ChivoxAI
sign in

/docs

en.pred.score

Kernel Description

  • Evaluate speech on paragraph reading, outputting scores of overall、fluency、accuracy、integrity、each sentence and each word.
    - Support specifying British or American pronunciation
    - Support detection of tone of stress, rising tone at end of sentence
    - Support difficulty adjustment
    - Support real-time feedback of spoken words and evaluation scores

Request Parameters

Parameter Type Required Description
coreType String true "en.pred.score" means English paragraph kernel.
accent int false Specify British or American pronunciation. 1, 2 or 3, Default value is 3.
  • 1. Accept British pronunciation.
  • 2: Accept American pronunciation.
  • 3: Accept both British and American pronunciation.
Now about 4000 words have been marked British and American pronunciation in our dictionary.
rank int true Default value is 100, means full score is 100. You can set any number.
precision float false Rating precision, 1 or 0.5 . Default value is 1
attachAudioUrl number false Enable to return audio url or not.
  • 0: Disable
  • 1: Enable
refText String true The English paragraph you want to envalate the recording with.
More than 20 words are recommended
Accepted text punctuations refer to Text Symbol Description
result object false Return result detailed settings
- details object false Detailed scoring options
- - word int false Enable to return detailed word score. Default value is 0.
  • 0: Disable.
  • 1: Enabled
- - connti number false Whether to enable detection of loss of plosion and liaison, Default value is 0.
  • 0: Disable
  • 1: Enable
- - gop_adjust number false Increase or reduce the evaluation scores.
The value range is between[-1,1] and the precision is 0.1 .
Default value is 0, which means using original scores.
If the value is between [0,1]. The larger the value is, the higher degree the score will be increased.
If the value is between [-1,0], the smaller the value is, the higher degree the scores will be reduced.
- - ext_cur_wrd number false Enable to return the words that have been read in real time，Default value is 0. Support multiple modes
0: Disable.
1: General mode, the returned results are output from front to back in the order of the text;
2: Recitation mode, the returned results are strictly in accordance with the text order from front to back; if the previous sentence is missed, even if the following sentence is read normally, it will not be returned;
3: Free mode, returns the content of reference text actually read by users, regardless of the order of text;
Stress、Rising tone(optional)
  • Symbols can be used to mark the position of stress tone, rising tone in sentences.
  • Stress、rising tone are not counted in the total score.
Type Symbols Sample Description
Stress (s:1) Are(s:1) you ok? "Are" is maked with stressed.
Rising tone (t:1) Are you ok(t:1)? "Ok" is marked with rising tone.
Both stress and rising tone (s:1,t:1) Are you ok(s:1,t:1)? Word “ok” is marked as both stress and rising tone, use commas to separate marks.
Loss of Plosion、Liaison(optional)
  • Need to be marked with a symbol in the refText field
  • Not counted in the total score
Type Sample Description
Loss of Plosion What(c:1) time is it now? Loss of Plosion occurred between What and time.
Liaison Nice to meet(c:1) you. Liaison occurred between meet and you.

Supported rules，Please refer toClick to view.

Sample Code

{
    "coreType": "en.pred.score",
    "refText": "It was Sunday. I never get up early on Sundays. I sometimes stay in bed until lunchtime. Last Sunday I got up very late. I looked out of the window. It was dark outside. ",
	"accent": 1,
    "rank": 100,
    "precision": 0.5,
    "attachAudioUrl": 1,
    "result": 
	{
		"details": 
		{
            "word": 1,
            "gop_adjust": 0,
		    "ext_cur_wrd": 0
		}
    }    
}

Introduction to Real-time return Results

Key Type Description Recommendations
eof int Whether the final evaluation result
0: Indicates that the result is returned in real time during the recording process;
1: Indicates that the result is the final evaluation result after the recording is completed;		 Suggest to display
result object Scoring result
- rec string recognized words in real time Suggest to display
- details object details of recognized words
- - - text string word Suggest to display
- - - score number word pronunciation score Suggest to display
- - - beginindex number The word first letter's index Suggest to display
- - - endindex number The word last letter's index. Suggest to display

Example of Real-time return Results

{
	"tokenId": "76****d8-****-****-****-7a****db****",
	"result": {
	    "rec": "it was sunday ",
		"details": [[{
			"score": 100,
			"text": "it",
			"endindex": 1,
			"beginindex": 0
		},
		......
		{
			"score": 97,
			"text": "sunday",
			"endindex": 12,
			"beginindex": 7
		}]],
		"phnspeed": 625
	},
	"eof": 0
}}

Example of Returned Results

  • The results are returned using the JSON structure,Example

Introduction to Returned Results

Key Type Description Recommendations
uuid string Unique id of audio. Save uuid in database in case of need. For reference
audioUrl string Audio url.

