Published: Jul 8, 2025 by Isaac Johnson

Today I wanted to dig in to Vertex AI Media Studio as well as Google AIStudio. They are very similar though the prior is pay-as-you-go and the latter has some free credits.

We’ll start with some simple animations of a logo. Then we will try and build out an intro using Video (Veo), Music (Lyria) and Voice (Chirp) (with some help from Midjourney).

I’m also going to explore what we can do for free with other AI video tools like Suno and ElevenLabs.io.

I’ll put two examples to together by the end and we can compare both the outputs and costs associated with them.

Vertex AI Media Studio

Let’s first look at taking the image we used with the PyBsPoster app and turning it into a video.

Here we can see the final render with the extended time

I can also take the final logo we used and feed that into Veo2 using the Vertex AI Studio’s Media Studio

which looks like

I like this one a lot better as the trunk is smooth, but I will admit the feet movements are less than desirable.

If all we are doing is playing, there are so many examples already. Here is just a sample from this morning on MJ

I began to think about videos I might use for creating a tutorial.

Perhaps a nice “fresh brewed” logo we zoom in on

or a laptop that also brews coffee

I then thought, what about a nice NPR style radio intro with video.

A smooth NPR radio introduction for a segment called “Fresh Brewed” with a light jazzing background. The video shows someone pouring a nice cup of dark coffee into a red cup

I got a couple good examples, but one of them you can see the person bare handing the carafe which would be a bit odd

There are no words, but there is a bit of relaxing jazz music

and the other

I can opt to extend the video, but doing so needs me to pick a bucket to store the output in

The four example 15s videos are interesting, but the mangled text would make them not usabe as they stand

Also, they lost the music in extending

I had another idea - what if we fed it one of the last frames of the prior video and went for a prompt from there

The results again, did not have sound, but only one I could see being usable.

Let’s review

The first and third videos has a magical espresso machine appear right after pouring coffee - that would make no sense. The second one moves to an expresso machine to perhaps top off with froth - i could see that perhaps. And the last one is my favourite. The bottom right video has a hand go to grab the cup but actually magically removes the handle in doing so!

I think we can keep that second video

I wanted a bit of tunes to go with these, so I used Lyria 2 to generate some backing tracks

Lastly, I used Chirp 3 to create some voice intros. I found that each run (clicking the play button) generated a slightly different inflection. So I would run it till it sounded right to my ears.

Brining together

I want to try a basic intro using some of the videos, sounds and intro audio files we saw.

Judge me all you want - I have been using PowerDirector for ages using my phone and I don’t plan to stop now.

Though I will use Windows Phone Link to take some screenshots here.

I’ll create a new project

From there, I often use Google Drive as a go-between

I upload all the generated assets into a temp folder from my desktop

Then I can pull them in and add them to a timeline

I tried to use a cam shot with a noisy background but found my talking intro really had to zoom into my head

I reshot it with a simple backdrop

which was a lot easier to chroma key out using the standard background remover

Bringing it altogether with some background music and voice overs looks a bit like this. I’m not thrilled, mind you, but it’s a pretty okay start

Will I make a Google CLI code video? I don’t really know. I’m not a tremendous fan of making video content, at least for blogging.

Costs

That was fun to make in Veo, but let’s really look at some costs. I had some credits mind you, but had I not, that would have been nearly US$60 to make

This is where the US$10/basic MJ plan might work in my favour

Cheaper

So let’s built out an option that should run me less than $60.

I did a video of a coffee pour in MJ and extended it

This is the one i liked best

And for Music, I could just use Suno

I laughed when I logged in and saw a 90s song I created (with the lyrics) when I was stuck by a Cloud Team that wouldn’t give me a project in GCP which was blocking our DR initiative.

On my ask now though, it missed the “hip hop”, however, the jazzy output was fine and i wouldn’t mind using it.

Additionally, I was just raw paying via the Vertex AI media studio.

One can get some free credits using the AIStudio instead

While it made videos, it didn’t provide audio

Perhaps a trial of Artlist would cut it

And a reason to really look. I typoed FSB and FBS, arg. So there went a credit.

They want me to pay just to download an MP3 which seems silly.

Windows key G and we have game capture…

But the one that blew me away was ElevenLabs.io

What is great is that by making just subtle tweaks, like capitalizing a letter or changing “FBS” to “F.B.S.”, I can get some really subtle but interesting variations on a theme

Run 1:

Run 2:

Run 3:

and I loved the outro

With such a voice, I had to do a new video. I ended up using Vidu to make some with free credits

I reworked it in PowerDirector a few times and managed to come up with this more country styled intro

Summary

Hopefully you saw some good things from commercial tooling like Midjourney (MJ) and Google Veo2 using the Vertex AI media studio. Additionally, you can get some free video generations using the AIStudio instead of the Vertext AI Media Studio.

I like how the Media Studio has all of our tools front and center - very easy to use.

I’m not as much a fan of how the costs aren’t clear and I blew through about US$60 in an hour. But maybe I should think of it as Pinz/Dave n Busters. I’m going to have some fun, come away with some silly stuff and blow a whole lotta money.

You be the judge… here is the expensive

And the one cobbled together from free or cheap solutions

I mentioned the spend on GCP (just over $60 when all totaled). I kept things in my basic US$10 MidJourney plan for video. I could move up to the $30/mo plan for more hours but since doing videos is not my main thing, I’m okay with basically blowing half my months credit on this.

Once thing about MJ I don’t see in Vidu or others is that I can buy more hours if I use my time up.

Which seems a much smarter option if doing videos just periodically.