Original subtitles

Da Vinci Resolve has been using AI to

transform the way that we edit. Things

that used to take experienced editors

hours can now be done within seconds

with just a simple click. So, in this

video, I'm going to show you the 10 best

AI tools that are built directly into Da

Vinci Resolve and how I use them when

I'm editing for businesses and YouTubers

who get millions of views. The first AI

tool works like magic. It's unbelievable

how good it is. Masking used to be one

of the most timeconsuming processes that

you could do in video editing. Another

word for it in the VFX world is called

rotoscoping. And they literally have

rotoscoping specialists whose job it is

just to cut out objects and people.

That's how difficult it is to do. But

now you can mask most things yourself

using Dinci Resolve's magic mask. And

it's really quick as well. So one thing

that you can use this for is to cut

people out of talking head footage. So

you can place things behind them. So for

example, I want a couple of images to

appear behind Ed. So I'll make a

duplicate of my clip by holding down the

option or alt key and dragging it up.

I'll then open it up on the Fusion page,

add the magic mask node, and then use

the onscreen picker to select what I

want to mask. So with just a couple of

clicks, I've got Ed selected. And then

I'll hit track forward and let it do its

thing. Once he's been cut out, I then

head back to the edit page. You can now

see this bottom clip is the background

and the top clip is the foreground, my

subject. I can now add my images in

between the two to make it look like

they're behind him. And I can also add a

drop shadow to my subject as well just

to sell the effect even further. This

would have taken me hours before and

well, in fact, I probably wouldn't have

even tried to mask it by hand cuz it

would have looked awful. So, there just

would have been no point. So, this tool

opens up a lot of options. The next AI

saves me hours and hours of time. If

you've ever had to edit large amounts of

interview footage into a short video,

then you'll probably know how difficult

it was when you've sent over the first

draft of the video and then you get

asked, "Oh yeah, can you just add in

that little bit where we talked about

X," for example, and then you'd have to

go back through all the hours of footage

to try and find where they spoke about

that one specific thing, and it's a bit

of a nightmare. But now you can find any

specific part of a video within seconds

using AI transcription. So let's say the

client for this video wants to include a

specific part where they talk about

wheelchair access. Instead of listening

through all of this footage, which is an

hour long, we can just open up the media

pool, right click on the timeline, come

down to AI tools, audio transcription,

and then hit transcribe. Da Vinci

Resolve will then transcribe the audio

for the whole hourong timeline. And then

when it's done, we've got the whole

transcription in this box. And you can

click on any word in this box. And

you'll see that the playhead will jump

to that point on the timeline. And even

better, to find the part where they

mention wheelchair access, I just come

to the search bar, type in wheelchair,

hit search, and it's highlighted where

he says the word wheelchair. And it's

also moved my play head to the exact

moment in the timeline. So now I can

just come down to the timeline and grab

the relevant footage for the video

wheelchair access as well. Or even

better yet, I can just highlight the

sentence that I want in the

transcription. Hit command C or control

C on Windows for copy and then paste in

that line wherever I need it. The brief

was to see 18 people comfortably,

obviously with wheelchair access as

well. This makes editing large amounts

of footage so easy. Now, and something

else you could do with the

transcription, which could be a whole

video on its own, is you can create

automatic subtitles with it. If you

click on these three little dots in the

top right of the transcription box and

then select create subtitles from audio,

you can then select how many characters

you want per line and if you want a

single line or two lines and then hit

create and it will add subtitles to your

entire timeline. These can be customized

however you like by opening the effects

panel, coming down to titles and then

subtitles. And then in here, you can

drag and drop any of these presets onto

your subtitle track like this. And then

if you click on a subtitle and come to

the inspector on the caption tab, you

can edit the lines here if there's any

mistakes. And then if you click on the

track tab, you can change the look of

your subtitles here. Also, Da Vinci now

has animated subtitles, so they can

animate word by word in lots of

different ways. Just drag and drop which

style you want onto the subtitle track,

and that's it. It will do that

animation.

>> The potential source of embarrassment

because obviously, like we've said, this

next AI tool is something that only a

couple of years ago would have been

impossible, but it's one that my clients

are always thankful I'm able to use. So,

sometimes you'll edit a video for a

client and they'll come back to you and

say, "Ah, we've said the wrong thing

there. We'll have to refilm that." And

this might mean they either have to

re-record that line themselves another

day, or if a production company has

filmed them, then it might mean having

to get someone out to then re-record

that whole line with them. It's just a

bit of a hassle. Thankfully, I've been

able to fix these situations on my own

using the AI voice convert tool. This

lets you train the AI to be able to

mimic someone's voice. So, the way that

I like to use this is so that I can

record a client's line for them and then

the AI will convert it into their voice.

So, here's how this works. You first

need a clip of the person speaking for

at least 10 minutes or a bunch of clips

of them speaking as long as there's at

least 10 minutes of audio. And then you

need to right click on those clips in

the media pool. Come down to AI tools

and then Da Vinci AI voice training.

