Datasets:

resproj007
/

severity_comparison_results

Modalities:

Text

Formats:

Size:

Libraries:

Dataset card Data Studio Files Files and versions

xet

Community

Dataset Viewer

Auto-converted to Parquet Duplicate

Split (1)

train · 1 rows

reference string	reference_distribution dict	datasets dict	ranking list	gt_ranking list	gt_reference_distribution dict	severity_labels list	classifier string
original_speech	{ "Mild": 0.1575, "Moderate": 0.35125, "Normal": 0.415, "Severe": 0.07625 }	{ "original_speech": { "name": "original_speech", "repo": "resproj007/original_pathological_speech", "total_samples": 800, "distribution": { "Mild": 0.1575, "Moderate": 0.35125, "Normal": 0.415, "Severe": 0.07625 }, "counts": { "Mild": 126, "Moderate": 281, ...	[ { "name": "original_speech", "jensen_shannon": 0, "cosine_similarity": 1, "kl_divergence": 0, "total_variation": 0, "n": 800 }, { "name": "orpheus_tts", "jensen_shannon": 0.041223, "cosine_similarity": 0.996001, "kl_divergence": 0.006633, "total_variation": 0.042764, ...	[ { "name": "original_speech", "jensen_shannon": 0, "cosine_similarity": 1, "kl_divergence": 0, "total_variation": 0, "n": 800, "accuracy": 0.5675, "speaker_agreement": 0.75 }, { "name": "baseline_orpheus", "jensen_shannon": 0, "cosine_similarity": 1, "kl_divergence...	{ "Mild": 0.0875, "Moderate": 0.25, "Normal": 0.325, "Severe": 0.3375 }	[ "Mild", "Moderate", "Normal", "Severe" ]	./outputs/severity_classifier_mms

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Severity Distribution Comparison

Compares predicted severity distributions of synthetic speech methods against original pathological speech.

Classifier: ./outputs/severity_classifier_mms (accuracy: 97.87%)

Reference: original_speech

Reference distribution: Mild: 15.8%, Moderate: 35.1%, Normal: 41.5%, Severe: 7.6%

Ranking (by Jensen-Shannon distance, lower = closer to original)

Method	JS Dist	Cos Sim	KL Div	TV Dist	N
original_speech	0.0000	1.0000	0.0000	0.0000	800
orpheus_tts	0.0412	0.9960	0.0066	0.0428	759
spark_tts	0.0514	0.9961	0.0106	0.0425	785
sesame_tts	0.0622	0.9908	0.0158	0.0717	759
sesame_knn_vc	0.2309	0.8584	0.1981	0.2313	759
spark_knn_vc	0.2586	0.8239	0.2565	0.2550	785
orpheus_knn_vc	0.2955	0.7785	0.3310	0.3085	759
knn_vc	0.3908	0.7668	0.9039	0.4574	799
baseline_orpheus	0.4319	0.7491	1.3982	0.4900	800

Per-Method Severity Distribution

Method	Mild	Moderate	Normal	Severe
original_speech	15.8%	35.1%	41.5%	7.6%
orpheus_tts	20.0%	34.5%	38.9%	6.6%
spark_tts	11.8%	34.8%	42.8%	10.6%
sesame_tts	13.2%	42.3%	39.7%	4.9%
sesame_knn_vc	15.9%	17.3%	36.2%	30.6%
spark_knn_vc	14.1%	14.0%	38.7%	33.1%
orpheus_knn_vc	7.2%	14.2%	40.1%	38.5%
knn_vc	0.8%	4.4%	69.8%	25.0%
baseline_orpheus	0.2%	1.6%	87.9%	10.2%

original_speech — Per-Speaker: Expected vs Predicted

Speaker	N	Expected	Predicted	Match	Mild	Moderate	Normal	Severe
211	30	Normal	Normal	✓	0%	0%	100%	0%
4014	30	Normal	Normal	✓	3%	0%	97%	0%
F02	200	Moderate	Moderate	✓	31%	68%	0%	0%
F04	70	Mild	Normal	✗	4%	4%	91%	0%
FC02	100	Normal	Normal	✓	0%	1%	99%	0%
M02	70	Severe	Severe	✓	4%	19%	19%	59%
M04	200	Severe	Moderate	✗	28%	63%	0%	9%
MC01	100	Normal	Normal	✓	1%	1%	97%	1%

