Request Payload

This sub-page helps the integrator to understand various different types of request payload based on the individual task or combination of tasks in that sequence that integrator wants to do.

Request Payload for Individual Task

ASR

{
    "pipelineTasks": [
        {
            "taskType": "asr",
            "config": {
                "language": {
                    "sourceLanguage": "xx"
                },
                "serviceId": "xxxxx--ssssss-d-ddd--dddd",
                "audioFormat": "wav",
                "samplingRate": 16000,
                "postprocessors": [
                    "itn"
                ]
            }
        }
    ],
    "inputData": {
        "input": [
            {
                "source": null
            }
        ],
        "audio": [
            {
                "audioContent": "{{generated_base64_content}}"
            }
        ]
    }
}

This response contains 2 major parameters listed below and detailed further down the section:

1. pipelineTasks

2. inputData

Parameter: pipelineTasks

Type: Array This parameter takes an array of tasks, in the form of dictionary of taskType and config, that are to be done by the integrator. In the above example, pipelineTasks takes only one dictionary (line 3-13) because integrator wants to do only ASR. taskType parameter takes String that takes the value asr

config parameter takes a Dictionary that contains following parameters:

Language

For ASR, language parameter only takes sourceLanguage which accepts ISO-639 Series Code of the language.

Service ID

serviceId parameter is obtained from the Pipeline Config Call response as described here.

Audio Format

audioFormat parameter accepts format of the audio which was recorded by the application.

• For Android, wav is preferred and

• For iOS, wav or flac is preferred.

However, the Server also accepts other well -known formats such as mp3.

Sampling Rate

Sampling Rate is determined by the application at which the audio is recorded. The Server accepts a minimum value of 8000 for samplingRate parameter.

Parameters other than taskType, serviceId and config are optional.

Parameter: inputData

inputData Parameter takes the actual input from the integrator on which the individual task has to be done. It can take the input either via input parameter or audio parameter depending on the task to be done. Since ASR is done on audio input data, for ASR,

• input parameter is optional, of no use for ASR but

• audio parameter is mandatory.

audio parameter takes audioContent parameter which accepts base64 String of the actual audio captured.

If audioFormat or/and samplingRate parameter is/are sent, integrator should make sure that these values correspond to the actual recorded audio.

Translation

{
    "pipelineTasks": [
        {
            "taskType": "translation",
            "config": {
                "language": {
                    "sourceLanguage": "hi",
                    "targetLanguage": "xx"
                },
                "serviceId": "xxxxx--ssssss-d-ddd--dddd"
            }
        }
    ],
    "inputData": {
        "input": [
            {
                "source": "मेरा नाम विहिर है और मैं भाषाावर्ष यूज कर रहा हूँ"
            }
        ],
        "audio": [
            {
                "audioContent": null
            }
        ]
    }
}

This response contains 2 major parameters listed below and detailed further down the section:

1. pipelineTasks

2. inputData

Parameter: pipelineTasks

Type: Array This parameter takes an array of tasks, in the form of dictionary of taskType and config, that are to be done by the integrator. In the above example, pipelineTasks takes only one dictionary (line 3-12) because integrator wants to do only Translation. taskType parameter takes String that takes the value translation

config parameter takes a Dictionary that contains following parameters:

Language

For Translation, language parameter takes both sourceLanguage and targetLanguage which accepts ISO-639 Series Code of the language.

Service ID

serviceId parameter is obtained from the Pipeline Config Call response as described here.

Parameter: inputData

inputData Parameter takes the actual input from the integrator on which the individual task has to be done. It can take the input either via input parameter or audio parameter depending on the task to be done. Since Transaltion is done on digital text input data, for Translation,

• input parameter is mandatory and

• audio parameter is optional and of no use for Translation.

input parameter takes source parameter which accepts digital text string.

TTS

{
    "pipelineTasks": [       
        {
            "taskType": "tts",
            "config": {
                "language": {
                    "sourceLanguage": "gu"
                },
                "serviceId": "xxxxx--ssssss-d-ddd--dddd",
                "gender": "female"
            }
        }
    ],
    "inputData": {
        "input": [
            {
                "source": "મારું નામ વિહીર છે અને હું ભાષાવર્ષનો ઉપયોગ કરી રહ્યો છું"
            }
        ],
        "audio": [
            {
                "audioContent": null
            }
        ]
    }
}

This response contains 2 major parameters listed below and detailed further down the section:

1. pipelineTasks

2. inputData

Parameter: pipelineTasks

Type: Array This parameter takes an array of tasks, in the form of dictionary of taskType and config, that are to be done by the integrator. In the above example, pipelineTasks takes only one dictionary (line 3-12) because integrator wants to do only TTS. taskType parameter takes String that takes the value tts

config parameter takes a Dictionary that contains following parameters:

Language

For TTS, language parameter only takes sourceLanguage which accepts ISO-639 Series Code of the language.

Service ID

serviceId parameter is obtained from the Pipeline Config Call response as described here.

gender

gender parameter takes a string input which can either be:

• male

• female

gender parameter tells the server that integrator is requesting the generated speech in either male or female voice.

