Skip to content

Samples

Urdu speech to text: a real sample transcript

What the free preview returns for a short Urdu recording (اردو), with the audio, the sources and the mistakes left in.

Updated

Urdu (اردو), 20 seconds, taken 2026-10-27. Sentences read aloud by a volunteer. Output of the free preview with the language set to Urdu, not corrected.

CC0 1.0

The transcript, as the engine returned it

  1. آئین کیا کہتا ہے
  2. ایشہ کا بغلط اشی سکوار سے سابر اور نانڈ روپ
  3. یعنی کوئی کافیت دائمی نہیں رہتی
  4. پاکستان میں سجا منفرد حیرت سیو
  5. اوپ نے بغلط اشی سے فتح چینی ویسٹنڈی سفاتے
  6. سڑکوں کی بندے سے تیل کی سپلائی میں تاتول ملک میں تیل کی بہوران کا خدشہ

Download this transcript as SRT (written by the same exporter the app uses; the other formats are TXT, timestamped TXT, VTT, Markdown, JSON and CSV).

Against the reference sentences

Word error rate 43.4% (23 edits over 53 words), character error rate 19.9%; normalisation: plain (explained on the samples page).

The sentences the speaker was asked to read:

  1. آئین کیا کہتا ہے۔
  2. ایشیا کپ بنگلہ دیشی اسکواڈ سے صابر اور انعام ڈراپ
  3. یعنی کوئی کیفیت دائمی نہیں رہتی۔
  4. پاکستان میں سجا منفرد ہیئراسٹائل شو
  5. ہوپ نے بنگلہ دیش سے فتح چھین لی ویسٹ انڈیز فاتح
  6. سڑکوں کی بندش سے تیل کی سپلائی میں تعطل ملک میں تیل کے بحران کا خدشہ
In the reference sentenceIn the transcript
ایشیاایشہ
کپکا
بنگلہبغلط
دیشیاشی
اسکواڈسکوار
صابرسابر
انعامنانڈ
ڈراپروپ
کیفیتکافیت
ہیئراسٹائلحیرت
شوسیو
ہوپاوپ
بنگلہبغلط
دیشاشی
چھین(missing)
لی(missing)
ویسٹچینی
انڈیزویسٹنڈی
فاتحسفاتے
بندشبندے
تعطلتاتول
کےکی
بحرانبہوران

Where the recording comes from

  • Source: Mozilla Common Voice 22.0, test split (CC0 1.0 public domain dedication, https://commonvoice.mozilla.org/en/datasets, https://creativecommons.org/publicdomain/zero/1.0/; copy used: https://huggingface.co/datasets/fsicoli/common_voice_22_0, whose card also says cc0-1.0).
  • Licence: CC0 1.0 (creativecommons.org/publicdomain/zero/1.0/).
  • Which recording: One volunteer speaker (Pakistan Urdu): the speaker with the most clips in the part of the test archive we read, their clips in the test file's order that are 2.5 to 7 seconds long and have no down-votes, until about 24 seconds were collected. Chosen by that rule, not by how well the engine did.
  • Clip ids: common_voice_ur_30643794, common_voice_ur_30643795, common_voice_ur_30643797, common_voice_ur_30643862, common_voice_ur_30643863, common_voice_ur_30643864.

How every sample was made, step by step, is on the samples page. To try your own Urdu recording, use the Urdu transcription page.

Questions

What did the engine get wrong in this Urdu clip?
21 words were substituted, 2 were missing and 0 were extra, out of 53 reference words. For example: ایشیا became ایشہ; کپ became کا; بنگلہ became بغلط.
Can I use the Urdu recording?
CC0 1.0. The source is named above.

Try it with your own file

Upload a recording and read the transcript.