orpheus_tts — Per-Speaker: Expected vs Predicted

Speaker	N	Expected	Predicted	Match	Mild	Moderate	Normal	Severe
211	30	Unknown	Normal	?	0%	0%	100%	0%
4014	30	Unknown	Normal	?	0%	0%	100%	0%
F02	182	Unknown	Moderate	?	48%	51%	0%	1%
F04	69	Mild	Normal	✗	0%	13%	87%	0%
FC02	98	Normal	Normal	✓	1%	22%	77%	0%
M02	69	Severe	Severe	✓	0%	32%	23%	45%
M04	184	Severe	Moderate	✗	34%	58%	0%	9%
MC01	97	Normal	Normal	✓	1%	11%	87%	1%

spark_tts — Per-Speaker: Expected vs Predicted

Speaker	N	Expected	Predicted	Match	Mild	Moderate	Normal	Severe
211	30	Unknown	Normal	?	0%	0%	100%	0%
4014	30	Unknown	Normal	?	0%	0%	100%	0%
F02	196	Moderate	Moderate	✓	32%	67%	0%	2%
F04	68	Moderate	Normal	✗	1%	3%	96%	0%
FC02	99	Unknown	Normal	?	0%	2%	98%	0%
M02	65	Mild	Severe	✗	18%	9%	26%	46%
M04	198	Severe	Moderate	✗	9%	67%	0%	25%
MC01	99	Unknown	Normal	?	1%	0%	98%	1%

sesame_tts — Per-Speaker: Expected vs Predicted

Speaker	N	Expected	Predicted	Match	Mild	Moderate	Normal	Severe
211	30	Unknown	Normal	?	0%	0%	100%	0%
4014	30	Unknown	Normal	?	0%	0%	100%	0%
F02	182	Unknown	Moderate	?	39%	60%	1%	0%
F04	69	Mild	Normal	✗	6%	26%	68%	0%
FC02	98	Normal	Normal	✓	0%	15%	85%	0%
M02	69	Severe	Severe	✓	1%	39%	17%	42%
M04	184	Severe	Moderate	✗	13%	82%	2%	3%
MC01	97	Normal	Normal	✓	0%	1%	96%	3%

sesame_knn_vc — Per-Speaker: Expected vs Predicted

Speaker	N	Expected	Predicted	Match	Mild	Moderate	Normal	Severe
211	30	Normal	Normal	✓	0%	0%	100%	0%
4014	30	Normal	Normal	✓	7%	0%	93%	0%
F02	182	Moderate	Moderate	✓	37%	52%	1%	11%
F04	69	Mild	Normal	✗	32%	7%	59%	1%
FC02	98	Normal	Normal	✓	5%	12%	83%	0%
M02	69	Severe	Severe	✓	14%	4%	12%	70%
M04	184	Severe	Severe	✓	1%	9%	4%	87%
MC01	97	Normal	Normal	✓	14%	1%	81%	3%

spark_knn_vc — Per-Speaker: Expected vs Predicted

Speaker	N	Expected	Predicted	Match	Mild	Moderate	Normal	Severe
211	30	Normal	Normal	✓	0%	0%	100%	0%
4014	30	Normal	Normal	✓	0%	0%	100%	0%
F02	196	Moderate	Moderate	✓	39%	44%	0%	17%
F04	68	Mild	Normal	✗	18%	3%	78%	1%
FC02	99	Normal	Normal	✓	1%	3%	96%	0%
M02	65	Severe	Severe	✓	18%	0%	14%	68%
M04	198	Severe	Severe	✓	2%	9%	0%	90%
MC01	99	Normal	Normal	✓	7%	2%	88%	3%

