diff --git "a/data/FysicsWorld-Omni/image-centric/Speech_Generation_from_Image.json" "b/data/FysicsWorld-Omni/image-centric/Speech_Generation_from_Image.json"
--- "a/data/FysicsWorld-Omni/image-centric/Speech_Generation_from_Image.json"
+++ "b/data/FysicsWorld-Omni/image-centric/Speech_Generation_from_Image.json"
@@ -1,1150 +1,904 @@
 [
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Sports",
-      "content_fine_category": "Career Highlights",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/97ec6dd4_G_VTkkb34gw_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/d73d2eb1_G_VTkkb34gw.jpg",
+    "image_path": "media_merge/image/1e2ff05ad5b046b986d30399bedd3ae7_id07312.jpg",
+    "audio_path": "media_merge/audio/1e2ff05ad5b046b986d30399bedd3ae7_id07312.wav",
     "index": 1
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Education",
-      "content_fine_category": "Engineering Projects",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/b0d6ac9c_d6b4OmUFt7I_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/197cf005_d6b4OmUFt7I.jpg",
+    "image_path": "media_merge/image/14a70a8dfdea45f681110e030673f6fd_id01989.jpg",
+    "audio_path": "media_merge/audio/14a70a8dfdea45f681110e030673f6fd_id01989.wav",
     "index": 2
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Sports",
-      "content_fine_category": "Career Highlights",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/9c794f85_G_VTkkb34gw_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/cc823d7f_G_VTkkb34gw.jpg",
+    "image_path": "media_merge/image/fdec9779f71a48df8d11e9508d289dc2_id04276.jpg",
+    "audio_path": "media_merge/audio/fdec9779f71a48df8d11e9508d289dc2_id04276.wav",
     "index": 3
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Sports",
-      "content_fine_category": "Career Highlights",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/7f3f9d52_G_VTkkb34gw_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/a66fa4d1_G_VTkkb34gw.jpg",
+    "image_path": "media_merge/image/bb830f83c73a44dfa9bd872ac2881b37_id08552.jpg",
+    "audio_path": "media_merge/audio/bb830f83c73a44dfa9bd872ac2881b37_id08552.wav",
     "index": 4
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Education",
-      "content_fine_category": "Engineering Projects",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/f4e6b2c9_d6b4OmUFt7I_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/ad8c66ef_d6b4OmUFt7I.jpg",
+    "image_path": "media_merge/image/0c64a0dd300348a88df57d63d8ca8c6f_id02057.jpg",
+    "audio_path": "media_merge/audio/0c64a0dd300348a88df57d63d8ca8c6f_id02057.wav",
     "index": 5
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Education",
-      "content_fine_category": "Engineering Projects",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/48b5643b_d6b4OmUFt7I_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/75db242d_d6b4OmUFt7I.jpg",
+    "image_path": "media_merge/image/d3388d2a619d454da49d7bd57e3a8fa8_id03178.jpg",
+    "audio_path": "media_merge/audio/d3388d2a619d454da49d7bd57e3a8fa8_id03178.wav",
     "index": 6
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Education",
-      "content_fine_category": "Engineering Projects",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/a5f99196_d6b4OmUFt7I_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/ee332948_d6b4OmUFt7I.jpg",
+    "image_path": "media_merge/image/4c2f3200fc2d43c8b1c123bff54e1977_id09017.jpg",
+    "audio_path": "media_merge/audio/4c2f3200fc2d43c8b1c123bff54e1977_id09017.wav",
     "index": 7
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Sports",
-      "content_fine_category": "Documentary Profiles",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/6b07feeb_jj5j3jksbUM_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/ccdec681_jj5j3jksbUM.jpg",
+    "image_path": "media_merge/image/978c3622f7104b498b65613baa30a28e_id04253.jpg",
+    "audio_path": "media_merge/audio/978c3622f7104b498b65613baa30a28e_id04253.wav",
     "index": 8
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Sports",
