audio audioduration (s) 5 8.72 | text stringlengths 80 191 | text_ts stringlengths 111 237 | preconditioning stringclasses 10
values | start_time stringlengths 12 12 | end_time stringlengths 12 12 | speech_duration float32 5 8.72 | word_timestamps stringlengths 942 2.54k | source_file stringclasses 2
values |
|---|---|---|---|---|---|---|---|---|
american fourteen seventy nine I think you failed the callsign again fly present heading maintain five thousand | <|0.00|>american fourteen seventy nine I think you failed the callsign again fly present heading maintain<|3.90|><|3.90|>five thousand<|4.72|> | 00:00:33.546 | 00:00:38.550 | 5.004 | [{"word": "american", "start": 0.07399999999999807, "end": 0.2539999999999978}, {"word": "fourteen", "start": 0.29400000000000404, "end": 0.5940000000000012}, {"word": "seventy", "start": 0.7539999999999978, "end": 0.9740000000000038}, {"word": "nine", "start": 1.054000000000002, "end": 1.1940000000000026}, {"word": "I... | test2.mp3 | ||
alright American fourteen seventy nine present heading five thousand we will expect two six thank you | <|0.00|>alright American fourteen seventy nine present heading five thousand we will expect<|3.56|><|3.56|>two six thank you<|4.86|> | american fourteen seventy nine I think you failed the callsign again fly present heading maintain five thousand | 00:00:39.882 | 00:00:45.270 | 5.388 | [{"word": "alright", "start": 0.1980000000000004, "end": 0.31800000000000495}, {"word": "American", "start": 0.3580000000000041, "end": 0.5579999999999998}, {"word": "fourteen", "start": 0.7379999999999995, "end": 1.078000000000003}, {"word": "seventy", "start": 1.2580000000000027, "end": 1.4580000000000055}, {"word": ... | test2.mp3 | |
american fourteen seventy nine turn left heading two niner zero intercept the two six localizer maintain four thousand | <|0.00|>american fourteen seventy nine turn left heading two niner<|3.86|><|3.86|>zero intercept the two six localizer maintain four thousand<|7.32|> | alright American fourteen seventy nine present heading five thousand we will expect two six thank you | 00:01:00.490 | 00:01:08.214 | 7.724 | [{"word": "american", "start": 0.10999999999999943, "end": 0.36999999999999744}, {"word": "fourteen", "start": 0.44999999999999574, "end": 0.7299999999999969}, {"word": "seventy", "start": 0.8500000000000014, "end": 1.0899999999999963}, {"word": "nine", "start": 1.2100000000000009, "end": 1.5300000000000011}, {"word": ... | test2.mp3 | |
we got a communication problem are you hearing this radio clear cause Im saying what I need to say but something is not happening | <|0.00|>we got a communication problem are you hearing this radio clear cause<|3.66|><|3.66|>Im saying what I need to say but something is not happening<|7.06|> | american fourteen seventy nine turn left heading two niner zero intercept the two six localizer maintain four thousand | 00:01:29.162 | 00:01:36.534 | 7.372 | [{"word": "we", "start": 0.7579999999999956, "end": 0.7979999999999876}, {"word": "got", "start": 0.8979999999999961, "end": 1.0180000000000007}, {"word": "a", "start": 1.0579999999999927, "end": 1.0579999999999927}, {"word": "communication", "start": 1.097999999999999, "end": 1.637999999999991}, {"word": "problem", "s... | test2.mp3 | |
american fourteen seventy nine you said cleared the approach I need the runway assignment which is runway two six not two five left or two five right you are cleared visual approach two six | <|0.00|>american fourteen seventy nine you said cleared the approach I need the runway assignment which<|3.62|><|3.62|>is runway two six not two five left or two five right you are cleared visual<|7.12|><|7.12|>approach two six<|8.26|> | we got a communication problem are you hearing this radio clear cause Im saying what I need to say but something is not happening | 00:01:36.874 | 00:01:45.590 | 8.716 | [{"word": "american", "start": 0.1460000000000008, "end": 0.36599999999999966}, {"word": "fourteen", "start": 0.42600000000000193, "end": 0.666000000000011}, {"word": "seventy", "start": 0.7460000000000093, "end": 0.9060000000000059}, {"word": "nine", "start": 0.945999999999998, "end": 1.0260000000000105}, {"word": "yo... | test2.mp3 | |
cleared for the visual approach to runway two six american fourteen seventy nine | <|0.00|>cleared for the visual approach to runway two six<|3.68|><|3.68|>american fourteen seventy nine<|6.00|> | american fourteen seventy nine you said cleared the approach I need the runway assignment which is runway two six not two five left or two five right you are cleared visual approach two six | 00:01:48.266 | 00:01:54.550 | 6.284 | [{"word": "cleared", "start": 0.23399999999999466, "end": 0.4539999999999935}, {"word": "for", "start": 0.5139999999999958, "end": 0.5939999999999941}, {"word": "the", "start": 0.7339999999999947, "end": 0.813999999999993}, {"word": "visual", "start": 1.2339999999999947, "end": 1.4140000000000015}, {"word": "approach",... | test2.mp3 | |
