File size: 9,244 Bytes
966f861
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0572d9a
966f861
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0572d9a
966f861
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
import numpy as np
import subprocess
import cv2

# with subprocess and an extra argument 'scene' and a 'resized image saved as png' we can call the server

# yt-dlp is instaled in .d4
# Download Part of Video
# yt-dlp https://www.youtube.com/watch?v=UZ9uyQI3pF0 --downloader ffmpeg --downloader-args "ffmpeg_i:-ss 997 -to 2512"
# ffmpeg -i Sandra\ Kotevska\,\ Painting\ Rose\ bush\,\ mixed\ media\,\ 2017.\ \[NMzC_036MtE\].mkv -f mp3 -ar 22050 -vn out44.wav -ac 1

# https://superuser.com/questions/583393/how-to-extract-subtitle-from-video-using-ffmpeg

def _shift(x):
    n = x.shape[0]
    i = np.random.randint(.24 * n, .74 * n)
    return np.roll(x, i)

#___________________________________________________________________________________________________
#   VIDEO FROM IMAGE with CAPTIONS
#
# UPLOAD to: Simaviro: Documents General WORK PACKAGES WP1 ContentRepository ANBPR_ROMANIA TTSvideos
# __________________________________________________________________________________________________

# TO DONLOAD SRT for youtub
# yt-dlp --write-sub --sub-lang en --convert-subs "srt" https://www.youtube.com/watch?v=F1Ib7TAu7eg&list=PL4x2B6LSwFewdDvRnUTpBM7jkmpwouhPv&index=2

# _voice = 'en_US/vctk_low#p330'
# _voice = 'en_US/cmu-arctic_low#lnh' #en_US/vctk_low#p249'  # 'en_US/vctk_low#p282'
# _voice = ''en_US/vctk_low#p351''
# _voice = 'en_US/vctk_low#p351'  # avoid 318 it does the ghhhhhh
# _voice = 'en_US/m-ailabs_low#judy_bieber'  # Nice voice for ('Arta culinara romaneasca - Groza Irina [phIF0NxgwlQ].mkv' 'Arta culinara romaneasca - Groza Irina [phIF0NxgwlQ].en-GB.srt'),
# _voice = 'en_UK/apope_low'
# _voice = 'en_US/m-ailabs_low#mary_ann'
# _voice = 'en_US/vctk_low#p351'
# _voice = 'en_US/hifi-tts_low#92'
# voice_str = f'_{_voice.replace("/", "")}'





# image/descriptions provided by other SHIFT tool or Human curator

# https://simaviro.sharepoint.com/sites/SHIFT/Shared%20Documents/Forms/AllItems.aspx?csf=1&web=1&e=JNK8dQ&cid=363c253d%2D4d61%2D4db1%2D8ffd%2Ddedda749da2d&RootFolder=%2Fsites%2FSHIFT%2FShared%20Documents%2FGENERAL%2FWORK%20PACKAGES%2FWP1%2FContent%20Repository%2Fshift%5FSPK%5Fuse%5Fcases%5Fshare%2F02%5Fuc%5Fspk%5FLandscape2Soundscape%2FLandscape2Soundscape%5F12%5FMasterpieces&FolderCTID=0x01200058F5037C0101524B82F6F0788C02A563
# STATIC_FRAME = 'uc_spk_Landscape2Soundscape_Masterpieces_pics/01_Schick_AII840_001.jpg' #'assets/image_from_T31.jpg'




PIC_DIR = 'uc_spk_Landscape2Soundscape_Masterpieces_pics/'