-      "content_fine_category": "Documentary Profiles",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/6ae1ce99_jj5j3jksbUM_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/86307fdf_jj5j3jksbUM.jpg",
+    "image_path": "media_merge/image/3d1cbb5120434af7a94611a19c676fea_id04030.jpg",
+    "audio_path": "media_merge/audio/3d1cbb5120434af7a94611a19c676fea_id04030.wav",
     "index": 9
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Lifestyle",
-      "content_fine_category": "Travel Tips",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/67238901_6vKvjzuICxg_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/e523f9bc_6vKvjzuICxg.jpg",
+    "image_path": "media_merge/image/db776900341b4f5f85da389ac3744ec6_id03127.jpg",
+    "audio_path": "media_merge/audio/db776900341b4f5f85da389ac3744ec6_id03127.wav",
     "index": 10
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Sports",
-      "content_fine_category": "Documentary Profiles",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/c36368e1_jj5j3jksbUM_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/f17e37da_jj5j3jksbUM.jpg",
+    "image_path": "media_merge/image/439ed1b3f7284fb2aa3cc5da8062f533_id06913.jpg",
+    "audio_path": "media_merge/audio/439ed1b3f7284fb2aa3cc5da8062f533_id06913.wav",
     "index": 11
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Lifestyle",
-      "content_fine_category": "Travel Tips",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/fd286127_6vKvjzuICxg_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/444d84bf_6vKvjzuICxg.jpg",
+    "image_path": "media_merge/image/628a88ba3bf94c179d802ec64d5e84a9_id03382.jpg",
+    "audio_path": "media_merge/audio/628a88ba3bf94c179d802ec64d5e84a9_id03382.wav",
     "index": 12
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "News & Politics",
-      "content_fine_category": "World News",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/8536ee5c_2u3tic9llpw_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/175165a5_2u3tic9llpw.jpg",
+    "image_path": "media_merge/image/c2190330195b4cfba84a94f7e0a74fa4_id07868.jpg",
+    "audio_path": "media_merge/audio/c2190330195b4cfba84a94f7e0a74fa4_id07868.wav",
     "index": 13
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "News & Politics",
-      "content_fine_category": "World News",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/dee1f1ca_2u3tic9llpw_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/fffba1d7_2u3tic9llpw.jpg",
+    "image_path": "media_merge/image/7abbf4d08b22436194f0dec8977d2112_id01892.jpg",
+    "audio_path": "media_merge/audio/7abbf4d08b22436194f0dec8977d2112_id01892.wav",
     "index": 14
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "News & Politics",
-      "content_fine_category": "World News",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/2ed8d85f_2u3tic9llpw_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/a27007e8_2u3tic9llpw.jpg",
+    "image_path": "media_merge/image/baed92a410e54be7842bf5304106b8ec_id00419.jpg",
+    "audio_path": "media_merge/audio/baed92a410e54be7842bf5304106b8ec_id00419.wav",
     "index": 15
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "News & Politics",
-      "content_fine_category": "World News",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/9bc35ba8_2u3tic9llpw_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/681f4222_2u3tic9llpw.jpg",
+    "image_path": "media_merge/image/7d842aefe722479e9ab3ef87d3073fd2_id05999.jpg",
+    "audio_path": "media_merge/audio/7d842aefe722479e9ab3ef87d3073fd2_id05999.wav",
     "index": 16
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Education",
-      "content_fine_category": "Historical Analysis",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/187ddd45_P6-I5fsCA9w_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/146d9622_P6-I5fsCA9w.jpg",
+    "image_path": "media_merge/image/459be473f2d247ce9bf87ea09560a053_id03178.jpg",
+    "audio_path": "media_merge/audio/459be473f2d247ce9bf87ea09560a053_id03178.wav",
     "index": 17
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Education",
