I’m a long-time YouTube video producer – these 3 AI tools help me do it better and faster

I all the time like discovering new methods to use synthetic intelligence (AI) instruments to my day-to-day productiveness duties. Final yr, I confirmed how I used generative AI to rescue some dangerous audio and in any other case tweak a brief how-to video. I used Photoshop’s Generative Fill, Adobe Podcast, and what was then a brand new background alternative characteristic in Closing Lower Professional.

This time, I am utilizing an AI gimbal to assist the digicam observe my actions, Apple’s Voice Memos AI transcription characteristic in MacOS Sequoia to transcribe an unscripted video, and ChatGPT to recommend titles, tags, and an outline for an unboxing video.

Let’s begin with the challenge. I do movies for my YouTube channel as usually as I can, however my major work product is writing. So I attempt to discover methods to optimize my restricted non-writing time for my varied YouTube initiatives.

The video I labored on most not too long ago was the unboxing of a multi-filament 3D printer. The Anycubic Kobra 3 Combo can print utilizing as much as 4 colours directly. Unboxing movies have all the time been in style with my viewers, so I needed to get the video performed shortly.

1. Computerized digicam: Hohem iSteady v3 gimbal

The problem with unboxing is that it is usually onerous to know what to movie as a result of I by no means know what’s contained in the field till I open it. One of the simplest ways to make certain I get good movie is to position a bunch of cameras throughout my work space, after which simply do my unboxing factor.

The issue is, I usually transfer across the workshop whereas unboxing. In earlier movies, I would usually wind up with pictures the place I am out of body, or coming out and in of body. I attempted some auto-follow gimbals previously, however they all the time received confused until I used to be dealing with the gimbal instantly always.

Not this time.

I picked up the Hohem iSteady v3 gimbal on sale at Amazon for $100. (It is often $129.) I watched a number of opinions of this gimbal, and started to understand that gimbal AI has come a great distance previously yr. This gimbal has a complete bunch of app-assisted options, however what I preferred most is that it has an “AI module” that orients the gimbal correctly, no matter whether or not you are working an app, and even what you are utilizing for a digicam.

Even when you do not have the app put in, the gimbal responds to a couple easy hand gestures. I’ve but to put in the app and I’ve made an awesome video with wonderful monitoring of my motion.

The setup is tremendous simple. Cost it by way of USB C, then pull out the little built-in tripod legs and insert your digicam. I used my previous iPhone SE within the little clamp. Lengthy urgent the ability button turns it on. It should auto-calibrate, setting your cellphone to movie in portrait mode.

To change to panorama mode, you merely level each your thumbs to the left. Then, give it the OK signal and it’ll monitor you as you stroll round.

This gimbal utterly solved my out-of-frame drawback proper out of the field as a result of the onboard machine studying within the AI module tracked me completely. It tracked me accurately after I moved behind a workbench and behind the massive field I used to be unboxing. It tracked me after I walked towards the digicam and after I rotated and walked away. The one time it misplaced monitor of me was after I walked utterly out of the room, and all I needed to do to get its consideration once more was maintain my hand up within the OK signal.

Along with the cellphone within the gimbal, I used a second iPhone pointed down from a excessive vantage level. I additionally used two iPads that have been filming from their front-facing cameras so I may watch what was on-frame whereas filming. Sure, the front-facing cameras are somewhat decrease in decision, nevertheless it’s well worth the trade-off to have a built-in monitor always.

2. Transcribing audio: MacOS Voice Memos

This video was solely off the cuff, so I did not have a pre-written script I may feed into YouTube for closed captions. I additionally did not have a script to provide to ChatGPT to assist me with website positioning and tag options.

As a substitute, I simply recorded my commentary into the DJI Mic 2, which was linked by way of Bluetooth to considered one of my iPads. After recording into all 4 iOS units, I transferred the video into Closing Lower Professional and used the multicam characteristic to match up the timing of all 4 digicam angles. That allowed me to simply change between angles throughout modifying by merely typing 1, 2, 3, or 4, comparable to whichever digicam I needed to indicate footage from at that time within the movie.