Parameter: inputData

inputData Parameter takes the actual input from the integrator on which the individual task has to be done. It can take the input either via input parameter or audio parameter depending on the task to be done. Since TTS is done on digital text input data, for TTS,

• input parameter is mandatory and

• audio parameter is optional and of no use for TTS.

input parameter takes source parameter which accepts digital text string.

Request Payload for Combination of Tasks in specific sequence

ASR+Translation

{
    "pipelineTasks": [
        {
            "taskType": "asr",
            "config": {
                "language": {
                    "sourceLanguage": "xx"
                },
                "serviceId": "xxxxx--ssssss-d-ddd--dddd",
                "audioFormat": "flac",
                "samplingRate": 16000
            }
        },
        {
            "taskType": "translation",
            "config": {
                "language": {
                    "sourceLanguage": "xx",
                    "targetLanguage": "yy"
                },
                "serviceId": "xxxxx--ssssss-d-ddd--mfkds"
            }
        }
    ],
    "inputData": {
        "input": [
            {
                "source": null
            }
        ],
        "audio": [
            {
                "audioContent": "{{generated_base64_content}}"
            }
        ]
    }
}

Parameter: pipelineTasks

Type: Array This parameter takes an array of tasks, in the form of dictionary of taskType and config, that are to be done by the integrator. In the above example, pipelineTasks takes two dictionaries:

• Line 3 to 13 i.e., ASR Dictionary

• Line 14 to 23 i.e., Translation Dictionary

because integrator wants to do ASR of the input voice followed by Translation of the digital text.

Line Number 7 and Line Number 18 are connected with below understanding. Consider a use-case described below:

Integrator wants to speak in say Hindi language and wants to see the translated output in Marathi. For this to happen, integrator has to:

• Convert the Audio integrator has spoken to digital text i.e., ASR of Hindi

• Translate this digital Hindi text to Marathi digital text i.e., Translation from Hindi to Marathi

Therefore, the language code for ASR that is to be inserted in Line 7, shall be hi, i.e., ISO 639 series code for Hindi. Once this Hindi digital text is generated, the same shall be translated to Marathi, therefore the source language code for Translation that is to be inserted in Line 18, shall also be hi, which means that language code in Line 7 and Line 18 shall be same.

For Target Language the code to be inserted in Line 19 shall be mr, i.e., ISO 639 series code for Marathi.

Understanding of all other parameters remains same as described above in Request Payload for Individual Task.

Translation+TTS

{
    "pipelineTasks": [
        {
            "taskType": "translation",
            "config": {
                "language": {
                    "sourceLanguage": "hi",
                    "targetLanguage": "yy"
                },
                "serviceId": "xxxxx--ssssss-d-ddd--dddd"
            }
        },
        {
            "taskType": "tts",
            "config": {
                "language": {
                    "sourceLanguage": "yy"
                },
                "serviceId": "xxxxx--ssssss-d-ddd--csdcxsa",
                "gender": "female"
            }
        }
    ],
    "inputData": {
        "input": [
            {
                "source": "मेरा नाम विहिर है और मैं भाषाावर्ष यूज कर रहा हूँ"
            }
        ],
        "audio": [
            {
                "audioContent": null
            }
        ]
    }
}

Parameter: pipelineTasks

Type: Array This parameter takes an array of tasks, in the form of dictionary of taskType and config, that are to be done by the integrator. In the above example, pipelineTasks takes two dictionaries:

• Line 3 to 12 i.e., Translation Dictionary

• Line 13 to 22 i.e., TTS Dictionary

because integrator wants to do Translation of a digital text followed by TTS.

Line Number 8 and Line Number 17 are connected with below understanding. Consider a use-case described below:

Integrator wants to translate say from Hindi to Marathi language and wants to hear the output in Marathi. For this to happen, integrator has to:

• Translate this digital Hindi text to Marathi digital text i.e., Translation from Hindi to Marathi

• Generate this Marathi text speech i.e., TTS of the Marathi digital text.

Therefore, the source language code for Translation that is to be inserted in Line 7, shall be hi, i.e., ISO 639 series code for Hindi. The target language code to be inserted in Line 8 shall be mr, i.e., ISO 639 series code for Marathi.

Speech shall be generated in Marathi which means the language code to be inserted in Line 17 shall be mr, same as Line 8.

Understanding of all other parameters remains same as described above in Request Payload for Individual Task.

ASR+Translation+TTS

{
    "pipelineTasks": [
        {
            "taskType": "asr",
            "config": {
                "language": {
                    "sourceLanguage": "xx"
                },
                "serviceId": "xxxxx--ssssss-d-ddd--dddd",
                "audioFormat": "flac",
                "samplingRate": 16000
            }
        },
        {
            "taskType": "translation",
            "config": {
                "language": {
                    "sourceLanguage": "xx",
                    "targetLanguage": "yy"
                },
                "serviceId": "xxxxx--ssssss-d-ddd--fwsd"
            }
        },
        {
            "taskType": "tts",
            "config": {
                "language": {
                    "sourceLanguage": "yy"
                },
                "serviceId": "xxxxx--ssssss-d-ddd--fvdfg",
                "gender": "female"
            }
        }
    ],
    "inputData": {
        "input": [
            {
                "source": null
            }
        ],
        "audio": [
            {
                "audioContent": "{{generated_base64_content}}"
            }
        ]
    }
}

Parameter: pipelineTasks

Type: Array This parameter takes an array of tasks, in the form of dictionary of taskType and config, that are to be done by the integrator. In the above example, pipelineTasks takes two dictionaries:

• Line 3 to 13 i.e., ASR Dictionary

• Line 14 to 23 i.e., Translation Dictionary

• Line 24 to 33 i.e., TTS Dictionary

because integrator wants to do ASR of the voice input, then Translation of a digital text followed by TTS.

