AI Matchstick Video Prompt: Yadda Ake Kirkirar Ashana Masu Waka da Rawa

 

Yadda Ake Kirkirar Bidiyon Ashana Masu Waka da Rawa Ta Amfani da AI

A yau fasahar Artificial Intelligence (AI) ta sauya yadda ake kirkirar hotuna da bidiyoyi. Abin da a da sai an yi amfani da manyan kayan aiki da lokaci mai yawa wajen samar da shi, yanzu ana iya yin sa cikin ‘yan mintuna ta amfani da AI. Daya daga cikin abubuwan da suka fi jan hankali shi ne mayar da hoto tsaye ya zama bidiyo mai motsi, musamman hotunan kirkira irin na mutanen da aka tsara da sandunan ashana.

Hoton da aka yi amfani da shi a wannan misalin yana nuna mutane uku da aka gina da sandunan ashana. Daya daga cikinsu yana da wutar ashana a kansa, yayin da sauran biyun suke a bangarorin hagu da dama. Manufar prompt din ita ce a sa wadannan halittun su zama kamar suna raye, su fara waka tare da rawa cikin yanayi mai ban sha'awa.

Wannan nau'in bidiyo na iya zama mai kyau sosai ga masu kirkirar content a TikTok, Facebook, YouTube Shorts, Instagram Reels da kuma shafukan yanar gizo.

Menene AI Video Prompt?

AI video prompt rubutu ne da ake bai wa wani AI video generator domin bayyana abin da ake son ya faru a bidiyon. Prompt din yana gaya wa AI irin motsin da characters za su yi, yadda camera za ta motsa, irin hasken da ake so, yanayin bidiyon da kuma yadda scene din zai kasance.

Misali, idan ka rubuta kawai:

“Three matchstick people dance.”

AI zai iya fahimtar cewa mutane uku na ashana za su yi rawa, amma sakamakon zai iya bambanta da abin da kake zato.

Saboda haka, idan kana son sakamako mai kyau, ya fi dacewa ka bayyana abubuwa kamar character, action, camera movement, lighting, style da audio. Google DeepMind ma yana bada shawarar a bayyana framing na camera, motsin camera, style, lighting, character, location da action domin samun karin iko akan sakamakon bidiyon.

Yadda Prompt Din Ashana Yake Aiki

A wannan misalin, an fara da hoton ashana guda uku. Wannan hoton shi ne reference image. AI video generator yana amfani da hoton domin fahimtar siffofin characters, matsayin su da yanayin scene.

Prompt din ya ce:

“Create a 10-second cinematic video using the uploaded image as the exact reference. The three matchstick characters come alive, singing a joyful song and dancing together rhythmically…”

Ma'anar wannan ita ce a yi amfani da hoton da aka saka a matsayin tushen bidiyon, sannan a sa characters din su fara motsi.

Wannan hanya tana da amfani saboda ba sai an fara kirkirar sabon scene daga sifili ba. Hoton da ka riga ka mallaka zai zama tushen animation.

Me Ya Sa Aka Yi Amfani da “Exact Reference”?

Kalmar “exact reference” tana nufin a yi kokarin kiyaye abubuwan da suke cikin hoton farko. Wannan yana da muhimmanci idan kana son characters din su kasance kamar yadda suke a hoton.

A irin wannan scene, abubuwan da ake son a kiyaye sun hada da:

• Mutane uku na ashana.
• Matsayinsu a cikin hoton.
• Siffar jikinsu.
• Sandunan ashana.
• Wutar da ke kan character na tsakiya.
• Background.
• Haske da inuwa.

Duk da haka, babu prompt da zai iya bada tabbacin cewa AI ba zai canza wani abu kwata-kwata ba. Saboda haka, idan generator ya sauya wani bangare, ana iya sake gyara prompt din ko a sake generate bidiyon.

Me Ya Sa Aka Ce Su Rawa?

A prompt din an bayyana cewa characters din za su “dance together rhythmically”. Wannan yana bai wa AI bayani cewa ba kawai za su tsaya suna motsi ba, amma za su yi rawa cikin yanayi mai daidaito.

