Travel Vlog Pronunciation Fix

[Assets]
@Image1 = Reference for protagonist's face, hairstyle, physique, casual clothes, yukata
@Image2 = Ryokan exterior, entrance, noren curtain, plants
@Image3 = Guest room (table, window, furniture layout)
@Image5 = Simmered golden eye snapper, dishware, plating (serve on table from @Image3, do not use background)

[Summary]
7 seconds | Vertical 9:16 | 24fps.
A 34-year-old female travel influencer arrives at 'Shionagian' hot spring ryokan in Atami, talks to selfie camera, then time jumps to dinner in guest room where she tastes a bite of golden eye snapper. Live-action travel vlog.

[Global Settings]
Basic Environment & Texture: September. First half: soft afternoon natural light at entrance. Second half: warm indoor lighting at dinner time. Emphasize extreme realistic physical textures.
Visuals: Live-action, shallow depth of field, natural transition from daylight to indoor light.
Camera Language: One camera movement per cut. First half: natural slight shake of selfie held by subject. Second half: fixed camera centered on face and shoulders.
Character: Strictly fix facial features, hairstyle, physique from @Image1 throughout. Casual clothes in first half, yukata in second. Retain real fine pores and skin texture.
Core Acting: Straightforward, friendly travel influencer. No shouting, soft natural Japanese. Distance feeling like talking to one friend. Do not use ad-read tone.
Prohibitions: No subtitles, no BGM, do not show filming equipment, do not add walking or opening doors actions, do not pose holding food with chopsticks.

[Timestamps]
0-3s [Arrival & Greeting]
Scene: In front of ryokan entrance from @Image2.
Action: Protagonist in casual clothes stops, takes selfie with one hand. Framing above chest, face slightly above center, entrance and noren behind. Smiles naturally while looking at lens and speaking.
Dialogue: "kyouno oyado wa, shionagian. mite, ii kanji"
(Speak in hiragana, not kanji)
Pronunciation Guide: "kyouno" normal speed, light.
Small pause after "oyado wa,". Take breath here.
"shionagian" proper noun, smooth in one breath, no splitting.
"mite, ii kanji" switch to bright bouncing tone.
Look lightly towards entrance at "mite", sharing background.

Pronunciation / Pitch:
KYOU[H]->NO[L] / kyoo->no-> (Today - existing word - kammuri type)
O[L]->YA[H]->DO[H] / o->ya->do-> (Ryokan/Oyado - existing word - heiban type. Particle 'wa' continues high)
SHI[H]->O[L]->NA[L]->GI[L]->AN[L] / shi->o->na->gi->an-> (Shionagian - coined word, no standard accent.
Specified as kammuri type for rhythmic direction. Rise at 'shio', keep 'nagian' flat and low, no phrase break)
MI[H]->TE[L] / mi->te-> (Look/Mite - existing word - kammuri type)
I[L]->I[H]->KAN[H]->JI[H] / i->i->kan->ji-> (Good Feeling/Ii Kanji - existing word - heiban type)
Intent: Expression showing natural joy of arrival. Cheeks and eyes move softly with mouth. Retain breathing, blinking, small eye movements.
Camera: Fixed selfie, natural perspective. Do not distort face/arms with wide angle.

[Between Cuts]
Skip hours until guest room, hot spring, post-bath, dinner service. Prohibit hard transitions, connect with natural cut changes.
Time progresses from noon to dusk, no reversal.

3-7s [First Bite & Reaction]
Scene: Table in guest room from @Image3. Protagonist in yukata seated. Simmered fish from @Image5 served.
Action: Start with bringing small flaked piece of fish to mouth with chopsticks in right hand. Put in mouth, lower chopsticks.
Gesture of eating fish that melts like butter. Do not chew hard, let it dissolve naturally in mouth.
Food Texture: Simmered white fish. Delicately cooked flesh that crumbles easily with chopsticks. Contains plenty of moisture, melts like butter.
Sound: Flesh dissolves softly in mouth, muffled small chewing sound containing moist fat and broth. Only faint wet 'mogu, mogu' sounds heard.
Intent: Chew slowly and naturally, swallow unhurried. After swallowing, return gaze from food to lens, mouth corners relax into speech.
Dialogue: "un, oishii"
Intent: Relaxed smile showing satisfaction.
Camera: Fixed, no movement.

[Conclusion]
Throughout: Unify voice quality, speaking distance, volume. Dialogue exactly as specified. Sync mouth movements.
Ambient Sound: First half: only outdoor wind. Second half: only quiet indoor sounds, small dish sounds, soft chewing sounds. No BGM.
Pure video. No subtitles. Retain real fine pores and skin texture. Do not change character's face/clothes.