2005

Broadcast Requirements– Backwards Compatibility

– No simulcast

– No quality loss for conventional receivers

– High Quality

– Stereo (artistic downmix)

– Multichannel (as discrete as possible)

– For all kind of material

– Low Extra Costs for Transmission

Page 5: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

5Olaf Korte © by Fraunhofer IIS, 9st May 2005

Spatial Audio Coding

Spatialparameters

SpatialMulti-channelreconstruction

ch1ch2

chN

SAC Decoder

Stereoor

monoch1ch2

chN

Stereoormono

Multi-Channel

N channels

Automaticdownmix(optional)

SpatialParametersEstimation

artisticdownmix(optional)

SAC Encoder

Page 6: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

6Olaf Korte © by Fraunhofer IIS, 9st May 2005

Spatial Audio Coding– Works on mono or stereo core and leave this untouched

– Spatial parameters contain most salient perceptual aspects of multi channel sound image, e.g.

– Interchannel level differences

– Interchannel time/Phase differences

– Interchannel correlation/coherence

– Scalable with regard to freq. bands, time resolution, …

– Scalable bitrate (quality) for Spatial parameters

– Most Digital Audio systems have capabilities to transport sideinformation transparently(MPEG layer 2, mp3, HE AAC: Ancilliary data)

Page 7: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

7Olaf Korte © by Fraunhofer IIS, 9st May 2005

Spatial Audio Coding - Bitrate/Quality Scalability

Multi ChannelQuality

Spatial Parameter Bitrate

Transparency

Matrixed Surround

Spatial Audio Coding

Page 8: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

8Olaf Korte © by Fraunhofer IIS, 9st May 2005

Spatial Audio Coding– “HD Radio Spatial Audio“ (96/16kbps)

Complete Realtime Broadcast Chain(FhG/Omnia/Telos, 2004/2005)

– “HD Radio Parametric Surround“HD Radio Codec Demo (CT, 2004)

– Eureka 147 “DAB 5.1“ (192/16kbps)Broadcasts for whole bavaria (BR/FhG, 2004)

– Eureka 147 DAB Parametric Surround (256/32kbps)(IRT/CT, 2004)

– “mp3surround“ (16kbps) (2004)

– HE AAC Spatial Surround (48/16kbps) (FhG, 2004)HE AAC

Page 9: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

9Olaf Korte © by Fraunhofer IIS, 9st May 2005

MPEG Standardisation

– Spring 2004: 4 concurrent systems are proposed to MPEG

– Oct 2004: MPEG Listening Test Results showed that the two

best systems performed equally

(“CT/Philips“ and “FhG/Agere-Systems“)

– Oct 2004: MPEG decided to develop a merged system

– RM0: First version of merged system

– RM0 test results available since April 2005

– Committee Draft in July 2005

Page 10: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

10Olaf Korte © by Fraunhofer IIS, 9st May 2005

Current Status – Release Milestone 0 („RM0“)

--31.65RM0 high_quality

--5.85RM0 low_rate

4.6811.7811.68RM0

Test Condition t3

(515, 48kbps)

Test Condition t2a

(515, 80kbps)

Test Condition t1a

(525, 160kbps)

Bitrate (kbps)

MPEG RM0 test cases (MUSHRA)

7 independent test sites with a total of 279 listening subjects)

Page 11: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

11Olaf Korte © by Fraunhofer IIS, 9st May 2005

Current Status – Release Milestone 0 („RM0“)

0

20

40

60

80

100

t1a(525)

t2a(515)

t3(48kbps)

t1LrHq(525)

CTP_CfP

FhG-A _CfP

R M0

D PL2

22 17 12 23 16 12 13 9 5 6 12 32

MO

S

Test case

Spatial bitra te(kbps)