-      "content_fine_category": "Marketing Strategies",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/66ef76c9_gleNOAoDG4U_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/b5a3617e_gleNOAoDG4U.jpg",
+    "image_path": "media_merge/image/0cc05eb924a54426b871fc96043fc34b_id03347.jpg",
+    "audio_path": "media_merge/audio/0cc05eb924a54426b871fc96043fc34b_id03347.wav",
     "index": 18
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Education",
-      "content_fine_category": "Historical Analysis",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/c09fe630_06b2PxvBoNU_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/e4b06935_06b2PxvBoNU.jpg",
+    "image_path": "media_merge/image/f71988d5a15942779359a34ca6447102_id07312.jpg",
+    "audio_path": "media_merge/audio/f71988d5a15942779359a34ca6447102_id07312.wav",
     "index": 19
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "News & Politics",
-      "content_fine_category": "Political News",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/f3ff0171_zI4QBIVHeIY_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/3a3beabf_zI4QBIVHeIY.jpg",
+    "image_path": "media_merge/image/195b71860a1940faa520046a03db22f3_id00817.jpg",
+    "audio_path": "media_merge/audio/195b71860a1940faa520046a03db22f3_id00817.wav",
     "index": 20
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Education",
-      "content_fine_category": "Marketing Strategies",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/e9b96560_gleNOAoDG4U_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/1b730eb1_gleNOAoDG4U.jpg",
+    "image_path": "media_merge/image/286126c4c058482789a2ad3f81f8e5a1_id03382.jpg",
+    "audio_path": "media_merge/audio/286126c4c058482789a2ad3f81f8e5a1_id03382.wav",
     "index": 21
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "News & Politics",
-      "content_fine_category": "Breaking News",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/8b810806_DTA5AbpHwEQ_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/60eea5dd_DTA5AbpHwEQ.jpg",
+    "image_path": "media_merge/image/778fce059549467eb37363e4fe21474f_id01593.jpg",
+    "audio_path": "media_merge/audio/778fce059549467eb37363e4fe21474f_id01593.wav",
     "index": 22
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "News & Politics",
-      "content_fine_category": "World News",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/f2d1d720_i9x_JohBBe4_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/1a8087b7_i9x_JohBBe4.jpg",
+    "image_path": "media_merge/image/0369064952034d09a7c20dd689fd524e_id06484.jpg",
+    "audio_path": "media_merge/audio/0369064952034d09a7c20dd689fd524e_id06484.wav",
     "index": 23
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "News & Politics",
-      "content_fine_category": "Social Commentary",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/1dc63799_ZtE20CoGkus_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/6b6713bb_ZtE20CoGkus.jpg",
+    "image_path": "media_merge/image/9761e9cd323f4ba09968cc571ba146b4_id00154.jpg",
+    "audio_path": "media_merge/audio/9761e9cd323f4ba09968cc571ba146b4_id00154.wav",
     "index": 24
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "News & Politics",
-      "content_fine_category": "Social Commentary",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/52806224_ZtE20CoGkus_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/24bebb45_ZtE20CoGkus.jpg",
+    "image_path": "media_merge/image/3a5086dabe7740e6976a33ee3cee27ee_id07414.jpg",
+    "audio_path": "media_merge/audio/3a5086dabe7740e6976a33ee3cee27ee_id07414.wav",
     "index": 25
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Education",
-      "content_fine_category": "Marketing Strategies",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/1c5756e0_jqJNYUb22Xk_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/843569e7_jqJNYUb22Xk.jpg",
+    "image_path": "media_merge/image/bdbe0f67aee64d49aada6b42509d0079_id07620.jpg",
+    "audio_path": "media_merge/audio/bdbe0f67aee64d49aada6b42509d0079_id07620.wav",
     "index": 26
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "News & Politics",
