LTX-2は、私の低スペックのパソコンでも、日本語のセリフ入りのAI動画を比較的短い時間で生成できて便利です。
１０秒程度のAI動画なら１０分強で生成できます。
ただ、問題は、日本語の精度で、思ったようには発音してくれません。
こちらのAI動画を作ったプロンプトは、以下
A quiet train station platform at dusk. Soft orange light. A young Japanese man and a young Japanese woman stand facing each other. The woman says softly in Japanese: 「いろいろと ありがとう…ほんとうに」 The man looks down and replies: 「そろそろいくね」 She continues: 「きをつけてね」 He nods and says: 「うん…またあえるよ」 A train approaches with a low rumble. Soft wind, distant station sounds, and a quiet melancholic piano in the background. Emotional, realistic, cinematic lighting, no text.
この程度なら、まだいい方
A Japanese mature woman and her young handsome boyfriend are strolling through a park in Tokyo. The man is enticing the woman, saying in Japanese ：「いいじゃないのお」 .
The woman pretends to be unwilling, saying: 「だめよだめだめ」.
But the man persists, saying：「がまんできないよ おくさん」. Then the man pulls the woman closer and kisses her.
Two young Japanese people sitting across from each other in a small quiet cafe. The woman says hesitantly in Japanese: 「ねえ…はなしがあるの」 The man looks at her and replies: 「なに？」 She continues softly: 「ちょっと、きになることになったの」 He falls silent for a moment, then says: 「…できたのか」 Soft cafe sounds, warm indoor lighting, emotional and realistic atmosphere, no text.
これは、かなりひどい
A huge moss-covered lantern stands on the edge of an old pond surrounded by greenery. A large frog sits atop the lantern. Suddenly, the frog leaps into the pond, sending up a tremendous splash about 10 meters high. I want the frog to be depicted with realistic colors and form. A Japanese narration plays: 「ふるいけや かわず とびこむ みずのおと」 .
The sound of the splashing water is also included.
数撃ちゃなんとかで、練習あるのみでしょうね。