This then asks you if you have

permission to clone this person's voice,

which hopefully you do have permission,

obviously. You then name their voice,

make sure training accuracy is on better

for the best results, and then press

start. This will probably take a while,

so it's probably best to leave this

running in the background. But once

that's done, you won't have to do it

again for that voice. So, for this

situation, Ed asked me to change one of

his lines. the original line. He says,

"This video, however, only got 2,900

views, but he now needs it to say 3,000

instead of 2,900." And he was unable to

record it himself. So, what I did was I

recorded that exact line for him. So,

here's what that sounded like. 3,000

views. I then dragged this into the

timeline, removed Ed's mistake, and then

put my line in its place. I then

rightcicked on the clip and selected

voice convert. I then selected Ed's

voice from the voice model dropdown and

then hit render. This then converted my

voice into Ed's, meaning that the line

now sounds like this. This video,

however, only got 3,000 views. For stuff

like this, it's best to have something

covering your speaker, like in this

situation where we've got this image on

top. Otherwise, most of the time it'll

be too noticeable if you've got them on

screen as well.

>> This video, however, only got 3,000

views. And if you really want, you can

actually also use this to make your

clients sing.

>> My name is Ed Lawrence and I like to

sing.

>> Moving on to the next AI tool, and this

one is pretty special. So, if you've

ever had to edit footage of people

talking in a public place or outside,

say, with a bit of wind going on, you'll

know full well that this can cause a lot

of issues in the edit because there's

just tons of background noise. Well, if

you use AI voice isolation, with

literally just a click of a button, you

can remove all of that unwanted noise.

It's insane. Just select your clip or

the audio track, open up the inspector,

come down to the audio tab, and turn on

AI voice isolation. By default, it will

be around 50, which I find normally

works quite well, but you can use this

slider to adjust the strength of it. You

basically don't want this setting to be

higher than you need it. So try and get

it as low as you can before you really

need it to kick in.

>> We came up with this idea of putting in

a a basically a video check-in wall,

which is what we've ended up doing. We

came up with this idea of putting in a a

basically a video check-in wall, which

is what we've ended up doing.

>> This is something I wish I had years ago

because it's insane how quick and how

good it is. Our next AI is also on the

audio tab. Now, when you're editing a

music track, you're able to remove

instruments or change the audio levels

of different instruments right here on

the edit page. So, for example, if you

have some music in the background of

your video, but you're finding that the

drums aren't really necessary.

>> They interface better with the with the

passengers coming in.

>> You can just turn on the AI music

remixer and click on mute drums. And now

the drums are removed whilst everything

else sounds the same. with this idea of

putting in a a basically a video

check-in wall.

>> You can do this for the vocals, bass,

guitar, and other instruments as well.

There's also slider controls for each of

these. So, you can adjust each of their

levels to what you like. So, if you

don't want to completely mute the drums,

you can just untick the mute button and

bring the slider level down a bit.

>> It's existed, but not all together. So,

we we brought together disperate systems

and have

>> and that's it. It's so quick and easy.

This next AI tool is just such a quality

of life improvement. Normally, when

you're editing a video where you've got

music in the background, it's pretty

rare that it's going to be the exact

right length that you need it to be.

Normally, it will either be too long or

too short, and then you'll have to copy

and paste it, line up the beats, and

then fade them into each other to extend

them or to bring the end of the track

further down the timeline where you need

it. But now with the AI music editor,

this can be done within seconds and with

total ease. This tool allows you to

change the music to whatever length you

need it to be. And then it will use AI

to seamlessly edit the song so that the

ending is where you want it. So, for

example, say we want this song to end

here, but it's currently too long and

ends over here. All I need to do is

simply hit the live trim checkbox, drag

the end of the track backwards, and then

AI will analyze the clip. And now look,

the ending is exactly where I want it.

You can see by these squiggly lines

where the AI has made its cuts. And it

also gives me four different versions of

it in case I don't like what it's done.

This can also be used in the same way to

extend tracks as well, which makes for a

cleaner timeline because in the past, I

would have a track copy and pasted three

or four times just to keep it going. And

while we're using AI to edit a music

track, we might as well use it to be

able to time our footage to it as well.

So, timing cuts to the beat of a song is

something that a lot of editors will

regularly do for many reasons, like

giving more energy to a video or to make

a video just flow better. And in the

past, you would have to manually use the

waveforms to find out where those beats

happened and then line that up with your

clips. But thanks to AI, this is now so

much easier. Now we just right click on

a clip and then click show music beats.

AI will then analyze the clip to find

where those beats are happening. And

then within seconds you can see that

each beat is visualized by these white

lines. So now if you have snapping

enabled on your timeline here then when

you drag clips over your music it will

just snap the start or the end of those

clips to the music beats.

