Audio formats that will be supplied to Recognizer.
|
Value |
String of audio parameters separated by semi-colons. |
|
Default |
audio/basic;rate=8000 |
|
How to set |
In Management Station set on the Nuance recognition service. (To set more than one value, use the pipe | symbol as a separator.) If not using Management Station, set in a Recognizer configuration file (User-nrsxx.xml). |
|
Usage |
Frequently used by system administrators during installation. The voice browser determines which formats are used. |
Recognizer uses this parameter to load matching recognition models during initialization.
At runtime, the browser specifies the audio format when it writes audio samples to the Recognizer (the browser can only specify formats declared by this parameter). By declaring the needed audio formats in advance, this parameter avoids the initialization time and memory cost of loading unneeded formats.
The parameter value has the following format:
media-type-prefix[;rate=sample-rate][;encoding=encoding][;orig-encoding=voip-std][;bitrate=voip-bitrate]
|
Format element |
Description |
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
media-type-prefix |
Indicates the format of the audio data:
Note: Aurora data is a description of audio features defined by the European Telecommunication Standards Institute (ETSI). Aurora is supported by Recognizer but due to size constraints the models are not shipped with the language packs. The Aurora models are available upon request. Contact your sales representative. |
|||||||||||
|
rate |
Optional. Indicates samples/second. The default is 8000. Other rates are allowed based on the media-type-prefix. |
|||||||||||
|
encoding |
Required if the media type is application/x-feature, in which case the encoding must be ES_202_050 (for the ES 202 050 standard). |
|||||||||||
|
orig |
Optional. This element is useful for installations with customized language packs for applications that receive voice over an IP network (VoIP). Use this argument to indicate the original encoding of the audio received by the application. Values: g723 or g729. It indicates that u-law audio has been decoded from VoIP audio. You cannot input VoIP encoded audio directly; you must always decode to u-law. You can use this element for all media types except application/x-feature. |
|||||||||||
|
bitrate |
Optional when orig-encoding is defined as g723. This is the bitrate of the original VoIp data stream. Values: 5.3 or 6.3. |
|||||||||||
The following tables show sample combinations of media-type-prefix, sample-rate, encoding, and bit-rate. Identical combinations of the rate, orig-encoding, and bitrate above can be used for audio/x-alaw-basic and audio/L16.
|
Sample media types for ulaw, 8,000 Hz. |
|---|
|
audio/basic;rate=8000 |
|
audio/basic;rate=8000;orig-encoding=g723 |
|
audio/basic;rate=8000;orig-encoding=g723;bitrate=5.3 |
|
audio/basic;rate=8000;orig-encoding=g723;bitrate=6.3 |
|
audio/basic;rate=8000;orig-encoding=g729 |
The table below shows Aurora combinations. If you use the 11,000 and 16,000 kHz rates, you must request the corresponding acoustic models as custom deliverables from Nuance.
|
Sample media types for Aurora |
|---|
|
application/x-feature;rate=8000;encoding=ES_202_050 |
|
application/x-feature;rate=11000;encoding=ES_202_050 |
|
application/x-feature;rate=16000;encoding=ES_202_050 |
Here are additional combinations:
|
Sample media types |
|---|
|
audio/L16;rate=8000 (8 KHz 16-bit linear) |
|
audio/L16;rate=8000 (8 KHz 16-bit linear) |
|
audio/x-alaw-basic;rate=8000 (8 KHz a-law) |
This parameter allows multiple values. If Aurora is one of the values, it must not be listed first.
Related parameter
