Spaces:
Sleeping
Sleeping
Spark Chou
commited on
Commit
·
a403a7e
1
Parent(s):
585029d
new
Browse files
app.py
CHANGED
|
@@ -48,7 +48,7 @@ print(sample1_audio_path)
|
|
| 48 |
DIMENSIONS_DATA = [
|
| 49 |
{
|
| 50 |
"title": "Semantic and Pragmatic Features",
|
| 51 |
-
"audio":
|
| 52 |
"sub_dims": [
|
| 53 |
"Memory Consistency: Human memory in short contexts is usually consistent and self-correcting (e.g., by asking questions); machines may show inconsistent context memory and fail to notice or correct errors (e.g., forgetting key information and persisting in wrong answers).",
|
| 54 |
"Logical Coherence: Human logic is naturally coherent and allows reasonable leaps; machine logic is abrupt or self-contradictory (e.g., sudden topic shifts without transitions).",
|
|
@@ -62,7 +62,7 @@ DIMENSIONS_DATA = [
|
|
| 62 |
},
|
| 63 |
{
|
| 64 |
"title": "Non-Physiological Paralinguistic Features",
|
| 65 |
-
"audio":
|
| 66 |
"sub_dims": [
|
| 67 |
"Rhythm: Human speech rate varies with meaning, occasionally hesitating or pausing; machine rhythm is uniform, with little or mechanical pauses.",
|
| 68 |
"Intonation: Humans naturally raise or lower pitch to express questions, surprise, or emphasis; machine intonation is monotonous or overly patterned, mismatching the context.",
|
|
@@ -73,7 +73,7 @@ DIMENSIONS_DATA = [
|
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"title": "Physiological Paralinguistic Features",
|
| 76 |
-
"audio":
|
| 77 |
"sub_dims": [
|
| 78 |
"Micro-physiological Noise: Human speech includes unconscious physiological sounds like breathing, saliva, or bubbling, naturally woven into rhythm; machine speech is overly clean or adds unnatural noises.",
|
| 79 |
"Pronunciation Instability: Human pronunciation includes irregularities (e.g., linking, tremors, slurring, nasal sounds); machine pronunciation is overly standard and uniform, lacking personality.",
|
|
@@ -83,7 +83,7 @@ DIMENSIONS_DATA = [
|
|
| 83 |
},
|
| 84 |
{
|
| 85 |
"title": "Mechanical Persona",
|
| 86 |
-
"audio":
|
| 87 |
"sub_dims": [
|
| 88 |
"Sycophancy: Humans assess context to agree or disagree, sometimes offering differing opinions; machines excessively agree, thank, or apologize, over-validating the other party and lacking authentic interaction.",
|
| 89 |
"Formal Expression: Human speech is flexible; machine responses are formally structured, overly written, and use vague wording."
|
|
@@ -92,7 +92,7 @@ DIMENSIONS_DATA = [
|
|
| 92 |
},
|
| 93 |
{
|
| 94 |
"title": "Emotional Expression",
|
| 95 |
-
"audio":
|
| 96 |
"sub_dims": [
|
| 97 |
"Semantic Level: Humans show appropriate emotional responses to contexts like sadness or joy; machines are emotionally flat, or use emotional words vaguely and out of context.",
|
| 98 |
"Acoustic Level: Human pitch, volume, and rhythm change dynamically with emotion; machine emotional tone is formulaic or mismatched with the context."
|
|
|
|
| 48 |
DIMENSIONS_DATA = [
|
| 49 |
{
|
| 50 |
"title": "Semantic and Pragmatic Features",
|
| 51 |
+
"audio": sample1_audio_path,
|
| 52 |
"sub_dims": [
|
| 53 |
"Memory Consistency: Human memory in short contexts is usually consistent and self-correcting (e.g., by asking questions); machines may show inconsistent context memory and fail to notice or correct errors (e.g., forgetting key information and persisting in wrong answers).",
|
| 54 |
"Logical Coherence: Human logic is naturally coherent and allows reasonable leaps; machine logic is abrupt or self-contradictory (e.g., sudden topic shifts without transitions).",
|
|
|
|
| 62 |
},
|
| 63 |
{
|
| 64 |
"title": "Non-Physiological Paralinguistic Features",
|
| 65 |
+
"audio": sample1_audio_path,
|
| 66 |
"sub_dims": [
|
| 67 |
"Rhythm: Human speech rate varies with meaning, occasionally hesitating or pausing; machine rhythm is uniform, with little or mechanical pauses.",
|
| 68 |
"Intonation: Humans naturally raise or lower pitch to express questions, surprise, or emphasis; machine intonation is monotonous or overly patterned, mismatching the context.",
|
|
|
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"title": "Physiological Paralinguistic Features",
|
| 76 |
+
"audio": sample1_audio_path,
|
| 77 |
"sub_dims": [
|
| 78 |
"Micro-physiological Noise: Human speech includes unconscious physiological sounds like breathing, saliva, or bubbling, naturally woven into rhythm; machine speech is overly clean or adds unnatural noises.",
|
| 79 |
"Pronunciation Instability: Human pronunciation includes irregularities (e.g., linking, tremors, slurring, nasal sounds); machine pronunciation is overly standard and uniform, lacking personality.",
|
|
|
|
| 83 |
},
|
| 84 |
{
|
| 85 |
"title": "Mechanical Persona",
|
| 86 |
+
"audio": sample1_audio_path,
|
| 87 |
"sub_dims": [
|
| 88 |
"Sycophancy: Humans assess context to agree or disagree, sometimes offering differing opinions; machines excessively agree, thank, or apologize, over-validating the other party and lacking authentic interaction.",
|
| 89 |
"Formal Expression: Human speech is flexible; machine responses are formally structured, overly written, and use vague wording."
|
|
|
|
| 92 |
},
|
| 93 |
{
|
| 94 |
"title": "Emotional Expression",
|
| 95 |
+
"audio": sample1_audio_path,
|
| 96 |
"sub_dims": [
|
| 97 |
"Semantic Level: Humans show appropriate emotional responses to contexts like sadness or joy; machines are emotionally flat, or use emotional words vaguely and out of context.",
|
| 98 |
"Acoustic Level: Human pitch, volume, and rhythm change dynamically with emotion; machine emotional tone is formulaic or mismatched with the context."
|