-      "content_fine_category": "Social Commentary",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/0ea15fd2_ZtE20CoGkus_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/70ed33be_ZtE20CoGkus.jpg",
+    "image_path": "media_merge/image/d41ace3040de440cbc7699afd71b9407_id01041.jpg",
+    "audio_path": "media_merge/audio/d41ace3040de440cbc7699afd71b9407_id01041.wav",
     "index": 27
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "Science & Technology",
-      "content_fine_category": "Climate Change",
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/0a8c96d4_b7LiW66cSM4_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/6cc50672_b7LiW66cSM4.jpg",
+    "image_path": "media_merge/image/35300db4491443dbbd1cf09f8c6749bd_id06104.jpg",
+    "audio_path": "media_merge/audio/35300db4491443dbbd1cf09f8c6749bd_id06104.wav",
     "index": 28
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": null,
-      "content_fine_category": null,
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/cbc98947_2q4_wFFFG_U_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/deb0913e_2q4_wFFFG_U.jpg",
+    "image_path": "media_merge/image/6d1d40e3469e4cf7a4429996e76237b9_id08374.jpg",
+    "audio_path": "media_merge/audio/6d1d40e3469e4cf7a4429996e76237b9_id08374.wav",
     "index": 29
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": null,
-      "content_fine_category": null,
-      "video_category": "News & Politics",
-      "video_duration": "30s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/468ee181_2q4_wFFFG_U_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/3dbf5cbe_2q4_wFFFG_U.jpg",
+    "image_path": "media_merge/image/8f271f7adbec45939518e3715e86ecc4_id01567.jpg",
+    "audio_path": "media_merge/audio/8f271f7adbec45939518e3715e86ecc4_id01567.wav",
     "index": 30
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/f14a3085_08km9Yqbt-A_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/6f05c7f4_08km9Yqbt-A.jpg",
+    "image_path": "media_merge/image/298f090ad0334264b176220c00b2a2f7_id05202.jpg",
+    "audio_path": "media_merge/audio/298f090ad0334264b176220c00b2a2f7_id05202.wav",
     "index": 31
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/bf074378_08km9Yqbt-A_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/6de3d129_08km9Yqbt-A.jpg",
+    "image_path": "media_merge/image/8bd97bf446bb4e8ebe366695e7ce10c0_id01000.jpg",
+    "audio_path": "media_merge/audio/8bd97bf446bb4e8ebe366695e7ce10c0_id01000.wav",
     "index": 32
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/02c36869_1NH5dJ9VRvU_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/ec2a69a2_1NH5dJ9VRvU.jpg",
+    "image_path": "media_merge/image/287d11adb1894120ba285a6e5a122478_id02725.jpg",
+    "audio_path": "media_merge/audio/287d11adb1894120ba285a6e5a122478_id02725.wav",
     "index": 33
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/4ab5309e_08km9Yqbt-A_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/f1d1258d_08km9Yqbt-A.jpg",
+    "image_path": "media_merge/image/75eb7879e4654f6082671ed8fb278f8b_id03969.jpg",
+    "audio_path": "media_merge/audio/75eb7879e4654f6082671ed8fb278f8b_id03969.wav",
     "index": 34
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/f7472433_1NH5dJ9VRvU_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/447f2678_1NH5dJ9VRvU.jpg",
+    "image_path": "media_merge/image/f60d773287b842bda932c47701b33174_id08696.jpg",
+    "audio_path": "media_merge/audio/f60d773287b842bda932c47701b33174_id08696.wav",
     "index": 35
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/eb941a8f_20gUIdXGuAs_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/8a5931c5_20gUIdXGuAs.jpg",
+    "image_path": "media_merge/image/be0b3b65c6dc4adc83681436832517df_id05594.jpg",
+    "audio_path": "media_merge/audio/be0b3b65c6dc4adc83681436832517df_id05594.wav",
     "index": 36
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/14e3d6f8_20gUIdXGuAs_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/526c0b5c_20gUIdXGuAs.jpg",
+    "image_path": "media_merge/image/213198562393461082c3a86877686230_id04862.jpg",
+    "audio_path": "media_merge/audio/213198562393461082c3a86877686230_id04862.wav",
