I searched the forum for "podcasting animations" and didn't find anyone asking this question so I will post it here.

My friends and I would like to do a live video podcast, talking about a particular topic. But instead of video feed of our actual face I'd like to create a talking head animation for each person. I'm not sure if a character can respond to real-time audio or if it needs to be edited after the fact.
In the end I'd like 3-5 talking heads on a screen having a conversation, mimicking the actual real live podcast audio. Anyone know how I can approach this? Which software would be involved?

you mean something like this ?
http://www.reallusion.com/CTCamSuite/30/Help_GNS/ENU/Default.asp
it doesn't exist.

• My Spotify  My Twitter  My Virtual Art Gallery
  
• My NFT    My Marketplace My Blog
  
• Music Sync Licensing

☯🐉

"To define Tao is to defile it" - Lao Tzu

Ok thank you for that idea; that software looks interesting but I guess you're saying it no longer exists. My other idea for a CrazyTalk-inspired podcast was to mimick what this person did: https://youtu.be/y6Z9nhbarf8

I would use CrazyTalk to create a character of each of podcast participant. I would then insert those characters onto a static scene. For example, if the podcast topic was about reviewing the last Star Wars movie I'd have each participant's character sitting inside a space ship. I wouldn't need too much animation, but I would need their character to talk in sync with their respective audio. In order for each participants' audio to sync with their respective character would I need an individual audio file from each podcast participant? If I only had one audio containing everyone's voices together I assume it would be ALOT of audio editing time and effort to sync the right voice with the appropriate character. I'm still learning the workflow so just trying to understand if a simple process exists regarding tying multiple audio tracks with their respective characters.

ya,  that's doable -  you can do that green screen in a video editor,  just do the characters one at a time in crazytalk, and composite / synch the vocals in your video editor -   I would use sony movie studio / vegas, being that it allows easy manipulation of multiple tracks ... chop up the different actor lines into segments -  create each segment individually per actor and make the parts in crazy talk-   then import that performance,  using the segment to realign to a main audio track to get it back in synch to the main performance.  -  the whole background of the scene should be 1 image,  and the individual actors overlays..

so track 1,  would be whole unedited audio performance,
track  2 - would be a background of your stage setting
track 3 - would be your actor green screend into the video - with the audio muted, you then use the muted wave file visualization to align it with the main audio track to get everything back in synch.
tracks 4 to ? = would be additional actors you green / screen - chroma into your video.


I did this video with ct6 in that way

• My Spotify  My Twitter  My Virtual Art Gallery
  
• My NFT    My Marketplace My Blog
  
• Music Sync Licensing

☯🐉

"To define Tao is to defile it" - Lao Tzu

Thanks for the thorough answer. I have basic knowledge in Adobe Premiere Pro and can probably synch the background, character overlays, and audio tracks within Premiere. That's a good idea. I assume CrazyTalk's purpose is to create the talking animation for the character's face. Will CrazyTalk allow me to upload the entire podcast audio clip (maybe 30 minutes) to create a character's entire animation? Or will I be forced to break it into segments?

I may want to give the characters a body instead of only a talking head. But from your answer I'm thinking about photoshopping the background scene WITH the bodies (but no heads). I would import the talking heads from Crazy Talk into Adobe Premiere as an overlay.

yes on segments,  the bigger the file,  the more problems, issues and accidents your project becomes prone to.

You may want to map this out in a story board first,  - for example,  jump right to the editing phase of your video,   then create the parts to put together as a puzzle.   I believe crazytalk will let you do up to 15 minute segments,  but I wouldn't advise going that big a shot.  the smaller the better,  or rather,  the smaller the easier it is to manage the project. 

in comparison to how you originally wanted to do the project,  this approach obviously will take longer - for the sake of stability ( and sanity lol - when and if you try a longer type project,  you'll know what i mean..) ,  but in the end the quality of production will be better because you paid more attention to the detail per segment.  -  and i'm sure it will get easier as you get used to the workflow -  you could find your sweetspot in terms of clip sizes. 

• My Spotify  My Twitter  My Virtual Art Gallery
  
• My NFT    My Marketplace My Blog
  
• Music Sync Licensing

☯🐉

"To define Tao is to defile it" - Lao Tzu

Ok, I like your workflow idea: I will start with the Video Editor part as far as importing the Podcast's master audio track (for purposes of syncing up smaller audio clips with it) and getting my static background imported(I don't plan on any moving background for simplicity's sake). (By the way, does Reallusion's Animator application handle the same process that a video editor would be doing?) The next step would be to create the smaller character animation segments within CrazyTalk (like you said, quality over convenience with this process). The challenge is that my vision for this initial "cartoonized podcast" project was not to create a story from scratch by chopping up my existing 30 minute podcast audio and rearranging them; I was hoping to just import the entire audio file and have CrazyTalk automatically animate, so I will just have to dive in and get my hands dirty to see how this workflow will work. If the podcast is a hit I would like to have a workflow in place as streamlined as possible.

Once again, I appreciate you sharing your experience and saving me some mistakes.

one thing you can do is watch some of your favorite podcasts,  and maybe take notes of how long their segments were. 

I've done the same thing in my case,  which is more music videos,  and I'll count how many seconds between edits.  Ive found that long segments can actually be exhausting to the viewer,   consider how naturally we don't stare at one thing for very long,  and if we do,  it actually becomes mesmerizing and can actually become distracting to your topic. 

So i'd suggest you watch a few podcasts that you enjoy very much,  and just take notes on timings between edits and camera angles. You will be surprised at how much that practice will improve the production.

and I've found that cutting small segments,  is actually much easier and faster than cutting long ones......especially since each segment usually comes from the same file -  so it's not like you have to create a whole new set and new characters ....it will go faster than you think.  Doing the story board just helps keep a goal in perspective -  you really don't want to be a creative and a techie at the same time - in my experience the techie aspect of making your scene always kills the creative aspect -  so this workflow will allow you to be completely creative in the story board stage,  and completely technical in the scene cutting stage.  -  i've gone through tons of writers blocks to learn this lol 

Best wishes on your project,  and please share when you are done,  I'd love to see what you come up with !!

• My Spotify  My Twitter  My Virtual Art Gallery
  
• My NFT    My Marketplace My Blog
  
• Music Sync Licensing

☯🐉

"To define Tao is to defile it" - Lao Tzu

great advice, I will share when I'm done