Shotcut tutorial · companion page
A real tilt-shift needs a tilt-shift lens. This fakes it: a second copy of the clip, an elliptical mask flipped to Subtract, and a Gaussian blur heavy enough that a city junction starts reading as a model railway.
Reading in another language? Use the translate picker in the header, or your browser’s own translate command — the Shotcut filter, control and value names are marked to stay in English so they still match what is on your screen.
Two filters on one duplicated clip. The mask decides where stays sharp; the blur decides how hard the rest falls away. Everything else in the video is scaffolding you throw away before export.
The illusion starts with perspective, not filters. A steep downward angle is how we look at models, so the brain is already half convinced before any blur lands. Ben says it plainly: imagine being a giant looking down.
The top copy gets an elliptical hole punched through it with Mask: Simple Shape on Subtract. The untouched clip below shows through the hole.
Miniature photography has almost no depth of field. That is why the blur here runs at 44.8 rather than the gentle edge softening used for a lens vignette — roughly three times as strong.
With two identical copies stacked, the mask is invisible while you shape it. A black colour clip in the middle makes the boundary obvious. It is a teaching aid; hide it before you export.
The Netflix vintage lens companion builds a soft-edged frame from the same two filters, in the same relationship. It is not a different technique. What separates a lens vignette from a miniature is entirely in the numbers.
| What | Vintage lens page | This page |
|---|---|---|
| Filters used | Blur: Gaussian + Mask: Simple Shape | The identical pair |
| Order in the filter list | Blur first, mask second | Mask first, blur second |
| Operation | Subtract | Subtract |
| Blur strength | 15.2 % | 44.8 |
| Sharp zone | Nearly the whole frame — a rim only | A band 41.55% wide by 14.61% tall |
| Sharp zone placement | Centred | Pushed low — 81.74% down the frame |
| Softness | 25.78 % | 45.21 % |
| Guide layer | Green colour clip, #55aa00 | Black colour clip, #FF000000 |
| Saturation or contrast push | None | None — see the note below |
| Recorded on | Shotcut 26.4.30 | Shotcut 20.09.27 |
Two things are worth flagging. First, it is not literally the same panel: the 2020 build used here expresses the ellipse as Horizontal / Vertical / Width / Height percentages, while the newer build on the vintage lens page uses Position and Size in pixels. Same filter, a different generation of the interface — the values compare in proportion, not field for field. Second, most miniature-faking recipes push saturation and contrast as well, on the theory that toys are made of bright plastic. This tutorial does neither. The clip carries exactly two filters from start to finish and no colour filter is ever opened. If your result looks flat next to the miniature photographs you have seen, that is the missing ingredient — and it would be your addition, not Ben’s.
[ that’s not how our eyeballs see toys ]
Ben Espanto · 8:57
Click any screencap or timestamp to play that chapter from its exact second, right here in the card. Quotes are transcribed verbatim from the audio. Five of these boundaries are YouTube’s own auto-generated chapter marks, kept at their original seconds but retitled; the rest are split where the screen actually changes and are marked derived from the frames.
“Of course this isn’t the real tilt-shift effect, because in order to do that you would need to do that in camera with a tilt-shift lens. But you can fake that effect in post.”
No interface on screen yet — this is the finished result playing full frame. A narrow band of the road stays sharp, everything above and below it is thrown out of focus, and real traffic starts reading as a model railway.
“This tutorial is done on Shotcut version 20.09.27.”
The About Shotcut dialog at 1:20 confirms the build: version 20.09.27. This is a 2020 release, and the Mask: Simple Shape panel it draws is laid out differently from current Shotcut — see the note under the settings table.
“If you’re filming your own file you need to make sure that you are looking down on your subject. Imagine you being a giant looking down on the cars and people below.”
The stock clip open in the player at 2:02 — a high, steep view down onto a junction. Ben picks the traffic in the lower middle of the frame as the thing to keep sharp. The perspective does half the work before a single filter is added.
“I’m also going to make a copy of this particular clip by clicking control C. I’m going to right click this and I’m going to insert another track right on top of it.”
At 2:32 the clip has just been dragged into the timeline on its own track. The copy goes on a new track directly above, pasted so the two start on the same frame.
“All I’m going to try to do is maintain the sharpness of my subject, which is this car for example, and then I’m going to try to blur everything else around it. And the way I’m going to do that is I’m going to mask the top video and basically punch a hole through it.”
Two identical copies stacked, aligned at zero. Nothing is masked or blurred yet — the preview looks exactly like the source, because it is.
“Now mind you this extra step has nothing to do with this technique. All it’s going to do is help me show you what I’m trying to do.”