Line Number 7 and Line Number 18 are connected and Line Number 19 and Line Number 28 with below understanding. Consider a use-case described below:

Integrator wants to speak in say Hindi language and wants to hear the translated output in Marathi. For this to happen, integrator has to:

• Convert the Audio integrator has spoken to digital text i.e., ASR of Hindi

• Translate this digital Hindi text to Marathi digital text i.e., Translation from Hindi to Marathi

• Generate this Marathi text speech i.e., TTS of the Marathi digital text.

Therefore, the language code for ASR that is to be inserted in Line 7, shall be hi, i.e., ISO 639 series code for Hindi. Once this Hindi digital text is generated, the same shall be translated to Marathi, therefore the source language code for Translation that is to be inserted in Line 18, shall also be hi, which means that language code in Line 7 and Line 18 shall be same.

The target language code to be inserted in Line 19 shall be mr, i.e., ISO 639 series code for Marathi.

Speech shall be generated in Marathi which means the language code to be inserted in Line 28 shall be mr, same as Line 19.

Understanding of all other parameters remains same as described above in Request Payload for Individual Task.

Pre-Processors and Post-Processors within Compute Request

ASR

In Automatic Speech Recognition (ASR) systems, preprocessors and postprocessors play a crucial role in refining the audio input and enhancing the textual output, respectively. Below, we provide details on the available preprocessors and postprocessors, along with an example of how to configure them in your request body.

Preprocessors

Voice Activity Detection (VAD)

• Syntax: "preProcessors": ["vad"]

• Function: VAD allows audio content longer than 30 seconds to be passed and processed. It helps identify voice activity to ensure that only the detected voice activity is processed, reducing the load and improving the efficiency of the ASR system.

Postprocessors

Inverse Text Normalization (ITN)

• Syntax: "postProcessors": ["itn"]

• Function: ITN converts spoken numbers and dates into their written forms. For example, the ASR would output "two thousand and twenty three" as "2023".

Punctuation

• Syntax: "postProcessors": ["punctuation"]

• Function: This postprocessor adds punctuations to the ASR output, making the text more readable and closer to natural written language.

  Example:

  • ASR Output: "hello how are you"

  • Punctuation Output: "Hello, how are you?"

The configuration of preprocessors and postprocessors can be included within the config section of the request body as shown below:

"config": {
    "language": {
        "sourceLanguage": "xx"
    },
    "serviceId": "xxxxx--ssssss-d-ddd--dddd",
    "audioFormat": "flac",
    "samplingRate": 16000,
    "preProcessors": ["vad"],
    "postProcessors": ["itn", "punctuation"]
}

Translation

In translation (NMT) systems, postprocessors play a crucial role in refining the textual output to meet specific needs. Below, we provide details on the postprocessor available for translation, along with an example of how to configure it in your request body.

Postprocessors

Glossary

• Syntax: "postProcessors": ["glossary"]

• Function: The glossary postprocessor allows users to create a list of glossary terms within Bhashini Udyat under the My Profile Section once logged in. Glossary terms created are unique for each Bhashini Inference API Key generated under app names. This postprocessor ensures that specific nouns and noun phrases have their translations overridden as per the user's glossary.

Example:

• Default Translation: "Digital India Bhashini Division" is translated to "डिजिटल इंडिया भैसिनी प्रभाग".

• With Glossary Term: If the glossary term between English and Hindi is entered as "डिजिटल इंडिया भाषिणी डिवीज़न", this will override the default translation.

Link to Access My Profile Page and Generate Keys and Glossary: Bhashini Udyat Profile Page

Glossary Terms Usage: Glossary terms help provide customized solutions for domain-specific translations, ensuring accuracy and context relevance in the translated output.

Example of Glossary Usage:

• Case sensitivity handling (Ex: Glossary entry - English to Hindi as IPO -> आईपीओ).

• Glossary entries will work by default for:

  1. Entered noun/noun phrase (e.g., IPO)

  2. Capitalized case (Ipo)

  3. Lower case (ipo)

  4. Upper case ( IPO)

  5. Reverse case (if आईपीओ is the source, the target is IPO when translating from Hindi to English).

Configuration Example

The configuration of the glossary postprocessor can be included within the config section of the request body as shown below:

"config": {
    "language": {
        "sourceLanguage": "hi",
        "targetLanguage": "xx"
    },
    "postProcessors": ["glossary"]
    "serviceId": "xxxxx--ssssss-d-ddd--dddd"
}