fourteen seventy nine do you need a phone number you can call in and we can talk about the tape but you misread the last three transmissions without a callsign and without a runway assignment | <|0.00|>fourteen seventy nine do you need a phone number you can call in and we can talk about the tape<|3.74|><|3.74|>but you misread the last three transmissions without a callsign and without<|7.50|><|7.50|>a runway assignment<|8.28|> | cleared for the visual approach to runway two six american fourteen seventy nine | 00:01:54.986 | 00:02:03.574 | 8.588 | [{"word": "fourteen", "start": 0.2539999999999907, "end": 0.4939999999999998}, {"word": "seventy", "start": 0.5939999999999941, "end": 0.7539999999999907}, {"word": "nine", "start": 0.7939999999999969, "end": 0.8739999999999952}, {"word": "do", "start": 0.9339999999999975, "end": 0.9539999999999935}, {"word": "you", "s... | test2.mp3 | |
winter haven we are trying to take the advance here papa on a three way final for runway five we are shooting the rnav gps for runway five approach winter haven we will be a touch and go | <|0.00|>winter haven we are trying to take the advance here papa on a three way final for runway five we are shooting<|3.80|><|3.80|>the rnav gps for runway five approach winter haven we will be a touch and go<|7.22|> | 00:00:01.290 | 00:00:08.950 | 7.66 | [{"word": "winter", "start": 0.1499999999999999, "end": 0.31000000000000005}, {"word": "haven", "start": 0.34999999999999987, "end": 0.44999999999999996}, {"word": "we", "start": 0.51, "end": 0.53}, {"word": "are", "start": 0.5700000000000001, "end": 0.6499999999999999}, {"word": "trying", "start": 0.69, "end": 0.83000... | test1.m4a | ||
winter haven traffic four four five uniform sierra is a mile north of the airport on a midfield crosswind to get to a left downwind | <|0.00|>winter haven traffic four four five uniform sierra is a mile<|3.52|><|3.52|>north of the airport on a midfield crosswind to get to a left downwind<|7.68|> | winter haven we are trying to take the advance here papa on a three way final for runway five we are shooting the rnav gps for runway five approach winter haven we will be a touch and go | 00:00:10.986 | 00:00:18.934 | 7.948 | [{"word": "winter", "start": 0.15399999999999991, "end": 0.3739999999999988}, {"word": "haven", "start": 0.5539999999999985, "end": 0.6739999999999995}, {"word": "traffic", "start": 0.8339999999999996, "end": 1.2139999999999986}, {"word": "four", "start": 1.3539999999999992, "end": 1.4939999999999998}, {"word": "four",... | test1.m4a | |
traffic cherokee eight five sierra papa on a short final for runway five we will be a touch and go | <|0.00|>traffic cherokee eight five sierra papa on a short final for runway<|3.50|><|3.50|>five we will be a touch and go<|4.90|> | winter haven traffic four four five uniform sierra is a mile north of the airport on a midfield crosswind to get to a left downwind | 00:00:58.058 | 00:01:03.382 | 5.324 | [{"word": "traffic", "start": 0.26200000000000045, "end": 0.5020000000000024}, {"word": "cherokee", "start": 0.5820000000000007, "end": 0.8419999999999987}, {"word": "eight", "start": 0.9420000000000002, "end": 1.0619999999999976}, {"word": "five", "start": 1.142000000000003, "end": 1.3419999999999987}, {"word": "sierr... | test1.m4a | |
how about you just shut the fuck up if you are in a pick up truck and ain’t worried about flying an airplane then dickhead | <|0.00|>how about you just shut the fuck up if you are in a pick up truck and ain’t worried about<|3.64|><|3.64|>flying an airplane then dickhead<|5.28|> | traffic cherokee eight five sierra papa on a short final for runway five we will be a touch and go | 00:01:40.586 | 00:01:46.230 | 5.644 | [{"word": "how", "start": 0.13400000000000034, "end": 0.21399999999999864}, {"word": "about", "start": 0.31400000000000716, "end": 0.534000000000006}, {"word": "you", "start": 0.6340000000000003, "end": 0.6940000000000026}, {"word": "just", "start": 0.7939999999999969, "end": 0.9140000000000015}, {"word": "shut", "star... | test1.m4a |
atc-test-5s
Speech dataset prepared with Trelis Studio.
Statistics
| Metric | Value |
|---|---|
| Source files | 2 |
| Validation samples | 11 |
| Total duration | 4.6 minutes |
Columns
| Column | Type | Description |
|---|---|---|
audio |
Audio | Audio segment (16kHz) - speech only, silence stripped via VAD |
text |
string | Plain transcription (no timestamps) - backwards compatible |
text_ts |
string | Transcription WITH Whisper timestamp tokens (e.g., `< |
start_time |
string | Segment start in original audio (HH:MM:SS.mmm) |
end_time |
string | Segment end in original audio (HH:MM:SS.mmm) |
speech_duration |
float | Duration of speech in segment (excluding silence) |
word_timestamps |
list | Word-level timestamps (relative to speech-only audio) |
source_file |
string | Original audio filename |
VAD Processing
Audio segments are processed with Silero VAD to match faster-whisper inference:
- Silence is stripped from audio (only speech regions remain)
- Timestamps are relative to the concatenated speech audio
- This ensures training data matches inference behavior
Training Usage
For Whisper timestamp training, use the two-bucket approach:
- Bucket A (50%): Use
text- plain transcription without timestamps - Bucket B (50%): Use
text_ts- transcription with Whisper timestamp tokens
Usage
from datasets import load_dataset
dataset = load_dataset("Trelis/atc-test-5s")
Prepared with Trelis Studio
- Downloads last month
- 10