thanks for getting back to me @jetdv In terms of switching to who is speaking - you're right - it does so by having each camera angle paired to a vocal track. You enter who is on which camera and what mic is picking up their volume. It works with up to 10 speakers and 10 cameras. So in my setup, I use 2 audio tracks and 3 cameras, 1 camera and audio track for each speaker on the pod, and a third camera as a wide capturing both. It then does a good job of understanding who to cut to based on the audio. The settings can be adjusted to cut to the wide shot more or less, depending on taste. I've seen it cut between an hour long podcast, featuring 3 speakers and 4 cameras, within 4 minutes, and require minimal adjustments to its cuts from my part. This gives me a great rough edit that frees up a ton of my time, and lets me focus on other aspects of the video (layout, graphics popups etc.) that really enhance the production value of the pod. If you're into audio, it basically functions as a Sidechain would in music, that ducks a bassline when the kickdrum hits. In this case, it "ducks" one video feed for another, based on which audio track features dialogue.

Considering Vegas has script capabilities and audio plugins, i feel like this should be something that can be created for it. I also think Vegas is focusing on smaller creatives, and i think tools that would help podcast editors falls into that, so I'm hopeful someone sees the value of what AutoPod is doing.

@Gastor I don't know of any tools in VEGAS that splits the audio per person. If you can get the split audio on the timeline in VEGAS, a script can certainly adjust the video to match (even with settings for a third camera). I don't know if the "whisper" audio translation tools used by some on this forum is capable of distinguishing between speakers and labeling/separating them.

Take a look at this tutorial:

I believe the person requesting this was actually using Premiere to split the audio and pick the different speakers. Then this script would take, for example, all of "Speaker 1" and move those blocks to a new track. Once "Speaker 1" and "Speaker 2" are on different tracks, then it's fairly trivial to select which camera to use. Instead of "ducking" cameras, what I typically do is add a new track and add the camera snippets to that new track. In fact, you can even add the other camera angles as "takes" making it easy to go to any section and Press "T" to switch between the angles.

You're correct. (Sorry if i wasn't explaining it clearly earlier.) They do require you use different audio tracks with each speakers/camera audio being captured separately. That way, matching the camera to the speaker is easy as the audio from the track that each camera angle is linked to, only has the dialogue from one speaker.

This promo video showing the edit, and the speed of the cuts are a real-time example of how fast it goes through the podcast. It's incredible. I'd love to see a version of this for Vegas.

The tutorial you shared is impressive. While it's not what's needed to create what AutoPod is doing, it does give me hope that it's possible. Thanks for sharing it. It opens up a few other ideas.

@Gastor, would you give me a sample on the VEGAS timeline so I can see what the original source would look like in order to generate the final output?

I'd need the original video/audio files as well for testing.

Yes I can @jetdv I'll have some stuff ready for you today

let me know if you got my DM with the link @jetdv I'm excited to hear your thoughts on what's possible

@Gastor, I did receive it, I have gotten the files downloaded (haven't had a chance to look at them yet) but it might be next week before I can really start to look at this. It's a busy season... Thanks for the files as they will definitely help as I start looking into this.

btw @jetdv it looks like someone is working on something similar for Resolve and Final Cut. https://www.podmate.io/

This one technically is a stand alone app that would render a final project file that could then be uploaded into Resolve or Final Cut. So not exactly the same as Autopod, which is a plugin for premiere

would love to see what you come up with for Vegas

@Gastor Cool. I wonder if it's the same person? Sounds like they're mainly Mac editors.

I spoke with Mo from podmate this past week. He's made some great progress. He's building something separately from the autopod team. I got to demo the current beta version of his podmate, which does the primary thing that autopod does - uses the vocal tracks from each video track to switch between active speakers, and also he's figured out how to also occasionally do a split screen with both speakers as an establishing shot. This is awesome for podcast editors. Unlike autopod, his won't be a plugin or a script, instead, its a separate app that spits out a new XML file that can be loaded in resolve or vegas. Would love to see what you're able to come up with @jetdv