A third track goes in, the video copy moves up to it, and Open Other → Color drops a colour clip into the middle track. The dialog at 4:06 still reads transparent — the reason for the aside in the audio. He goes on to set it to black.
“So while we’re on this top layer… we’re going to choose mask. And in this instance I’m going to choose mask simple shape.”
Typing mask into the filter search returns four entries: 360: Equirectangular Mask, Mask: Apply, Mask: From File and Mask: Simple Shape. The last one is the one to take.
“So in this particular mask we want the ellipse shape. And then I’m going to increase the width of it… and then I’m going to move it down and I’m going to find that car that we want to focus on.”
At 5:52 the shape is Ellipse, Width is opened to 41.55% and the band is dragged down the frame. The black colour clip underneath is what makes the shape visible — everything outside the ellipse is showing black, not blur.
“Because if this black layer did not exist, if I made it invisible, there’s no way for you to tell where the mask is. Because I have two identical videos on top of each other.”
The same moment with the black layer switched off, at 6:10. The mask is still there and doing exactly the same thing — but with two identical copies stacked, cutting a hole in the top one changes nothing you can see. That is the whole argument for the black clip.
“I’m then going to choose another filter, click the plus button, and we’re going to choose Gaussian blur. Now what’s happened… if I increase the blur, what’s happening is it’s blurring the ellipse in the middle.”
Second filter stacked under the mask. Amount starts at 20.0. Note what the preview shows at 6:55: the blur is trapped inside the ellipse, which is backwards. The mask is still on Overwrite.
“But in fact what I want it to do is I want it to blur everything else except for this middle part. So we actually want to go back to the mask simple shape, and instead of this operation set to overwrite, we want to go to subtract.”
One dropdown. At 7:32 the ellipse becomes a hole instead of a window: the black shows through the middle and the untouched footage surrounds it. The blur is sitting at 0.0 at this instant, which is why nothing looks soft yet.
“And so now that the selection has been created with the help of that black background, I don’t need that black background anymore. So I can actually just make that invisible by clicking this eyeball here.”
The Hide toggle on the middle track is clicked off at 7:58 and the sharp original on the bottom track shows through the hole. Blur is still 0.0 here, so the frame is entirely sharp — the composite is correct, the effect just has not been dialled in.
“As you can see in my image I’m blurring everything else except for that oval shape in the middle… the edge of the oval is a little too harsh. So we actually want to go back to the mask simple shape and go into the softness setting.”
At 8:20 the blur is up at 67.2 and the miniature illusion arrives. He then pushes it as far as 92.2, decides the oval rim is too abrupt, opens Softness from 20.00% to 45.21%, and finally pulls the blur back down. Full path in the tuning table below.
“But if I add that additional layer on top of it, this is where that toy… it actually helps that my computer is stuttering a little bit. So it creates kind of that herky jerky movement, when in actuality it should actually be playing in full speed.”
The top layer back on, playing at 9:52. The stutter Ben mentions is his machine struggling with the preview, not the effect — and he points out it accidentally flatters the look, because real stop-motion miniatures move in jerks.
“So let me just set the preview scaling to 360 so that you can see it in full speed.”
The status button beside Source / Project now reads Preview scaling is ON at 360p. This is a playback-performance setting: it lowers the resolution Shotcut renders into the preview window so playback keeps up. It changes nothing about the filters and nothing about the exported file. Do not carry it into the recipe.
“Now let me render this so that you see what this whole thing looks like after it’s been exported. So there it is, creating the tilt-shift effect in post editing using Shotcut.”
The exported file playing full frame at 10:50. Chapters stop at 10:11, but the video runs to 11:16: the render, the result, and the closing subscribe card with the bentacular handle.
Every number below was read off the screen in the video. Nothing here is a recommended-sounding guess — if a value is not stated, it is because it was never shown.
This is the one decision no filter can rescue. A high, steep angle onto small moving things — traffic, crowds, boats — is what makes the brain read a model. Ben uses a stock clip of a city junction filmed from well above street level.
Full source properties: H.264 / AVC / MPEG-4 AVC, yuv420p, aspect ratio 16:9, scan mode Progressive, colour range Broadcast Limited (MPEG). Speed left at 1.000000 x.
Drag the clip in. Select it, copy with Ctrl+C, right-click the track head and insert a video track above, then paste so both copies begin on the same frame.
If the two copies are offset by even one frame the sharp band will not line up with the blurred surround, and the boundary will smear as the shot moves.
Move the top video copy up to the new top track, then on the middle track use Open Other → Color and set the colour to black. The Select Color dialog reads #000000 with Alpha channel 255; the timeline labels the clip #FF000000.