To get an audio file appropriate for transcription, all you must do is open the finished video file produced by Closing Lower in QuickTime Participant. Underneath the File menu, choose Export As > Audio. You are not given a selection of codecs, so that you’re caught with m4a. Luckily, it will work for our functions.

Subsequent, open the Voice Memos app in Sequoia. This may not work on earlier variations of MacOS. There is not any import possibility in Voice Memos, however in case you drag and drop your m4a audio file onto the checklist of recordings, you will briefly see a inexperienced plus signal and it is going to be accepted into the checklist of clips. Observe that Voce Memos locations your clip chronologically based mostly on when it was recorded, not based mostly on if you insert it into Voice memos.

As soon as it is imported, click on the very tiny grey transcribe icon.

Wait a minute and it’ll generate a transcript.

Let’s be clear. It is a poor transcription. It received my title unsuitable, it received the product names unsuitable, and it did not have any idea of paragraphs or line breaks. It does not appear to make use of any type of customized on-device dictionary culled from the tens of millions of phrases I’ve typed on the Mac it is working on.

It is nothing like what would come from the industrial Rev.com service, however at two bucks a minute for human transcription, this little video would have value over $20. Utilizing this Apple Voice Memos hack was free (though you do get what you pay for). I am not knocking Rev.com. I exploit the service anytime that high quality is essential for shopper work.

However for my little field opening? It simply wasn’t value the price.

To get the textual content out of Voice memos, hit the Edit button and duplicate. You may want to stick it into your textual content editor of selection and put it aside for later. Take a quick second to make some edits. You may wish to search and paste in your title and product names, so a minimum of they’re appropriate within the transcript.

3. AI and YouTube Studio: ChatGPT

We will use this transcript for a number of issues on YouTube Studio. First, as soon as your video is uploaded, go into YouTube Studio and click on the Subtitles tab. In case your video has been within the system for some time, YouTube is more likely to have generated computerized captions, as proven with my video.

However above the Computerized Captions entry, there’s often a collection of three dots the place you’ll be able to enter your individual “English (video language)” transcript. Mine’s full right here, however you’d click on on the arrow and add your considerably corrected textual content file from earlier.

YouTube makes use of this to assist produce closed captions, evaluating what you add with what it creates internally. I’ve additionally heard from different YouTubers that having a full set of uploaded captions will get you a bit extra website positioning juice as a result of YouTube has extra perception into what your video is about, and the algorithm is reputed to maximise publicity based mostly on that.

Subsequent up are three easy-to-write parts of the video itemizing: the headline, the outline, and the tags. As a author, that is the simplest a part of the entire challenge for me, however as an AI researcher, this is one other alternative to see what we will get an LLM like ChatGPT to do for us.

I used to be pleasantly stunned. The AI would not instantly “watch” my video, however ChatGPT Plus did ingest my transcript. I gave it the immediate:

Learn the next after which wait for extra directions.

Then I gave it this immediate:

It is a transcript from a YouTube video. Please give me 10 high-impact attainable YouTube video titles.

It returned the next 10 video titles:

• Unboxing the Anycubic Kobra 3 Combo: 4-Shade 3D Printing Made Straightforward
• Multi-filament Magic: Contained in the Anycubic Kobra 3 Combo
• First Have a look at the Anycubic Kobra 3 Combo – 4-Shade 3D Printing Unboxed!
• Setting Up the Anycubic Kobra 3 Combo: Unboxing and Meeting Information
• Anycubic Kobra 3 Combo Unboxing – Multi-filament Energy in One Field
• Behind the Scenes of 3D Printer Unboxings: Kobra 3 Combo Version
• Final 3D Printing: Unboxing the Anycubic Kobra 3 Combo
• Multi-color 3D Printing on a Finances? Anycubic Kobra 3 Combo Overview
• Unboxing the Way forward for 3D Printing: Anycubic Kobra 3 Combo
• How the Anycubic Kobra 3 Combo Adjustments Multi-filament 3D Printing

For some purpose, ChatGPT likes capitalizing all of the phrases in these bullet objects. It is a pet peeve of mine, however since I did not inform it to not, I can dwell with the affront to my author sensibilities, for now. I used to be pretty impressed that ChatGPT picked up on the main target of my unscripted, poorly transcribed video, nevertheless it did fairly nicely.

