MPEG-2 Audio Encoder

The Microsoft Media Foundation MPEG-2 audio encoder is a Media Foundation transform that encodes mono or stereo audio to MPEG-1 audio (ISO/IEC 11172-3) or MPEG-2 audio (ISO/IEC 13818-3).

The encoder supports Layer 1 and Layer 2 audio. It does not support MPEG-Layer 3 (MP3) audio. For MPEG-2, the encoder supports the Low Sampling Frequencies (LSF) portion of MPEG-2 audio. It does not support the multichannel extensions. The MFT outputs an MPEG elementary stream. It cannot generate packetized elementary streams, program streams, or transport streams.

Class Identifier

The class identifier (CLSID) of the MEPG-2 audio encoder is CLSID_CMPEG2AudioEncoderMFT, defined in the header file wmcodecdsp.h.

Output Types

The output type must be set first, before the input type. The following table lists the required and optional attributes for the output media type.

If the optional attributes are not set, the encoder adds them to the media type after the type is set.

Input Types

The following table lists the required and optional attributes for the input media type.

The encoder does not support sample-rate conversion or stereo/mono conversion. If the optional attributes are not set, the encoder adds them to the media type after the type is set.

Codec Properties

The encoder supports the following properties through the ICodecAPI interface.

To get a pointer to the ICodecAPI interface, call QueryInterface on the MFT.

The MFT implements the following ICodecAPI methods:

• GetParameterRange
• GetParameterValues
• GetValue
• IsModifiable
• IsSupported
• SetValue

All other ICodecAPI methods return E_NOTIMPL.

Remarks

Each MPEG audio frame contains either 384 (Layer 1) or 1152 (Layer 2) audio samples per channel. However, each input buffer to the encoder may contain any number of PCM samples. The size of each input buffer must be a multiple of the block alignment. The encoder caches input samples until it has enough for an MPEG audio frame.

Each output buffer contains one raw MPEG frame. The size of each output buffer depends on the bit rate and the sample rate.

Configuring the Encoder

To change any of the default settings on the encoder, perform the following steps:

1. Create an instance of the encoder MFT.

2. Call IMFTransform::GetOutputAvailableType to get the list of the preferred output types. The encoder enumerates all sample rates for both mono and stereo. Select one of these media types, based on the sample rate and number of channels. The MF_MT_AUDIO_AVG_BYTES_PER_SECOND attribute indicates the default bit rate, in bytes per second.

3. Optional: You can override the default bit rate by setting a new value for MF_MT_AUDIO_AVG_BYTES_PER_SECOND on the output media type. Valid bit rates depend on the sample rate, number of channels, and audio layer.

   Note

   At this point in the configuration process, the encoder defaults to Layer 2 audio and will only accept Layer 2 bit rates. You will be able to switch the encoder to Layer 1 in a later step (see step 7). In that case, leave the default bit rate for now; you can change it again in step 8.

    

4. Call IMFTransform::SetOutputType to set the output media type. If you set your own value for MF_MT_AUDIO_AVG_BYTES_PER_SECOND and the MFT rejects the output media type, it is likely because you specified an invalid bit rate.

5. Call IMFTransform::GetInputAvailableType to enumerate the input media type. Because the sample rate and number of channels must be identical to the output type, only two options are enumerated: 32-bit floating-point PCM input and 16-bit integer PCM input. Select one of these.

6. Call IMFTransform::SetInputType to set the input media type.

7. Optional: To encode Layer 1 audio, set the CODECAPI_AVEncMPALayer property to eAVEncMPALayer_1.

8. Optional: To change the bit rate, set the CODECAPI_ AVEncCommonMeanBitRate property. The bit rate must be one of the valid bit rates listed in the MPEG-1 or MPEG-2 LSF specifications. Alternatively, you can call ICodecAPI::GetParameterValues to get a list of valid bit rates, based on the current settings.

9. Optional: With 2-channel audio, you can set the CODECAPI_ AVEncMPACodingMode property to change the coding mode to dual channel or joint stereo. You can call ICodecAPI::GetParameterRange to get the valid options. (For 1-channel audio, the only option is mono.)

10. Optional: Set any of the other ICodecAPI properties listed previously.

It is important to follow the order of these steps. In particular, set the output media type before changing any ICodecAPI properties. Also, you must set ICodecAPI properties before the MFT receives the first input sample. After the MFT receives input, the codec properties are read-only, and ICodecAPI::SetValue returns the value S_FALSE.

Supported Bit Rates

The encoder supports the following bit rates.

MPEG-1

MPEG-2

Layer 1

Layer 2

Layer 1

Layer 2

32

32*

32

8

64

48*

38

16

96

56*

56

24

128

64

64

32

160

80*

80

40

192

96

96

48

224

112

112

56

256

128

128

64

288

160

144

80

320

192

160

96

352

224**

176

112

384

256**

192

128

416

230**

224

144

448

384**

256

160

\* Mono only \*\* Stereo only

Example Media Types

Here is an example of the media types that are needed to encode 16-bit integer PCM, 48-kHz stereo audio at the default bit rate.

Output media type:

Input media type:

Requirements

See also

Codec Objects