orpheus_knn_vc — Per-Speaker: Expected vs Predicted

Speaker	N	Expected	Predicted	Match	Mild	Moderate	Normal	Severe
211	30	Unknown	Normal	?	0%	0%	100%	0%
4014	30	Unknown	Normal	?	3%	0%	97%	0%
F02	182	Moderate	Moderate	✓	23%	44%	0%	33%
F04	69	Moderate	Normal	✗	9%	1%	87%	3%
FC02	98	Unknown	Normal	?	4%	22%	73%	0%
M02	69	Mild	Severe	✗	3%	0%	29%	68%
M04	184	Severe	Severe	✓	0%	2%	0%	98%
MC01	97	Unknown	Normal	?	0%	2%	96%	2%

knn_vc — Per-Speaker: Expected vs Predicted

Speaker	N	Expected	Predicted	Match	Mild	Moderate	Normal	Severe
211	30	Unknown	Normal	?	0%	0%	100%	0%
4014	30	Unknown	Normal	?	0%	0%	100%	0%
F02	200	Unknown	Normal	?	3%	8%	74%	16%
F04	70	Mild	Normal	✗	0%	0%	99%	1%
FC02	100	Normal	Normal	✓	0%	6%	92%	2%
M02	70	Severe	Normal	✗	0%	0%	99%	1%
M04	200	Severe	Severe	✓	0%	6%	12%	82%
MC01	99	Normal	Normal	✓	0%	1%	98%	1%

baseline_orpheus — Per-Speaker: Expected vs Predicted

Speaker	N	Expected	Predicted	Match	Mild	Moderate	Normal	Severe
211	30	Normal	Normal	✓	0%	0%	100%	0%
4014	30	Normal	Normal	✓	0%	0%	100%	0%
F02	200	Moderate	Normal	✗	0%	2%	84%	14%
F04	70	Mild	Normal	✗	3%	0%	89%	9%
FC02	100	Normal	Normal	✓	0%	1%	91%	8%
M02	70	Severe	Normal	✗	0%	1%	93%	6%
M04	200	Severe	Normal	✗	0%	3%	82%	14%
MC01	100	Normal	Normal	✓	0%	0%	92%	8%

Ground-Truth Label Comparison

Compares severity distributions based on dataset labels (from speaker metadata) rather than classifier predictions.

Reference GT distribution: Mild: 8.8%, Moderate: 25.0%, Normal: 32.5%, Severe: 33.8%

GT Ranking (by Jensen-Shannon distance on labels)

Method	GT JS Dist	GT Cos Sim	GT KL Div	GT TV Dist	N	Clf Accuracy	Speaker Agreement
original_speech	0.0000	1.0000	0.0000	0.0000	800	0.5675	0.7500
baseline_orpheus	0.0000	1.0000	0.0000	0.0000	800	0.3538	0.5000
knn_vc	0.0006	1.0000	0.0000	0.0008	799	0.5382	0.6000
spark_tts	0.0029	1.0000	0.0000	0.0037	785	0.5924	0.2500
spark_knn_vc	0.0029	1.0000	0.0000	0.0037	785	0.7159	0.8750
orpheus_tts	0.0110	0.9996	0.0005	0.0144	759	0.4717	0.6000
sesame_tts	0.0110	0.9996	0.0005	0.0144	759	0.5046	0.6000
orpheus_knn_vc	0.0110	0.9996	0.0005	0.0144	759	0.7088	0.5000
sesame_knn_vc	0.0110	0.9996	0.0005	0.0144	759	0.7141	0.8750

GT Per-Method Severity Distribution

Method	Mild	Moderate	Normal	Severe
original_speech	8.8%	25.0%	32.5%	33.8%
baseline_orpheus	8.8%	25.0%	32.5%	33.8%
knn_vc	8.8%	25.0%	32.4%	33.8%
spark_tts	8.7%	25.0%	32.9%	33.5%
spark_knn_vc	8.7%	25.0%	32.9%	33.5%
orpheus_tts	9.1%	24.0%	33.6%	33.3%
sesame_tts	9.1%	24.0%	33.6%	33.3%
orpheus_knn_vc	9.1%	24.0%	33.6%	33.3%
sesame_knn_vc	9.1%	24.0%	33.6%	33.3%

Downloads last month: 15

Number of rows:

Total file size:

51.3 kB