DESCRIPTIONS = [
    # 1
    [
        '01_Schick_AII840_001.jpg',                               # image
        '01_Schick_AII840_001.txt',                               # text
        'Statue in shire hill on autumn beach.',                  # audiocraft
        'Gottlieb Schick - Bildnis der Heinrike Dannecker - 1802', # cv2 puttext title
        'en_US/m-ailabs_low#mary_ann',
     ],  
    # 2
    [
        '02_Constable_AI555_001.jpg',
        '02_Constable_AI555_001.txt',
        'Meadows country farm village in sight',
        'John Constable - Dorf an dem Flusse Stour - 1804',
        'en_US/m-ailabs_low#mary_ann',
    ],
    # 3
    [
        '03_Schinkel_WS200-002.jpg',
        '03_Schinkel_WS200-002.txt',
        'Arriving at the shore on horses',
        'Karl Friedrich Schinkel - Gotische Kirche auf einem Felsen am Meer - 1815',
        'en_US/m-ailabs_low#mary_ann',
    ],
    #
    [
        '04_Friedrich_FV317_001.jpg',
        '04_Friedrich_FV317_001.txt',
        'Land steppes',
        'Friedrich Caspar David - Der Watzmann - 1824/1825',
        'en_US/m-ailabs_low#mary_ann',
    ],
    #
    [
        '05_Blechen_FV40_001.jpg',
        '05_Blechen_FV40_001.txt',
        'fjords',
        'Blechen - Carl Unwetter in der römischen Campagna - 1829',
        'en_US/m-ailabs_low#mary_ann',
    ],
    # 6
    [
        '06_Menzel_AI900_001.jpg'
        '06_Menzel_AI900_001.txt',
        'Olive trees in Seville',
        'Adolph Menzel - Bauplatz mit Weiden - 1846',
        'en_US/m-ailabs_low#mary_ann',
    ],
    # 7
    [
        '07_Courbet_AI967_001.jpg',
        '07_Courbet_AI967_001.txt',
        'Storm at the strand of waves Tsunami',
        'Gustave Courbet - Die Welle - 1869/1870',
        'en_US/m-ailabs_low#mary_ann',
    ],
    # 8
    [
        '08_Monet_AI1013_001.jpg',
        '08_Monet_AI1013_001.txt',
        'Mai flowers blossom picnic',
        'Claude Monet - Sommertag - 1874',
        'en_US/m-ailabs_low#mary_ann',
    ],
    # 9
    [
        '09_Blechen_AII823_001.jpg',
        '09_Blechen_AII823_001.txt',
        'Cascade in Africa',
        'Carl Blechen - Wasserfälle bei Tivoli - 1832',
        'en_US/m-ailabs_low#mary_ann',
    ],
    # 10
    [
        '10_Boecklin_967648_NG2-80_001_rsz.jpg',
        '10_Boecklin_967648_NG2-80_001.txt',
        'Hades ades at it sisland',
        'Arnold Böcklin - Toteninsel - 1883',
        'en_US/m-ailabs_low#mary_ann',
    ],
    # 11
    [
        '11_Liebermann_NG4-94_001.jpg',
        '11_Liebermann_NG4-94_001.txt',
        'Tavern at the waterfront',
        'Max Tiebermann - Gartenlokal an der Havel. Nikolskoe - 1916',
        'en_US/m-ailabs_low#mary_ann',
    ],
    # 12
    [
        '12_Slevogt_AII1022_001.jpg',
        '12_Slevogt_AII1022_001.txt',
        'toy sailing yachts pool',
        'Max Slevogt - Segelboote auf der Alster am Abend -1905',
        'en_US/m-ailabs_low#mary_ann',
    ],
]


SILENT_VIDEO = '_silent_video.mp4'


# SILENT CLIP


for _img_, _text_, soundscape_text, _title_, _voice_ in DESCRIPTIONS[:20]:

    # cv2put txt
    im = cv2.imread(PIC_DIR + _img_)  # IMG must have EVEN shape
    h, w, _ = im.shape
    im = im[(h%2):, (w%2):, :]  # assure even image
    print(im.shape, "GLOBAL IM\n\n\n\n")
    fram = np.zeros((94, im.shape[1], 3), dtype=np.uint8)
    h, w, _ = fram.shape
    font                   = cv2.FONT_HERSHEY_SIMPLEX
    bottomLeftCornerOfText = (240, 74)  # w,h
    fontScale              = 2
    fontColor              = (255, 255, 255)
    thickness              = 4
    lineType               = 2
    cv2.putText(fram, _title_, #'LandScape 2 SoundScape',
        bottomLeftCornerOfText,
        font,
        fontScale,
        fontColor,
        thickness,
        lineType)
    offset_h = 24
    im[offset_h:h+offset_h, :w, :] = (.4 * im[offset_h:h+offset_h, :w, :] + .6 * fram).astype(np.uint8)
    # cv2.imshow('i', im); cv2.waitKey(); cv2.destroyAllWindows()