I did, nevertheless, like one of many titles. I settled on “Unboxing the way forward for 3D printing: Anycubic Kobra 3 Combo. Then I requested it to jot down a YouTube description. That is usually one thing I would crank out, however I needed to see how the AI carried out. I gave it this immediate:

We’re utilizing “Unboxing the way forward for 3D printing: Anycubic Kobra 3 Combo.” Give me a very good YouTube description.

The primary consequence was very quick and never significantly compelling. So I advised it:

Make it longer

As with the titles, I used to be pretty impressed. Do not forget that its solely supply materials was that authentic transcript. And this is what it produced:

Welcome to my unboxing of the Anycubic Kobra 3 Combo, a revolutionary 3D printer designed to raise your multi-color printing initiatives! On this video, I am going to information you thru each step of the unboxing course of, showcasing the 4-color filament switcher and the way it comes packaged. From the preliminary setup to a take a look at the included parts, I am going to spotlight the benefit of meeting and share insights on the digicam setup I exploit for most of these movies. Whether or not you are simply diving into the world of 3D printing otherwise you’re an skilled maker, this unboxing provides you with an in-depth preview of what to anticipate from the Kobra 3 Combo.

Keep tuned for upcoming movies the place I am going to take you thru take a look at prints, superior settings, and a full assessment of the printer’s efficiency. This combo unit is a game-changer for anybody trying to streamline multi-filament prints with out sacrificing high quality or ease of use.

Do not forget to love, remark, and subscribe for extra content material on 3D printing, sensible house tech, and every little thing superior geekery!

Lastly, I requested it for tags. I am going to say right here that ChatGPT isn’t any substitute for a industrial tag evaluation device like TubeBuddy. However this set wasn’t too dangerous, though the self-referential tags with my title and my channel’s title did not purchase us something:

Anycubic Kobra 3, Kobra 3 Combo, 3D printing, multi-color 3D printing, 3D printer unboxing, 3D printing setup, 3D printer assessment, multi-filament 3D printer, Anycubic, Kobra 3 unboxing, coloration filament switcher, 3D printing tutorial, superior geekery, 3D printing know-how, DIY 3D printing, Kobra 3 setup, Anycubic Kobra 3 assessment, filament switcher, tech unboxing, 3D printer meeting, sensible house tech, maker tech, David Gewirtz

What is the backside line, Dave?

General, of the three AIs used on this challenge, I would give the next grades:

• Hohem iSteady v3 gimbal: A
• MacOS Voice Memos transcription: C
• ChatGPT: B+

All of them did their job nicely sufficient. You may discover all of them useful until you are manner past needing their assist. I’ll use the gimbal once more — that is enormous for me. If I desire a transcript, I am going to in all probability pay Rev.com if it is a high-leverage challenge. And, it is a lot simpler and sooner for me to jot down my very own titles and physique copy for a YouTube video than it’s to persuade ChatGPT of what I need.

However in case you’re not an expert writer who spews phrases onto pages as recurrently as my canine yaps at any noise he hears, instruments like ChatGPT could be very useful to get you over the hurdle of manufacturing workable supporting textual content on your YouTube posts.

This is the way it all got here collectively:

Do you employ AI assist on your YouTube movies? What AIs do you employ? Have you ever tried AI transcription utilizing Apple’s tech? How do you employ ChatGPT? Tell us within the feedback beneath.

You’ll be able to observe my day-to-day challenge updates on social media. Make sure to subscribe to my weekly replace e-newsletter, and observe me on Twitter/X at @DavidGewirtz, on Fb at Fb.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, and on YouTube at YouTube.com/DavidGewirtzTV.