Worked example · stage 1 of 15
Transcript segments
Diarised, timecoded transcript units — the semantic source of truth.
This material is fictional. It was invented for the package to demonstrate the method.
[
{
"unit_id": "u1",
"source_id": "audio_host_iso",
"speaker_id": "host",
"range": {
"start": {
"value": 173232,
"rate": 24
},
"duration": {
"value": 70,
"rate": 24
},
"source_timecode": "00:12:03.200",
"end_timecode": "00:12:06.100"
},
"level": "sentence",
"text": "So you are saying the transcript itself is not enough?",
"verbatim_text": "So you are saying the transcript itself is not enough?",
"language": "en",
"confidence": 0.98,
"speaker_confidence": 0.99,
"speech_acts": [
"question"
],
"propositions": [
{
"id": "p1",
"text": "The transcript may be insufficient for editing.",
"polarity": "positive",
"modality": "possible"
}
],
"entity_ids": [
"transcript_system"
],
"claim_ids": [],
"context_dependencies": [],
"emotion": {
"labels": [
"curiosity"
],
"valence": 0,
"arousal": 0.35,
"control": 0.2,
"certainty": 0.3
},
"links": [
{
"relation": "answered_by",
"target_id": "u2",
"confidence": 0.99
}
],
"review_state": "approved",
"human_corrections": [],
"metadata": {}
},
{
"unit_id": "u2",
"source_id": "audio_guest_iso",
"speaker_id": "guest",
"range": {
"start": {
"value": 173316,
"rate": 24
},
"duration": {
"value": 312,
"rate": 24
},
"source_timecode": "00:12:06.480",
"end_timecode": "00:12:19.500"
},
"level": "speaker_turn",
"text": "Not exactly. The words are enough to search, but not enough to make every cut. A sentence can look complete and still depend on what ‘it’ means, which question was asked, or whether the speaker corrects themselves ten seconds later.",
"verbatim_text": "Not exactly. The words are enough to search, but not enough to make every cut. A sentence can look complete and still depend on what ‘it’ means, which question was asked, or whether the speaker corrects themselves ten seconds later.",
"language": "en",
"confidence": 0.97,
"speaker_confidence": 0.99,
"speech_acts": [
"answer",
"qualification",
"claim"
],
"propositions": [
{
"id": "p2",
"text": "Words are sufficient for search.",
"polarity": "positive",
"modality": "asserted"
},
{
"id": "p3",
"text": "Words alone are insufficient for every edit.",
"polarity": "negative",
"modality": "asserted"
},
{
"id": "p4",
"text": "Meaning can depend on referents, question scope, or later correction.",
"polarity": "positive",
"modality": "asserted"
}
],
"entity_ids": [
"transcript_system",
"editor"
],
"claim_ids": [
"claim_context_01"
],
"context_dependencies": [
{
"type": "question_scope",
"resolved": true,
"target_id": "u1",
"severity": "high",
"note": "This turn answers the host’s question."
}
],
"emotion": {
"labels": [
"certainty"
],
"valence": 0.1,
"arousal": 0.4,
"control": 0.7,
"certainty": 0.9
},
"links": [
{
"relation": "answers",
"target_id": "u1",
"confidence": 0.99
},
{
"relation": "exemplified_by",
"target_id": "u4",
"confidence": 0.96
}
],
"review_state": "approved",
"human_corrections": [],
"metadata": {}
},
{
"unit_id": "u3",
"source_id": "audio_host_iso",
"speaker_id": "host",
"range": {
"start": {
"value": 173760,
"rate": 24
},
"duration": {
"value": 22,
"rate": 24
},
"source_timecode": "00:12:20.040",
"end_timecode": "00:12:20.950"
},
"level": "sentence",
"text": "Give me an example.",
"verbatim_text": "Give me an example.",
"language": "en",
"confidence": 0.99,
"speaker_confidence": 0.99,
"speech_acts": [
"question",
"request"
],
"propositions": [],
"entity_ids": [],
"claim_ids": [],
"context_dependencies": [],
"emotion": {
"labels": [
"curiosity"
],
"valence": 0.1,
"arousal": 0.4,
"control": 0.3,
"certainty": 0.5
},
"links": [
{
"relation": "answered_by",
"target_id": "u4",
"confidence": 0.99
}
],
"review_state": "approved",
"human_corrections": [],
"metadata": {}
},
{
"unit_id": "u4",
"source_id": "audio_guest_iso",
"speaker_id": "guest",
"range": {
"start": {
"value": 173789,
"rate": 24
},
"duration": {
"value": 442,
"rate": 24
},
"source_timecode": "00:12:21.210",
"end_timecode": "00:12:39.600"
},
"level": "speaker_turn",
"text": "Yesterday I said, ‘We doubled conversion.’ That sounds like a result. But the next sentence was, ‘Only in the returning-customer group, and the sample was small.’ If an editor takes the first sentence by itself, the words are real, but the claim has changed.",
"verbatim_text": "Yesterday I said, ‘We doubled conversion.’ That sounds like a result. But the next sentence was, ‘Only in the returning-customer group, and the sample was small.’ If an editor takes the first sentence by itself, the words are real, but the claim has changed.",
"language": "en",
"confidence": 0.98,
"speaker_confidence": 0.99,
"speech_acts": [
"example",
"claim",
"qualification"
],
"propositions": [
{
"id": "p5",
"text": "Conversion doubled.",
"polarity": "positive",
"modality": "asserted"
},
{
"id": "p6",
"text": "The result applied only to returning customers.",
"polarity": "positive",
"modality": "asserted"
},
{
"id": "p7",
"text": "The sample was small.",
"polarity": "positive",
"modality": "asserted"
},
{
"id": "p8",
"text": "Removing the qualifications changes the claim.",
"polarity": "positive",
"modality": "asserted"
}
],
"entity_ids": [
"returning_customers",
"sample"
],
"claim_ids": [
"c1",
"c2"
],
"context_dependencies": [
{
"type": "question_scope",
"resolved": true,
"target_id": "u3",
"severity": "medium",
"note": "The turn is explicitly an example."
},
{
"type": "qualification",
"resolved": true,
"target_id": "c2",
"severity": "critical",
"note": "c2 qualifies c1 and must remain linked."
}
],
"emotion": {
"labels": [
"certainty",
"concern"
],
"valence": -0.1,
"arousal": 0.5,
"control": 0.7,
"certainty": 0.9
},
"links": [
{
"relation": "answers",
"target_id": "u3",
"confidence": 0.99
},
{
"relation": "summarized_by",
"target_id": "u5",
"confidence": 0.95
}
],
"review_state": "approved",
"human_corrections": [],
"metadata": {
"fictional_example": true
}
}
]