Digital Multi-Channel Audio Compression and Metadata Dolby E and Dolby Digital (AC3) surround sound, concept of Dolby metadata #1 assuredcommunications Feb 19, 20
Digital Audio Compression Multi-Channel Programming Enhances the home listening experience Desired by consumers Extra features Additional languages, SAP Emergency audio services Descriptive comments for viewing impaired and so on Backward compatibility to Mono, Stereo #2 assuredcommunications Feb 19, 20
Implementation L R C Ls Rs LFE Lt Rt VTR / Server Codec / Encoder Typical 8 Channels (4 AES pr.) L R C Ls Rs LFE Lt Rt Metadata Using un-compressed signals, all channels of VTR / Server used. Requires 8 levels of audio router. External connection of serial Metadata stream (routing?) #3 assuredcommunications Feb 19, 20
Implementation 5.1 (+2) Lt/Rt VO EFX SAP-1 SAP-2 Other? Other? VTR / Server Codec / Encoder Typical 8 Channels (4 AES pr.) 5.1 (+2) Lt/Rt VO EFX SAP-1 SAP-2 Other? Other? Metadata Using compressed 5.1 signals, allows more programs and monitoring of Lt/Rt downmix without decoder. #4 assuredcommunications Feb 19, 20
Digital Audio Compression Convenience Digitally encoded and compressed signals less susceptible to impairments due to signal path Impulse Noise Ground Loops Random Noise Amplitude/Phase Distortion Single stream can be transported in video ancillary Reduce physical routing layers ($$) A/V timing advantages #5 assuredcommunications Feb 19, 20
Digital Audio Compression Perceptual Coding (AC-3) Reduce the data rate of a digital audio signal without introducing any perceivable audible changes. Several physiological limitations of the human hearing system. Predicts which sounds your ears will and will not hear, and only encodes audible sounds. Uses the human ear s - hearing threshold phenomenon. Ear is not equally sensitive at all frequencies Detect quiet signals in the 2 khz - 4 khz midrange Less sensitive to quiet signals at very low or very high frequencies #6 assuredcommunications Feb 19, 20
Digital Audio Compression Perceptual Coding (AC-3) Examples: Louder Sounds Mask Quiet Sounds (Relative) Orchestra, flutes quiet during loud passage - only hear flutes if you are in the orchestra Out of doors - listen to a bird sing - truck passes Some sounds under threshold of hearing #7 assuredcommunications Feb 19, 20
Digital Audio Compression Perceptual Coding (AC-3) Low Frequencies db Speech and Primary Frequencies Harmonics and Incidental High Frequency Content Eliminated with Perceptual Coding 0 Hz 100 Hz 500 Hz 1 Khz 10K Hz 20K Hz Frequency Threshold of Hearing #8 assuredcommunications Feb 19, 20
Metadata Metadata used to shape consumer listening experience to particular requirements Transmission Bitstream - Intended for the transmission of audio to the home through digital television broadcast (either high or standard definition), Set Top Box, DVD, or other media. Defines a single channel of audio through a full 5.1-channel program, including Metadata in both D-TV and DVD. AC-3 designed for maximum fidelity and space efficiency, and only passes through one encode/decode cycle. #9 assuredcommunications Feb 19, 20
Metadata Additional control information carried with the encoded audio program and provides essential information about the audio to an AC-3 decoder. Data describing the audio data format Created at time of program origination Provides many important functions including the three D s : Dynamic range control Dialnorm Downmix Complete Control of final audio listening environment! #10 assuredcommunications Feb 19, 20
Metadata Flow Receiver Metadata Modifies Listening Levels Audio Mixer Audio Multi-Channel Encoder Audio Metadata VTR / Server Audio Metadata MPEG Encoder Mix set in Production Video #11 assuredcommunications Feb 19, 20
Metadata Dialog Level (Dialog Normalization or Dialnorm) Dialnorm (loudness uniformity) Describes the average program volume Level variations are undesirable Between different programs Between program segments (station breaks and commercials) Metadata contains Dialnorm value, used by decoder #12 assuredcommunications Feb 19, 20
Metadata Dialog Level (Dialog Normalization or Dialnorm) Set by the program producer or the broadcaster Defined as the level of normal spoken dialogue with respect to Full Scale Digital Dialnorm values range between -31(no level shift in the home decoder) to -1 (maximum level shift in the home decoder) Dialnorm also applies to other types of program material, like music videos and concerts #13 assuredcommunications Feb 19, 20
Metadata Dialog Level (Dialog Normalization or Dialnorm) Turn It Down! Output Level Program 1 Dialog Level Program 2 Dialog Level Comfortable Listening Level Input Level #14 assuredcommunications Feb 19, 20
Metadata Dialog Level, Dialog Normalization, or Dialnorm Why -27dB? Film Soundtrack Dialog Level 78 db AC-3 has 105 db of dynamic range, Loudest level is "0 db, Quietest level is "-105 db". -27 db aligns with movie soundtracks in that 78 db (above silence) 105-27=78 is an accepted level for speech. Dialnorm = value means the level that dialogue is lower than the peak (0 db), Value of "-31" is 31 db below the peak (the value at which no volume adjustment is performed by a consumer decoder). A value of -27 causes the decoder to reduce the program by 4 db AC-3 Output Level #15 assuredcommunications Feb 19, 20
Metadata Downmix Three types of downmix: Surround downmix Lt / Rt Left total / Right total, for Pro Logic compatibility No Monaural Compatibility Stereo downmix Lo / Ro Left only / Right only, possibly for headphones Correct audio phase, is monaural compatible Mono downmix (DVB-H) From Lo / Ro Metadata controls C and S mix level #16 assuredcommunications Feb 19, 20
Metadata DRC, Dynamic Range Control Some listeners want full dynamic range Some listeners do not! listening conditions vary ambient noise problems late night listening the kids are asleep the neighbors are complaining #17 assuredcommunications Feb 19, 20
Metadata DRC, Dynamic Range Control The level of audio that falls above the dialog area (as defined by the dialnorm value) is cut. -31 consumer dialnorm Output Level Dialog Level Audio that falls within the dialog area (a.k.a. the null band ) is unaffected. The level of audio that falls below the dialog area (as defined by the dialnorm value) is boosted. Input Level #18 assuredcommunications Feb 19, 20
Metadata Compression (AC-3 decoder) RF Link Heavy Protect peak levels, transmission paths with small dynamic range Line Light Use in noisy environments, small reduction of dynamics None Automatic or User Controlled (Set top box design) #19 assuredcommunications Feb 19, 20
AC3 Metadata Variables #20 assuredcommunications Feb 19, 20
Summary Metadata - Specific Data about the audio data Controls listening experience Contains Information about the content Program Name Stream Type Time Reference Controls downmixing Center Surround Provides some user control #21 assuredcommunications Feb 19, 20
Questions? TVM Series VTM Series #22 assuredcommunications Feb 19, 20