    # logo aud

    logo = cv2.imread('assets/audeering_logo.jpg')[:740, :, :]
    logo = cv2.resize(logo, (logo.shape[1]//2, logo.shape[0]//2))
    h, w, _ = logo.shape
    offset_h = im.shape[0] - h
    im[offset_h:h+offset_h, :w, :] = (.23 * im[offset_h:h+offset_h, :w, :] + .77 * logo).astype(np.uint8)

    # logo SMB

    logo = cv2.imread('assets/SMB_logo.png')#[:740, :, :]
    logo = cv2.resize(logo, (logo.shape[1]//2, logo.shape[0]//2))
    h, w, _ = logo.shape
    offset_h = im.shape[0] - h
    # fill logo SMB with the pixels of im - where SMB is empty
    ptc = im[offset_h:h+offset_h, :w, :]
    logo[logo == 0] = ptc[logo == 0]  # fill empty
    im[offset_h:h+offset_h, :w, :] = (.13 * im[offset_h:h+offset_h, :w, :] + .86 * logo).astype(np.uint8)

    # # logo shift

    # logo = cv2.imread('assets/shift_logo.png')#[:740, :, :]
    # logo = cv2.resize(logo, (logo.shape[1]//2, logo.shape[0]//2))
    # h, w, _ = logo.shape
    # offset_h = im.shape[0] - h #-274
    # offset_w = im.shape[1] - w #400
    # # # fill logo SMB with the pixels of im - where SMB is empty
    # ptc = im[offset_h:h+offset_h, :w, :]
    # # msk = np.tile(logo[:, :,0:1] > 252, [1,1,3])
    # # logo[msk] = ptc[msk]  # fill empty
    # im[offset_h:h+offset_h, offset_w:w+offset_w, :] = (.0 * im[offset_h:h+offset_h, offset_w:w+offset_w, :] + 1 * logo).astype(np.uint8)

    # silent video - img
    # im = cv2.resize(im, (700, 700))
    cv2.imwrite('pic_logo_emb.png', im)
    
    
    

    # raw, _ = soundfile.read(soundscape_file)  # 12345, 2

    # # fill
    # soundscape = []
    # for _replica in range(math.ceil(len(total) / raw.shape[0])+1):
    #     soundscape.append(raw)  # _shift non defined for stereo
    # soundscape = np.concatenate(soundscape, 0)

    # total = .36 * np.concatenate([total[:, None], 
    #                              total[:, None]], 1) + .64 * soundscape[:len(total), :]

    # outfile

    OUT_FILE = _img_.split('/')[-1].replace('.','__') + '.mp4'  # assets / -1
    print(f'{OUT_FILE=}\n')
    # call API passing img
    
    subprocess.run(
            [   
             "python",
             "tts.py",
             "--text", PIC_DIR + _text_,
             '--image', 'pic_logo_emb.png',
              # "--title", _title_,
              # '--soundscape_text', soundscape_text,
             '--voice', _voice_,
             '--out_file', OUT_FILE,
                ])

    # soundfile.write(AUDIO_TRACK, total, 22050)
    # subprocess.call(
    #     ["ffmpeg",
    #         "-y",
    #         "-i",
    #         SILENT_VIDEO,
    #         "-i",
    #         AUDIO_TRACK,
    #         #"-c:v",
    #         #"copy",
    #         "-map",
    #         "0:v:0",
    #         "-map",
    #         " 1:a:0",
    #         "-vf",
    #         "pad",
    #         OUT_FILE])