     "index": 37
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/62b0c467_1NH5dJ9VRvU_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/cb1b0e1e_1NH5dJ9VRvU.jpg",
+    "image_path": "media_merge/image/b658c85fda484b03b16435a482fe0691_id03041.jpg",
+    "audio_path": "media_merge/audio/b658c85fda484b03b16435a482fe0691_id03041.wav",
     "index": 38
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/02f9f6b8_40BlVzjxu-I_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/51432dcd_40BlVzjxu-I.jpg",
+    "image_path": "media_merge/image/ec8c9a1f7c5c412fa41e2f8584ee086c_id03839.jpg",
+    "audio_path": "media_merge/audio/ec8c9a1f7c5c412fa41e2f8584ee086c_id03839.wav",
     "index": 39
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/9f4e5938_40BlVzjxu-I_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/bd147ca6_40BlVzjxu-I.jpg",
+    "image_path": "media_merge/image/dea777986ee244979a5c3a8655893c6d_id05055.jpg",
+    "audio_path": "media_merge/audio/dea777986ee244979a5c3a8655893c6d_id05055.wav",
     "index": 40
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/61d46b34_20gUIdXGuAs_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/c54e079e_20gUIdXGuAs.jpg",
+    "image_path": "media_merge/image/9f5116191ec745f799893bb316e6381c_id03981.jpg",
+    "audio_path": "media_merge/audio/9f5116191ec745f799893bb316e6381c_id03981.wav",
     "index": 41
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/e656bd4d_3fQxvz4ENDg_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/8fcb7278_3fQxvz4ENDg.jpg",
+    "image_path": "media_merge/image/3df5a8a22e434a1fbee7707fe790fd47_id00061.jpg",
+    "audio_path": "media_merge/audio/3df5a8a22e434a1fbee7707fe790fd47_id00061.wav",
     "index": 42
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/a5f2ed4d_40BlVzjxu-I_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/2d34ccf7_40BlVzjxu-I.jpg",
+    "image_path": "media_merge/image/6a1ac9e88fe04dc4907724e2a70da6b4_id05124.jpg",
+    "audio_path": "media_merge/audio/6a1ac9e88fe04dc4907724e2a70da6b4_id05124.wav",
     "index": 43
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/1eb1b1fc_3fQxvz4ENDg_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/03b3be14_3fQxvz4ENDg.jpg",
+    "image_path": "media_merge/image/4b62524efe5d41d0a9188ac7d9f6c0d2_id02576.jpg",
+    "audio_path": "media_merge/audio/4b62524efe5d41d0a9188ac7d9f6c0d2_id02576.wav",
     "index": 44
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/02bd1373_3uV8XZcIBbk_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/263c71c5_3uV8XZcIBbk.jpg",
+    "image_path": "media_merge/image/33a762741fe940a193863278bee398f1_id00081.jpg",
+    "audio_path": "media_merge/audio/33a762741fe940a193863278bee398f1_id00081.wav",
     "index": 45
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/a28667d2_3fQxvz4ENDg_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/e09ec463_3fQxvz4ENDg.jpg",
+    "image_path": "media_merge/image/6ca28de8fa5e42f8b57e92569bd89b9c_id04627.jpg",
+    "audio_path": "media_merge/audio/6ca28de8fa5e42f8b57e92569bd89b9c_id04627.wav",
     "index": 46
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/91d085df_54H8ppxnp8I_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/f8e2dadf_54H8ppxnp8I.jpg",
+    "image_path": "media_merge/image/f3c0f4bc40d64a939847954b3d5a0655_id05015.jpg",
+    "audio_path": "media_merge/audio/f3c0f4bc40d64a939847954b3d5a0655_id05015.wav",
     "index": 47
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/c10964a5_3fQxvz4ENDg_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/a476f2c9_3fQxvz4ENDg.jpg",
+    "image_path": "media_merge/image/2882daf534334e66932bb06ab299d8c2_id03347.jpg",
+    "audio_path": "media_merge/audio/2882daf534334e66932bb06ab299d8c2_id03347.wav",
     "index": 48
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/a01199b1_8TNPeimqOO0_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/feda04e8_8TNPeimqOO0.jpg",
+    "image_path": "media_merge/image/0b6a372818934e38922117129da70c5d_id08456.jpg",
+    "audio_path": "media_merge/audio/0b6a372818934e38922117129da70c5d_id08456.wav",
     "index": 49