Page 12: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

12Olaf Korte © by Fraunhofer IIS, 9st May 2005

HD Radio with MPEG Spatial Audio Coding

HDAudio

Decoder

HDBitstreamDe-Mux

L

R

HD Stereo Receiver

HDAudio

Encoder HDBitstream

Multiplexer

L

R

L

R

HD Station

Page 13: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

13Olaf Korte © by Fraunhofer IIS, 9st May 2005

HD Radio with MPEG Spatial Audio Coding

Enhanced HD Surround Receiver

HDAudio

Decoder

HDBitstreamDe-Mux

L

R

HD Stereo Receiver

HDAudio

Decoder

MPEGSpatialAudio

Decoder

HDBitstreamDe-Mux

L

R

LFRFC

LFELS

RSMPEG Spatial

Bitstream

HDAudio

Encoder

MPEGSpatial

Encoder

HDBitstream

Multiplexer

L

R

MPEG SpatialBitstream

L

R

HD Station with artistic downmix

LFRFC

LFELS

RS

Page 14: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

14Olaf Korte © by Fraunhofer IIS, 9st May 2005

HD Radio with MPEG Spatial Audio Coding

Enhanced HD Surround Receiver

HDAudio

Decoder

HDBitstreamDe-Mux

L

R

HD Stereo Receiver

HDAudio

Decoder

MPEGSpatialAudio

Decoder

HDBitstreamDe-Mux

L

R

MPEG SpatialBitstream

HDAudio

Encoder

Downmixand

MPEGSpatial

Encoder

HDBitstream

MultiplexerLFRFC

LFELS

RS

L

R

MPEG SpatialBitstream

HD Station with automatic downmix

LFRFC

LFELS

RS

Page 15: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

15Olaf Korte © by Fraunhofer IIS, 9st May 2005

MPEG Spatial Audio RM0 demo for HD Radio

MPEG Spatial Audio RM0Current Implementation Status

– Encoder

– Mathlab Code available for generation of test items

– Realtime Implementation under development

– Decoder

– Mathlab Code available

– Realtime Implementation for PC code available

Page 16: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

16Olaf Korte © by Fraunhofer IIS, 9st May 2005

MPEG Spatial Audio RM0 demo for HD Radio

Non-Realtime Encoding (prepared for the demo)

Downmixand

MPEGSpatial

Encoder*.wav

discrete5.1 Files

*.bsSpatial Data

Bitstream

Stereodownmixwave files

5.1

MPEG Spatial Audio Parameters Bitstream5,85kbps

L

R

Page 17: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

17Olaf Korte © by Fraunhofer IIS, 9st May 2005

MPEG Spatial Audio RM0 demo for HD Radio

Non-Realtime Encoding (prepared for the demo)

Stereodownmixwave files

HDAudio

Encoder

HDAudio

Decoder*_hdc.wav

HD stereo qualitywave files

L

R

HDC AudioBitstream

90kbpsStereo

Encoding and Decoding done byCoding Technologies

Page 18: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

18Olaf Korte © by Fraunhofer IIS, 9st May 2005

MPEG Spatial Audio RM0 demo for HD Radio

Non-Realtime Encoding (prepared for the demo)

Downmixand

MPEGSpatial

Encoder*.wavediscrete5.1 Files

*.bsSpatial Data

Bitstream

Stereodownmixwave files

HDAudio

Encoder

HDAudio

Decoder*._hdc.wav

HD stereo qualitywave files

5.1

MPEG Spatial Audio Parameters Bitstream5,85kbps

L

R

L

R

HDC AudioBitstream

90kbpsStereo

Encoding and Decoding done byCoding Technologies

Page 19: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

19Olaf Korte © by Fraunhofer IIS, 9st May 2005

MPEGSpatial

Decoder

MPEG Spatial Audio RM0 demo for HD Radio

*.bsSpatial Data

Bitstream

*_hdc.wavHD stereo quality

wave files

MPEG Spatial AudioParameters Bitstream

5,85kbps

L

R

Realtime Decoding at the demo

External6chUSB

Soundcard

5.1Audio

LF

RF

C

LFE

LS

RS

Page 20: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

20Olaf Korte © by Fraunhofer IIS, 9st May 2005

MPEGSpatial

Decoder

MPEG Spatial Audio RM0 demo for HD Radio

*.bsSpatial Data

Bitstream

*_hdc.wavHD stereo quality

wave files

MPEG Spatial AudioParameters Bitstream

5,85kbps

L

R

Realtime Decoding at the demo

External6chUSB

Soundcard

5.1Audio

*_dec.wavDecoded 5.1

wave files

LF

RF

C

LFE

LS

RS

Page 21: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

21Olaf Korte © by Fraunhofer IIS, 9st May 2005

WavSwitch

MPEG Spatial Audio RM0 demo for HD Radio

Realtime ComparisonOriginal / RM0 5.1 / HD stereo

External6chUSB

Soundcard

LF

RF

C

LFE

LS

RS

*_dec.wavRM0 5.1

wave files

*_hdc.wavHD stereo quality

wave files

*.wavdiscrete 5.1wave files

Original 5.1

Spatial 5.1

Downmix

Page 22: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

22Olaf Korte © by Fraunhofer IIS, 9st May 2005

Conclusion

– MPEG Spatial Audio offers perfect multichannel for HD Radio

– Highly Scalable (matrix, high quality, transparent)

– Backwards compatible (perfect downmix)

– Future looking Technology (e.g. 7.1, 10.2, …)

– Open MPEG standard

– Is a must for HD Radio!

(for sure will also be used in many other applications)

Page 23: MPEG Spatial Audio Coding Multichannel Audio for … archive...– Highly Scalable (matrix, high quality, transparent) – Backwards compatible (perfect downmix) ... 050509_NRSC-SSATG-SAC_May9th.ppt

MPEG Spatial Audio –Multichannel Audio for HD Radio

Fraunhofer Institut Integrierte Schaltungen

23Olaf Korte © by Fraunhofer IIS, 9st May 2005

Outlook

– MPEG RM0 offers high quality and flexiblity

though it is only the basic profile of the new standard!

– A Digital Solution for a Digital System:

Choose the best available system for HD Radio!

– We strongly recommend an independent MUSHRA testing by a

neutral party for direct comparison of proposed systems

– Cost of this testing shall be be divided among the proponents