Download audio in http protocol: Just add ".mp3" at the end of the URL.
Download audio in https protocol: Remove the port number, add ".mp3" at the end, and add "https://" to the prefix.
e.g.https://download.cloud.chivox.com/XXXXXXXX.mp3

Tips:
  • Chivox save audio for one week. Please save audio to your own server for long-term storage.
  • The domain name will be different for different access areas, and the actual return shall prevail.
 For reference
result object Scoring result
- wavetime number Audio duration (Unit:milliseconds) For reference
- overall number Overall score Suggest to display
- integrity number Score of integrity. Suggest to display
- accuracy number Score of accuracy(pronunciation evaluation of the read parts). Suggest to display
- fluency object Score of fluency.
- - pause number Number of pauses. Suggest to display
- - overall number Overall score of fluency Suggest to display
- - speed number Speech rate, which is the number of words read per minute. Suggest to display
- oov_words array The word is not in Chivox’s dictionary.
Please feedback these words to Chivox technical support engineer.		 For reference
- details array Sentence detail score
- - text string sentence text For reference
- - score number Sentence pronunciation score. Suggest to display
- - is_leak number Whether the sentence is missing reading. Values are 0，1。
0: no missing reading。
1: missing reading 。		 Suggest to display
- - start number The start time of the word in audio (Unit:milliseconds). For reference
- - end number The end time of the word in audio (Unit: milliseconds). For reference
- - words array For detailed word scores, you need to enable detailed word scores in the request parameter("word": 1)
- - - text string word For reference
- - - score number pronunciation score. For reference
- - - is_err number Type of error。
0: Normal reading
1: superfluous reading
2: missing reading		 Suggest to display
- - - accent int 0, 1, 2 or 3:
  • 0: Its pronunciations are not marked with American or British accent.
  • 1: British pronunciation.
  • 2: American pronunciation.
  • 3: British pronunciation is the same as American pronunciation.
 Suggest to display
- - - beginindex number The word first letter's index. Suggest to display
- - - endindex number The word last letter's index. Suggest to display
- - - start number The start time of the word in the audio (Unit:milliseconds). For reference
- - - end number The end time of the word in audio (Unit: milliseconds). For reference
- - - stressref number 0 or 1:
0: Not marked to be stressed.
1: Marked to be stressed.		 Suggest to display
- - - stressscore number 0 or 1:
0: Not stressed in audio.
1: Stressed in audio.		 Suggest to display
- - - toneref number 0 or 1:
0: Not marked to rising tone.
1: Marked to rising tone.		 Suggest to display
- - - tonescore number 0 or 1:
0: Not reading with rising tone.
1: Reading with rising tone.		 Suggest to display
- - - conn_ref number Rule type(Loss of Plosion or Liaison)
0, No rules
1,Indicates liaison
2, Indicates loss of plosion
3, It can be either liaison or loss of plosion.
Note：When conn_ref is 3, please choose according to the actual needs of investigation, whether it is type 1 or type 2		 Suggest to display
- - - conn_type number Detect result(Loss of Plosion or Liaison)
0，No liaison or loss of plosion.
1，liaison
2，loss of plosion
conn_ref=conn_type，Indicates that the user’s pronunciation is consistent with the mark.		 Suggest to display
- info object
- - tipId number Audio quality
For details, please refer to TipId field description		 For reference
- - tips string Error information

1. "post proc failed"
2. "empty or too short wav!"
3. "fa forceout!"
4. "fa decode failed!"
5. "rec decode failed!"

These errors may be that the audio was not recorded successfully or the engine did not detect the user's valid voice, which can prompt the user to re-record.		 For reference

Note: The other returned parameters that are not shown above are internal ones, which you don't need to pay much atttention to.

SDK, API, MCP and Function Calling documentation on this site.