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/9388db8b_8TNPeimqOO0_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/64f13ccf_8TNPeimqOO0.jpg",
+    "image_path": "media_merge/image/210f5f96f7a6435ab0b4e0a92815169e_id07621.jpg",
+    "audio_path": "media_merge/audio/210f5f96f7a6435ab0b4e0a92815169e_id07621.wav",
     "index": 50
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/dc08daa9_8TNPeimqOO0_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/ac8a8efb_8TNPeimqOO0.jpg",
+    "image_path": "media_merge/image/d86deb609b6b4049a40f76571ec68484_id07426.jpg",
+    "audio_path": "media_merge/audio/d86deb609b6b4049a40f76571ec68484_id07426.wav",
     "index": 51
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/7460e8d0_990ci2H9BBc_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/8da76641_990ci2H9BBc.jpg",
+    "image_path": "media_merge/image/5bdd419f2f154690a2c4f7119fe9a227_id01437.jpg",
+    "audio_path": "media_merge/audio/5bdd419f2f154690a2c4f7119fe9a227_id01437.wav",
     "index": 52
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/2dd2a7a7_9QXXL96y3A0_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/b01f024f_9QXXL96y3A0.jpg",
+    "image_path": "media_merge/image/731769f5b85f43df96ed45ab2e76c314_id07961.jpg",
+    "audio_path": "media_merge/audio/731769f5b85f43df96ed45ab2e76c314_id07961.wav",
     "index": 53
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/6e569510_8TNPeimqOO0_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/a74686c5_8TNPeimqOO0.jpg",
+    "image_path": "media_merge/image/aa7fac75ad3f48b5b3218ed740fac81b_id06104.jpg",
+    "audio_path": "media_merge/audio/aa7fac75ad3f48b5b3218ed740fac81b_id06104.wav",
     "index": 54
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/21d39268_9Y-YJEtxHeo_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/0490b965_9Y-YJEtxHeo.jpg",
+    "image_path": "media_merge/image/69a49a8154ae436e8874f5a46534d7bb_id05176.jpg",
+    "audio_path": "media_merge/audio/69a49a8154ae436e8874f5a46534d7bb_id05176.wav",
     "index": 55
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/836f05b6_9Y-YJEtxHeo_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/79725b2b_9Y-YJEtxHeo.jpg",
+    "image_path": "media_merge/image/452d7851ab474980863b443a2003ed13_id07802.jpg",
+    "audio_path": "media_merge/audio/452d7851ab474980863b443a2003ed13_id07802.wav",
     "index": 56
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/b34c55cf_9Y-YJEtxHeo_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/e3e512a4_9Y-YJEtxHeo.jpg",
+    "image_path": "media_merge/image/7a445fcd224345709654e545692575ae_id04366.jpg",
+    "audio_path": "media_merge/audio/7a445fcd224345709654e545692575ae_id04366.wav",
     "index": 57
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/d27f09ba_AFFZjzuAHMg_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/6654acd7_AFFZjzuAHMg.jpg",
+    "image_path": "media_merge/image/f0befbfeeab443d8b91c6e3f79634087_id08701.jpg",
+    "audio_path": "media_merge/audio/f0befbfeeab443d8b91c6e3f79634087_id08701.wav",
     "index": 58
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/8493fb53_AFFZjzuAHMg_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/9c80c881_AFFZjzuAHMg.jpg",
+    "image_path": "media_merge/image/b4fc8c6087174550a753766bf10380db_id06209.jpg",
+    "audio_path": "media_merge/audio/b4fc8c6087174550a753766bf10380db_id06209.wav",
     "index": 59
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/2f7b01a1_CKYw_RG7i7w_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/1ef16416_CKYw_RG7i7w.jpg",
+    "image_path": "media_merge/image/63ea503b9bc04ed785122a1360b6a0cc_id01460.jpg",
+    "audio_path": "media_merge/audio/63ea503b9bc04ed785122a1360b6a0cc_id01460.wav",
     "index": 60
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/60a2f879_CKYw_RG7i7w_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/dbadaf68_CKYw_RG7i7w.jpg",
+    "image_path": "media_merge/image/bf597fa1e39a400c9a660cc7ddc8a33f_id03980.jpg",
+    "audio_path": "media_merge/audio/bf597fa1e39a400c9a660cc7ddc8a33f_id03980.wav",
     "index": 61
