Descript Overdub Tutorial: CLONE Your Voice with AI
Вставка
- Опубліковано 2 лип 2024
- In this Descript tutorial, I guide you through the process of creating, cloning and customizing your own AI voice in Descript. I start by creating a new AI speaker from scratch, submitting my training data and verification phrase. This will both verify and train the voice.
Then I create a new Descript project, add some text using the Write tool and generate speech with the newly created AI voice using Overdub. I cover key functionalities such as assigning the AI voice to text, generating speech, adding gaps, controlling the Overdub files/segments, converting dynamic overdubs into editable audio files.
Chapters:
00:00 Introduction
00:22 Navigating Descript to Create an AI Speaker
00:55 Recording Your Verification Statement
02:25 Fine-Tuning and Submitting Your AI Voice
03:11 Testing and Sharing Your New AI Voice
03:54 Advanced Editing and Overdub Features
08:31 Converting Overdubs to Audio Files
10:08 Final Thoughts and Tips
------------------------------------------------------------------------------------------
Join my Media Makers Club, a membership where you get:
* Access to the community to ask questions, interact with other editors
* Free access to all of my courses and mini courses (all of them!)
* Free access to live webinars and Q&A Sessions
* Early access to ad-free UA-cam videos and experiments
* Free downloads (Camtasia assets, music, templates)
* 1-on-1 Coaching Session Discounts
Join here: www.mediamakers.club
Would you like to learn one-on-one with me?
cotovan.com/coaching
More Descript tutorials:
• Descript Tutorials
Subscribe for goodies, tips, live webinar announcements, early access and more:
cotovan.com/newsletter
Join our Descript Community on Facebook:
/ descriptusers
Are my tutorials useful? Buy me a coffee! I LOVE coffee:
www.buymeacoffee.com/cristi
Cristi's website and blog:
cotovan.com
-- SOFTWARE I use and recommend * --
Camtasia (includes 10% OFF coupon: CRISTI10): cotovan.com/camtasia
Snagit (includes 10% OFF discount code CRISTI10): cotovan.com/snagit
Transcription and Captions: cotovan.com/descript
Remote recording: cotovan.com/riverside
Audio clean up and processing: cotovan.com/auphonic
* Disclosure: Some of these links take you to my website and some of them are affiliate links where I will learn a small commission if you make a purchase. This is at no additional cost to you and they help support my work. Thank you!
#cristicotovan #descript #overdub
Descript courses, minicourses, Q&A sessions, coaching calls with me, all available in the MediaMakers.club
Thanks Cristi for this info!!!
Thanks, glad it helped.
That was really cool..
Thanks a lot!
thank you
You're welcome 😊
Hi Cristi! I'm trying to decide on a video editing tool as a beginner. Thanks for all the info. Super helpful! Quick question: I have a subscription to Elevenlabs and want to use AI generated voice recordings from this tool to narrate my videos. Can I insert these elevenlabs audio files into videos I edit on Descript OR do I have to use Descript generated audio files?
You can use external audio (narration) in Descript and add it but it's quite difficult to sync it to existing footage. Descript makes it very difficult to stitch and match audio to existing footage AND footage to existing audio.
I think it's because of the way scenes work - since Descript is mainly text-driven, everything you add has to be anchored in text somehow. Now they have scenes and whatnot, but effectively, it's all based on the transcript or the script track at the bottom. Therefore, it can be very hard to bring new audio and make it play at very precise points in the timeline where footage exists.
It is also hard because you can't really see all the frames of the existing video in the timeline and again, syncing it to existing footage is hard.
For what you describe, I would use Camtasia any day. And it's easy to learn too.
Tried many times it keeps rejecting the recordings and saying we didnt match the statement, stupid thing.
I saw that in the past, I think it was a bug. You should reach out to support about it
Hey @Graphicious do you know if there is a limit to the amount of AI voices you can create? Thanks in advance for the info!
I don’t think so, I haven’t seen a limit mentioned anywhere. How many do you need? :)
@@Graphicious I have some clients I make content for and I wanted to have like 2-3 voices per clients, for social media content. Excited, informative, serious, etc. So, as of right now, about 12.
Also, I didn't know about the 10k cap for the AI voice generation, that's words I'm assuming?
I think the cap of 10k refers to ‘overdubs’. Previously, Descript created one overdub per sentence. With the last iteration of the AI voice, they are now generating one overdub per paragraph. Helps with better context and intonation, they say. So I am no longer certain what that limit means now because they change things every 5 minutes, nobody can keep track. I am pretty sure they are having trouble too :)
As for your question about the voice, you can ask your clients to create their voices using their own Descript accounts and share them with you.
@@Graphicious LoL, got it and very true about the changes. I'm loving the new UI and features, cuts out extra tools I had before.
My clients are the last tech savvy people in the world, but I'll give it a shot!
Thanks as always my man, and have a good one!
Yes, this all works as stated . . .
But, with the Creator plan they only have a vocabulary of 1000 words, all other words are filled with "Jibber," and "Jabber." This is VERY deceptive on Descripts part. They go on to state you must update to the Pro plan to make your cloned voice worth using.
I find them deceptive in other ways as well.
That’s correct, only the Pro plan allows full speech generation