And when you're done with them, you can

just turn them off by clicking on show

music beats again. It is such a handy

tool for working with music. Moving on

to our next AI tool. And this one is

another huge timesaver. Cutting A-roll

on talking head footage like this is

just a long and tedious process. And

most of your time is spent removing the

silent parts in between sentences. So in

this case, most people will move the

play head to where the waveforms are

showing that the speaker has stopped

talking. Just here. make a cut and then

move the playhead to just before where

they start speaking again. Make a cut

there and then ripple delete that clip

so there's no longer a big gap. But

doing this over and over again is just

not a good use of your time. So this is

where Da Vinci Resolve's new AI powered

ripple delete silence comes in. You

simply select your clip in the timeline

and then come up to clip then audio

operations and then ripple delete

silence. This box will then pop up with

five controls on it. The four sliders

will help you refine which parts of your

clip get removed, which are shown by

these red boxes. So, the first control

is the threshold slider. This is going

to tell Da Vinci to remove any parts of

your clip that are quieter than this

amount. So, if I set this to minus 34

dibbels, you can see that when the

person on camera is talking, their voice

is over the minus 34 dibels. So, it's

not going to delete those bits. But when

they finish speaking, the audio drops to

under that amount. So, those bits will

be removed. The next two controls are

prehead and posttail. These are

determining how many frames it's going

to leave as a buffer before and after a

cut. So, if I set the prehead to two,

it's going to make the cut two frames

before the audio comes in. And then, if

I set post tail to four, it's going to

make the cut four frames after the audio

has stopped. I find this is normally a

good setting just so it lets the end of

the sentence breathe a little bit rather

than instantly cutting away. The last

slider is minimum to strip. This is

determining the minimum amount of frames

that a section needs to be in order to

be allowed to delete it. So, if I set

this to 10 frames, then any quiet parts

that are less than 10 frames won't be

deleted. This helps to avoid having too

many cuts at unnecessary moments

mid-sentence and leaves it to just

cutting out the big gaps. Lastly, you

will see that there's a check box for

crossfade audio. What this will do is it

will add a audio crossfade between every

cut. This can be good in some situations

where there's a lot of other noises

going on and the cut needs to be

smoother. But for normal talking head

footage like this, I just leave it

unchecked. Then you hit remove. And you

can see that it's removed all of the

silent parts of the video. Meaning you

can now just go through the footage and

remove any mistakes the speaker's made.

I am so glad this is built into Resolve.

Now, now after cutting your A-roll, you

might find that you're left with some

jump cuts. This is when one clip cuts to

another clip like this. And because of

the dead space deleted, it now looks

quite abrupt when going from one clip to

the next. So this next AI tool is a

great way of fixing this. All we need to

do is add the smooth cut transition.

This transition uses AI to create new

frames that go between each cut to blend

them together so they smoothly

transition from one clip to the other.

So instead of a jarring cut like this,

clarity and purpose, which makes it feel

like they're never getting anywhere, it

gives you a nice smooth cut like this.

clarity and purpose, which makes it feel

like they're never getting anywhere,

which you can't even tell that there

even is a cut. I normally set the

duration to around 0.2 or 0.3 seconds

because I found that if you have it much

longer, then it starts going over the

end and the start of the sentences,

which can make it more noticeable. If

you find that the transition is still

covering the start of the next clip

where the speaker is talking, something

you can try is holding down the option

or the alt key on your keyboard, left

clicking on the cut point and moving it

a couple of frames to the left. If

there's enough of a gap where the voice

stops at the end of the previous clip,

then it can get the transition done

before the subject starts speaking in

the new clip. The smooth cut transition

won't work on all clips. For example, if

the head position is too different, then

it will give a weird effect like this.

Or if the speaker is using their hands a

lot, then it can also give some weird

artifacts. So, just watch out for that.

Oh, and if you're enjoying this video,

make sure to hit subscribe and give it a

thumbs up so more people can see it. So,

earlier I spoke about retiming music.

But what about retiming video? As you

might know, if you want to slow a clip

down to say 50% of its speed, you'll

need the clip to have been filmed at at

least double the frame rate of your

timeline you're working in. Because if

it's not, then it will only be able to

show one frame for every two frames in

the timeline, which results in slow

motion that looks choppy like this. But

now, thanks to Da Vinci's speed warp AI,

this isn't a problem anymore. After

you've slowed down your clip, you can

just open the inspector, come down to

rettime and scaling, and then set the

rettime process to optical flow and the

motion estimation to AI speed warp.

You'll see a faster option and a better

option. The better option is slower to

render, but should normally give you a

better result than the faster option.

Although, there's been a couple of times

where I prefer the look of the faster

option. So, just have a play with both.

And now you can see that the clip plays

back in super slow motion. This is

something that wouldn't have even been

possible not long ago. So, it's so good

to have this built directly into the

program. These AI tools have really sped

up my editing and make my life so much

easier. But there was one other major

change that I had to make to save even

more hours editing. So, watch this video

to find out what that big change was

that dramatically accelerated my editing

speed beyond what I thought was even

possible.