SYSTEM NOTICE

Auto translation by AI. Be sure, accuracy, nuances and authorial intent may not be fully reflected.
見出し画像

I just added music and YAML. My MV is complete thanks to ChatGPT Work


"I've made a song with AI, but making a music video seems like a lot of work."

I imagine many of you feel that way.

Creating an MV requires a lot of tasks, such as adjusting lyric timing, creating images, adding subtitles, synchronizing with the music, and exporting the video.

However, this time, I was able to complete an MV with almost no manual work by combining "Lyrics to SRT Converter V2 + MV Block" and "ChatGPT Work".

The work I did was to load the music file and lyrics, output the MV design YAML, and pass it to ChatGPT Work.

After that, ChatGPT Work handled the image creation, subtitle insertion, synchronization with the audio, and exporting to MP4.

In this post, I will introduce that process.


Tools used

I used the following two tools this time.

  • Lyrics to SRT Converter V2 + MV Block

  • ChatGPT Work

I just added music and YAML. My MV is complete thanks to ChatGPT Work

With Lyrics to SRT Converter V2 + MV Block, you can load music and lyrics and adjust the timing of the lyrics display.

Furthermore, by selecting "MV Design YAML Output," you can export a YAML file that summarizes information such as the song structure, lyric timing, required images, and subtitles.

This YAML serves as the blueprint for MV production.


Step 1: Load music and lyrics

First, load the music file into Lyrics to SRT Converter V2 + MV Block.

Next, load the lyric text or SRT file.

Check the timing of the lyrics while playing the music, and adjust as necessary.

By grouping the instrumental sections of the song or lyrics you want to keep together as the same scene, it becomes easier to organize the MV scene structure.


Step 2: Output the MV design YAML

Once you have finished timing the lyrics and dividing them into blocks, select "MV design YAML output".

Then, a YAML file containing the information necessary for MV production will be exported.

The YAML mainly contains the following information:

  • Total length of the song

  • Start and end times for each lyric

  • Intro, interlude, and ending

  • Image file name for each scene

  • Timing for displaying subtitles

  • File name of the completed video

  • Required delivery files

At this stage, the blueprint for the MV is almost complete.


Step 3: Input the music and YAML into ChatGPT Work

Next, open ChatGPT Work.

I attached the following two files to the conversation screen:

  • Music file

  • MV design YAML file

Then, I requested it to create the images and complete the MP4.

From this point on, ChatGPT Work read the contents of the YAML and proceeded with the MV production.


Tasks performed by ChatGPT Work

Here are the main tasks that ChatGPT Work performed this time.

  • Check the length of the audio source

  • Analyze the content of the song and emotional changes

  • Generate images for the MV

  • Assign each image to the corresponding lyric scene

  • Create lyric subtitles

  • Synchronize subtitles with the music

  • Export the MP4 video

  • Create a thumbnail

  • Create an SRT file for editing

  • Create title suggestions for YouTube

  • Create the description and tags

  • Package images and videos into a ZIP file

Just by providing the music and YAML, an MV combining images, subtitles, and audio was completed.


What I learned from actually making it

The process up to creating the initial video proceeded almost automatically.

However, I did make some adjustments to finish it into a more natural and readable MV.

For example, in areas such as the following:

  • Adding more images so that the same image does not repeat

  • Change lyric subtitles to white text

  • Adjust pink border to 4 pixels

  • Add a 1-pixel black shadow

  • Correct to show shoes off while indoors or on the bed

  • Create a lightweight MP4 version for easy downloading

It was effective not to just leave everything to the AI, but to have a human check the finished video and have specific parts that were concerning corrected.

Instead of saying 'make the subtitle border a bit thinner,' it is more accurate to use numbers, such as 'make the pink border 4 pixels.'


Completed MV

Here is the completed MV.

Song Title: I Won't Search for You Anymore

I was able to complete an MV with images from music and YAML.


Reference video from the instructor

Here is the instructor's video that inspired this production method.

I think it is easier to understand if you check the steps in the video before actually performing the operations.


Convenient aspects of this method

A major benefit of this method is that you don't have to sync the lyrics one by one in video editing software.

Here are the points I found particularly convenient.

  • Lyric timing can be organized in YAML

  • MV scene composition can be accurately conveyed to the AI

  • Everything from image generation to MP4 creation can be done in a single workflow

  • You can communicate revisions through text

  • Easy to try even if you are not used to video editing

I think this significantly lowers the barrier to entry for those who have previously felt that MV production was difficult.


Summary

The production steps for this project can be summarized as follows:

  1. Load music and lyrics into Lyrics to SRT Converter V2 + MV Block

  2. Organize lyric timing and scene blocks

  3. Output the MV design YAML

  4. Attach the music and YAML to ChatGPT Work

  5. Request image generation, subtitles, synchronization, and MP4 creation

  6. Check the finished video and revise only the necessary parts

Even if you can create music, sometimes video production is so difficult that you cannot proceed to publication.

However, by summarizing the MV design information into a YAML file and passing it to the AI, I was able to significantly reduce the time-consuming tasks that were previously required.

"I want to turn the music I made with AI into an MV that I can publish on YouTube."

For those people, I think this is a production method worth trying.

*Processing time and output results may vary depending on the usage environment and file size. After completion, be sure to check the synchronization of subtitles, images, and audio.


いいなと思ったら応援しよう!

Mii よろしければ応援お願いします! いただいたチップはクリエイターとしての活動費に使わせていただきます!