Wannan ya fi amfani fiye da rubuta kawai “make them move”.

Idan kana son rawa ta zama mai kyau, zaka iya bayyana motsin hannu, kafa da jiki. Misali, zaka iya cewa character na tsakiya yana daga hannunsa yayin waka, yayin da na hagu da dama suke motsa kafafunsu da hannayensu bisa rhythm.

Google DeepMind yana bayyana cewa idan ana son karin iko a kan complex action, yana da kyau a bayyana abin da zai faru cikin dalla-dalla.

Me Ya Sa Aka Saka “Cinematic”?

Kalmar “cinematic” tana taimakawa wajen bayyana irin salon bidiyon da ake so. Ana son bidiyon ya kasance kamar wani gajeren fim ko music video maimakon animation mai sauki.

A cikin prompt din an hada da:

Cinematic lighting

Dynamic camera movement

Detailed shadows

4K quality

Wadannan kalmomi suna taimakawa wajen bayyana yanayin visual da ake nema.

Amma yana da kyau kada a dogara da keywords kawai. Jagororin Runway sun nuna cewa prompt mai sauki, kai tsaye kuma mai bayyana abin da zai motsa a scene na iya taimakawa wajen samun sakamako mai sarrafuwa.

Camera Movement

Camera movement wani muhimmin bangare ne na video prompt.

Misali, zaka iya amfani da:

• Slow push-in.
• Smooth camera tracking.
• Gentle camera pan.
• Medium shot.
• Wide shot.
• Close-up.
• Dynamic camera movement.

A wannan bidiyon, camera na iya fara da kallon characters uku gaba daya, sannan ta dan matsa kusa da su yayin da suke fara waka da rawa.

Idan kana amfani da AI video generator wanda ke fahimtar camera instructions, wannan zai iya taimaka wajen sa bidiyon ya zama kamar an shirya shi da gangan maimakon motsi marar tsari.

Google ya kuma bada misalan umarnin camera kamar static, locked-off, push-in da sauran motsin camera a jagororinsa na prompting.

Wakar Characters Din

Abin da ya fi ban sha'awa a wannan prompt shi ne cewa characters din ba rawa kawai suke yi ba; suna kuma waka.

A prompt din an ce:

“The three matchstick characters come alive, singing a joyful song and dancing together rhythmically.”

Wannan yana bayyana cewa AI ya kamata ya hada movement da musical performance.

Idan AI video generator din yana da damar samar da audio, zaka iya kara bayanin irin sautin da kake so. Misali:

Audio: cheerful rhythmic music, light percussion, playful singing and synchronized foot movements.

Google DeepMind ya nuna cewa ana iya hada bayanin visuals da sound design a prompt domin samar da bidiyon da ke da yanayin sauti mai dacewa da scene.

Yadda Za Ka Yi Amfani da Prompt Din

Mataki na farko shi ne ka ajiye ko ka samu hoton ashana mai kyau. Hoton ya kamata ya kasance clear kuma ba shi da manyan kuskure.

Mataki na biyu shi ne ka bude AI video generator wanda ke goyon bayan image-to-video.

Misali, Google DeepMind yana da Veo, wani video generation model da ake iya gwadawa ta wasu kayayyakin Google.

Google DeepMind – Veo

Ka saka hoton a matsayin reference image, sannan ka saka prompt din da aka shirya.

Mataki na uku shi ne ka zabi tsawon bidiyon idan kayan aikin da kake amfani da shi yana baka damar zabar duration. A wannan misalin, manufar ita ce bidiyo na kusan seconds 10.

Mataki na hudu shi ne ka generate bidiyon.

Bayan an gama, ka duba ko characters din suna motsi yadda ake so, ko suna rawa tare, ko wutar ashana ta kasance mai kyau, kuma ko camera movement ya dace.

Me Za Ka Yi Idan Sakamakon Bai Yi Kyau Ba?

Kada ka damu idan generation na farko bai yi daidai ba. AI video generation yana bukatar gwaji da gyara.

Idan characters din suna motsi da yawa, zaka iya rage prompt din ka mayar da hankali kan motsi mai sauki.