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/b19ba8be_E-YnmCB6yjo_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/b3f64faf_E-YnmCB6yjo.jpg",
+    "image_path": "media_merge/image/9c1994b01edd4b92b5939f431f53b5df_id01509.jpg",
+    "audio_path": "media_merge/audio/9c1994b01edd4b92b5939f431f53b5df_id01509.wav",
     "index": 62
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/0cbb811e_E-YnmCB6yjo_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/1a08a299_E-YnmCB6yjo.jpg",
+    "image_path": "media_merge/image/31164d88ef6e434b8211961de6aae6cc_id07802.jpg",
+    "audio_path": "media_merge/audio/31164d88ef6e434b8211961de6aae6cc_id07802.wav",
     "index": 63
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/87b319f7_FkZn5u3MIiY_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/32922eb8_FkZn5u3MIiY.jpg",
+    "image_path": "media_merge/image/3f9d583af1a54858b6d3abd1e89891a2_id01541.jpg",
+    "audio_path": "media_merge/audio/3f9d583af1a54858b6d3abd1e89891a2_id01541.wav",
     "index": 64
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/fb042dc3_E-YnmCB6yjo_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/41716b95_E-YnmCB6yjo.jpg",
+    "image_path": "media_merge/image/e705553afa694730aea77465d0615f5b_id08456.jpg",
+    "audio_path": "media_merge/audio/e705553afa694730aea77465d0615f5b_id08456.wav",
     "index": 65
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/124bcfbe_FcP0mzWFCQU_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/031168c3_FcP0mzWFCQU.jpg",
+    "image_path": "media_merge/image/923e1da6e2fe4ac1a7327d21ffa28d4e_id04006.jpg",
+    "audio_path": "media_merge/audio/923e1da6e2fe4ac1a7327d21ffa28d4e_id04006.wav",
     "index": 66
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/f4db15bf_FkZn5u3MIiY_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/003f8119_FkZn5u3MIiY.jpg",
+    "image_path": "media_merge/image/eb1a6655e8ec40eeb254dd7d25fd9511_id00812.jpg",
+    "audio_path": "media_merge/audio/eb1a6655e8ec40eeb254dd7d25fd9511_id00812.wav",
     "index": 67
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/dec8eb80_FkZn5u3MIiY_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/833a9085_FkZn5u3MIiY.jpg",
+    "image_path": "media_merge/image/7ac170f4e8d841619f7998b742c59430_id04570.jpg",
+    "audio_path": "media_merge/audio/7ac170f4e8d841619f7998b742c59430_id04570.wav",
     "index": 68
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/e81da04a_HK13nVxBYxI_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/9c1daa1e_HK13nVxBYxI.jpg",
+    "image_path": "media_merge/image/fdec6e118d12474caf0fdccfefcf266e_id04950.jpg",
+    "audio_path": "media_merge/audio/fdec6e118d12474caf0fdccfefcf266e_id04950.wav",
     "index": 69
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/895ed456_N1cdUjctpG8_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/4c157c29_N1cdUjctpG8.jpg",
+    "image_path": "media_merge/image/c77f3cfbe1734f4b9bbac40b06f26b97_id02685.jpg",
+    "audio_path": "media_merge/audio/c77f3cfbe1734f4b9bbac40b06f26b97_id02685.wav",
     "index": 70
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/c0b0e549_N1cdUjctpG8_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/9d8c2a87_N1cdUjctpG8.jpg",
+    "image_path": "media_merge/image/8f072c4b968f4fcdb59d41c9fdef24b2_id08696.jpg",
+    "audio_path": "media_merge/audio/8f072c4b968f4fcdb59d41c9fdef24b2_id08696.wav",
     "index": 71
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/924bcd7a_NEFG7YcIDcI_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/bac6a8e3_NEFG7YcIDcI.jpg",
+    "image_path": "media_merge/image/301cf13f7b77441592ba2e835e409f1c_id00866.jpg",
+    "audio_path": "media_merge/audio/301cf13f7b77441592ba2e835e409f1c_id00866.wav",
     "index": 72
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/cfc5b2b1_K3ksKkCOgTw_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/dfcc3bd1_K3ksKkCOgTw.jpg",
+    "image_path": "media_merge/image/557344e04b9e4635b14ff9d5fa703242_id02445.jpg",
+    "audio_path": "media_merge/audio/557344e04b9e4635b14ff9d5fa703242_id02445.wav",
     "index": 73
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/323fe88c_NEFG7YcIDcI_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/a459e994_NEFG7YcIDcI.jpg",
