Datasets:
audio audioduration (s) 60 60 | transcript stringclasses 4
values | timecodes stringclasses 2
values | accent stringclasses 3
values |
|---|---|---|---|
Speaker 1 [00:00:00.000 - 00:00:00.790]: Alright, right.
Speaker 2 [00:00:01.670 - 00:00:02.630]: Here we are again.
Speaker 1 [00:00:01.760 - 00:00:02.130]: Get.
Speaker 2 [00:00:03.290 - 00:00:03.410]: [sniffling]
Speaker 2 [00:00:05.170 - 00:00:07.170]: Your, uh, your other account got kicked out.
Speaker 2 [00... | overlapping | Arabic-influenced English | |
Speaker 1 [00:00:00.000 - 00:00:00.660]: You know.
Speaker 2 [00:00:00.760 - 00:00:02.500]: Okay, yeah, I gotcha.
Speaker 2 [00:00:04.400 - 00:00:05.840]: Okay, choose one of the three.
Speaker 2 [00:00:06.870 - 00:00:13.400]: Would you rather be-- which, this is kind of stupid. Again, okay, in this scenario you're ... | overlapping | Arabic-influenced English | |
Speaker 2 [00:00:00.000 - 00:00:02.150]: If, if this is recording or not.
Speaker 1 [00:00:05.750 - 00:00:08.390]: Do you want me to be a narrator?
Speaker 1 [00:00:08.990 - 00:00:09.020]: [clicking]
Speaker 2 [00:00:09.990 - 00:00:11.250]: What narrator?
Speaker 1 [00:00:11.970 - 00:00:13.000]: I will be-
Speaker... | overlapping | Filipino English | |
00:00:00,060 --> 00:00:09,660 [Speaker 0]
He's not, the not-- that number has expired, so this week, they will, they will up-- they will, they will do it. He has sent it for them to replace it for him. So they, they will up-
00:00:09,780 --> 00:00:10,620 [Speaker 1]
Which number?
00:00:10,700 --> 00:00:17,320 [Speake... | non-overlapping | West African English |
Accented English Conversational Audio
This dataset contains conversational English audio recordings with human-verified transcripts, grouped by speaker accent. The recordings are taken from the yapdo dataset. The conversations feature natural, spontaneous speech between multiple speakers, captured across three accent communities.
Each sample is the first 60 seconds of a longer recording.
Transcription Process
All transcripts go through at least two full human review passes. First, a native-speaker transcriber reviews and revises an AI-generated transcript using our custom-built transcript editor. Transcribers are provided the conversation as channel-separated audio, which enables precise speaker diarization correction and makes it easier to identify and label each speaker accurately. They work through the entire audio file to produce a full verbatim transcript with precise timestamps down to the millisecond, labeled speaker turns for diarization, and audio event tags for non-speech sounds such as [phone buzzing], [laughing], or [door closing]. Native fluency allows transcribers to accurately capture code-switching, overlaps, disfluencies, colloquialisms, and other conversational details that are difficult to capture without deep familiarity with the speakers and language variety.
Once the first pass is complete, the file is handed off to a senior reviewer with a proven track record of producing high-quality transcripts. The senior reviewer listens through the full audio file again and manually corrects any remaining transcription errors, spelling issues, timestamp inconsistencies, speaker label issues, or missed audio events. In addition, a senior native project lead performs rigorous spot checks across completed files to monitor quality, enforce consistency, and identify any recurring issues that need to be corrected across the project.
Timecodes
Transcripts are labeled as either overlapping or non-overlapping based on their timecode structure. In overlapping transcripts, speaker turns may have timestamps that overlap, capturing moments where multiple speakers talk simultaneously. In non-overlapping transcripts, timecodes are strictly sequential — no two speaker turns share the same time range — though the underlying audio may still contain overlapping speech. We provide transcripts tailored to any style guide, and our customers have requested a wide range of formatting and annotation conventions.
- Downloads last month
- 188