Idan camera tana motsi fiye da kima, zaka iya rubuta:

“Smooth controlled camera movement, stable framing.”

Idan kana son rawa ta zama synchronized, zaka iya cewa:

“The three characters perform simple synchronized dance movements.”

Idan kana son character na tsakiya ya zama singer, zaka iya cewa:

“The center character leads the song while the two side characters dance.”

Runway ma yana bada shawarar a yi amfani da kalmomi masu kyau da kai tsaye wajen bayyana abin da ake son ya faru, maimakon dogaro da negative prompts.

Me Ya Sa Hoton Farko Yake Da Muhimmanci?

Ingancin reference image yana da matukar muhimmanci. Idan hoton yana da blur, distortion ko abubuwan da ba su da kyau, AI na iya samun matsala wajen fahimtar abin da yake gani.

Saboda haka, kafin ka yi image-to-video, ka tabbatar hoton yana da:

• Kyakkyawan resolution.
• Characters masu bayyana.
• Background mai tsabta.
• Kyakkyawan lighting.
• Babu abubuwan da suka rikitar da AI.

Wannan ya dace da shawarwarin Runway na amfani da high-quality input image wajen image-to-video.

Misalin Prompt Mai Kyau

Ga prompt din da za ka iya amfani da shi:

“Create a 10-second cinematic video using the uploaded image as the exact reference. The three matchstick characters come alive, singing a joyful song and dancing together rhythmically. The middle character leads the singing while the two characters on the sides dance beside him. Keep the original matchstick design, composition, flame, background and character positions consistent. Smooth synchronized movements, realistic matchstick physics, dynamic but controlled camera movement, beautiful studio lighting, detailed shadows, cinematic atmosphere, ultra-detailed 4K quality, joyful musical energy.”

Wannan prompt ya hada abubuwa masu muhimmanci: reference image, characters, action, singing, dancing, camera, lighting, consistency da quality.

Ina Za Ka Iya Gwada Irin Wannan?

Idan kana son koyon karin bayani game da Veo da yadda ake rubuta prompts, zaka iya amfani da official guide na Google DeepMind:

Veo 3 Prompt Guide – Google DeepMind

Haka kuma, Runway yana da cikakken jagorar rubuta prompts na video, musamman wajen bayyana visual description da motion description.

Runway – Text to Video Prompting Guide

Kammalawa

Wannan misalin na mutanen ashana yana nuna yadda AI zai iya mayar da hoto mai sauki zuwa wani gajeren bidiyo mai ban sha'awa. Sirrin samun sakamako mai kyau ba wai rubuta dogon prompt kawai ba ne; abu mafi muhimmanci shi ne a bayyana abin da characters za su yi, yadda za su motsa, irin salon camera, lighting, environment da audio idan ana bukatarsa.

Idan kana amfani da hoton ashana a matsayin reference, ka tabbatar hoton yana da kyau sannan ka rubuta prompt mai bayyana movement a sarari. Ka fara da simple version, ka generate, sannan ka gyara abin da bai yi daidai ba. Wannan hanyar gwaji da gyara tana taimakawa wajen samun sakamakon da ya fi kusa da abin da kake so.

Saboda haka, wannan 10-second matchstick singing and dancing prompt zai iya zama kyakkyawan misali ga masu son koyon AI video creation, musamman masu kirkirar gajerun bidiyoyi na nishadi. Ana iya amfani da irin wannan dabara wajen kirkirar characters na ashana, takardu, toys, abubuwan gida ko wasu abubuwa sannan a sa su motsa, su yi waka ko su yi wani labari.

Fasahar AI video tana ci gaba da samun sabbin damar sarrafa characters, scenes, camera da audio. Saboda haka, yana da kyau mai kirkirar content ya rika gwada sabbin prompt techniques kuma ya koyi yadda kowanne AI video tool yake fahimtar umarni.

A karshe, ka tuna cewa kyakkyawan reference image + bayyanannen action + camera direction + lighting + style + audio = prompt mai karfi. Ka yi amfani da wannan tsarin wajen kirkirar naka bidiyon ashana, sannan ka yi iterations har sai ka samu sakamakon da ya dace da manufarka.