+    "image_path": "media_merge/image/baf0ef8abd064f1898ee3854554c3da3_id09017.jpg",
+    "audio_path": "media_merge/audio/baf0ef8abd064f1898ee3854554c3da3_id09017.wav",
     "index": 74
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/1007bb02_QTA8j5wSTx4_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/11dd86d0_QTA8j5wSTx4.jpg",
+    "image_path": "media_merge/image/5bc0b1700efe4e22a6eaa568831fb40e_id05999.jpg",
+    "audio_path": "media_merge/audio/5bc0b1700efe4e22a6eaa568831fb40e_id05999.wav",
     "index": 75
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/996e5b83_RBTiTcHm_ac_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/84341535_RBTiTcHm_ac.jpg",
+    "image_path": "media_merge/image/6962e3be4a984ce0bd2ec2a8c91aeeac_id04030.jpg",
+    "audio_path": "media_merge/audio/6962e3be4a984ce0bd2ec2a8c91aeeac_id04030.wav",
     "index": 76
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/a62d430c_RBTiTcHm_ac_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/9522f63a_RBTiTcHm_ac.jpg",
+    "image_path": "media_merge/image/2ce49529505e478cbcd448cd863f25b1_id05459.jpg",
+    "audio_path": "media_merge/audio/2ce49529505e478cbcd448cd863f25b1_id05459.wav",
     "index": 77
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/d6eaa023_UyaKyS2wTQk_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/4aaf850e_UyaKyS2wTQk.jpg",
+    "image_path": "media_merge/image/e3c5c17e95b14dbb8270d4e201a78e84_id01822.jpg",
+    "audio_path": "media_merge/audio/e3c5c17e95b14dbb8270d4e201a78e84_id01822.wav",
     "index": 78
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/3d18edac_UZvydHZKyww_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/ae6a40e8_UZvydHZKyww.jpg",
+    "image_path": "media_merge/image/28b27542193c4be7849cc98168dfd185_id03978.jpg",
+    "audio_path": "media_merge/audio/28b27542193c4be7849cc98168dfd185_id03978.wav",
     "index": 79
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/60e8eefd_UyaKyS2wTQk_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/f98319ce_UyaKyS2wTQk.jpg",
+    "image_path": "media_merge/image/1235553cdc324fbe9f428085c3347172_id05850.jpg",
+    "audio_path": "media_merge/audio/1235553cdc324fbe9f428085c3347172_id05850.wav",
     "index": 80
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/bdbb6943_VQKpMmBDtZo_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/bddf90f4_VQKpMmBDtZo.jpg",
+    "image_path": "media_merge/image/1441683b694d4cc19fb3495812645a5d_id02086.jpg",
+    "audio_path": "media_merge/audio/1441683b694d4cc19fb3495812645a5d_id02086.wav",
     "index": 81
   },
   {
-    "prompt": "Analyze the character shown in the image and generate a line of spoken dialogue that matches their identity, personality, age range, gender expression, emotional tone, and likely speaking style. Ensure the line of dialogue fits the character's appearance and possible role suggested by the image.",
+    "prompt": "Given an image of a person, generate a natural human voice that matches the person's gender, age, and overall appearance.\nThe timbre, pitch, and speaking style should be consistent with the character implied by the image.\nSpeak a short, neutral English utterance (2-3 sentences).\nOutput only an audio file (e.g., WAV or MP3), with no text.",
     "task": "Speech_Generation",
     "other_info": {
-      "content_parent_category": "No content_parent_category provided",
-      "content_fine_category": "No content_fine_category provided",
-      "video_category": "News & Politics",
-      "video_duration": "60s"
+      "content_parent_category": "Speech",
+      "content_fine_category": "daily life"
     },
-    "video_path": "media_merge/video/Speech_Generation_from_Image/b2d92114_RBTiTcHm_ac_video.mp4",
-    "audio_path": "",
-    "image_path": "media_merge/image/Speech_Generation_from_Image/2246817d_RBTiTcHm_ac.jpg",
+    "image_path": "media_merge/image/aeb55c1d58e540259b9cab0eed1ab29c_id07396.jpg",
+    "audio_path": "media_merge/audio/aeb55c1d58e540259b9cab0eed1ab29c_id07396.wav",
     "index": 82
   }
 ]
\ No newline at end of file