Stretch the colour clip to the full length of the video clips. V3 keeps its default Blend mode of Over; nothing on this page changes a blend mode.
Select the clip on the top track, open Filters, press + and type mask. Four filters come back; take Mask: Simple Shape — not Mask: Apply, not Mask: From File, not 360: Equirectangular Mask.
It arrives on defaults: Operation Overwrite, Shape Rectangle, Horizontal and Vertical both 50.00%, Width and Height both 10.00%, Rotation 0.0 deg, Softness 20.00%.
Open Width until the ellipse spans most of the frame horizontally, then drag Vertical down until the sharp band sits over whatever you chose to keep in focus — in the video, the traffic in the lower middle. Height controls how deep that band is.
Vertical at 81.74% is the miniature-specific choice. A centred band gives you a lens effect; a band low in the frame gives you a viewer leaning over a table.
With the same clip still selected, press + again and add Blur: Gaussian. It has one control, Amount, and it opens at 20.0.
This looks wrong because it is wrong, and that is the teaching moment. The mask is still on Overwrite, so the blur is confined to the shape rather than excluded from it.
Go back to Mask: Simple Shape and change Operation from Overwrite to Subtract. The ellipse stops being a window and becomes a hole: everything outside it is what the blur now owns, and the sharp original underneath shows through the middle.
This one dropdown is the effect. Everything before it is setup and everything after it is taste.
Click the eye on the middle track to hide the colour clip. Now judge the effect on actual pixels: raise Amount until the surround reads as macro-lens shallow, then open Softness so the band does not have a visible rim. The video settles on 44.8 blur and 45.21% softness.
Hide or delete the colour clip before you export. It is opaque black; leave it visible and it will render.
At 10:11 Ben sets preview scaling to 360p. This is worth separating out, because it looks like part of the recipe and is not.
Two heavy filters on a stacked composite are expensive to preview, and Ben’s machine was stuttering. Preview scaling lowers the resolution Shotcut renders into the preview window only so playback keeps up. It does not touch the filters, does not touch the project, and does not touch the exported file. Turn it on freely while you work; it will not change your output either way.
Read directly off the frames of the video. Where a value changed on screen, the tuning path is shown separately below.
| Setting | Value | Where it lives |
|---|---|---|
| Shotcut version | 20.09.27 | Help → About Shotcut |
| Source clip | video01.mp4 | Playlist |
| Source resolution | 1920 × 1080 | Properties → Video |
| Source frame rate | 23.976024 | Properties → Video |
| Source codec | H.264 / AVC / MPEG-4 AVC | Properties → Video |
| Source format | yuv420p, 16:9, Progressive | Properties → Video |
| Colour range | Broadcast Limited (MPEG) | Properties → Video |
| Clip duration | 00:00:05:23 (timeline 00:00:05:22) | Properties |
| Track stack, final | V3 masked + blurred copy / V2 black colour clip / V1 original | Timeline |
| Blend mode | Over (default, unchanged) | V3 properties |
| Guide layer colour | #000000, Alpha channel 255 · labelled #FF000000 | Color clip on V2 |
| Guide layer at export | Hidden — eye toggled off at 7:53 | Timeline track head |
| Filter 1 | Mask: Simple Shape | Top copy only, first in the list |
| Filter 2 | Blur: Gaussian | Same clip, stacked after the mask |
| Shape | Ellipse | Mask: Simple Shape |
| Operation | Overwrite → Subtract | Mask: Simple Shape |
| Horizontal, final | 50.00 % | Mask: Simple Shape |
| Vertical, final | 81.74 % | Mask: Simple Shape |
| Width, final | 41.55 % | Mask: Simple Shape |
| Height, final | 14.61 % | Mask: Simple Shape |
| Rotation | 0.0 deg | Mask: Simple Shape |
| Softness, final | 45.21 % | Mask: Simple Shape |
| Amount, final | 44.8 | Blur: Gaussian |
| Saturation / contrast filter | None — never added | — |
| Preview scaling | ON at 360p | Player status bar — playback only |
The Mask: Simple Shape panel in Shotcut 20.09.27 expresses the ellipse as four percentages — Horizontal, Vertical, Width, Height. Newer builds replace these with Position and Size in pixels. If your copy shows pixel fields, translate by proportion: Width 41.55% means the ellipse is about 42 per cent of the frame’s width.
| At | What changed | Operation | Shape | Horizontal | Vertical | Width | Height | Softness | Amount |
|---|---|---|---|---|---|---|---|---|---|
| 5:00 | Mask filter added, defaults | Overwrite | Rectangle | 50.00 % | 50.00 % | 10.00 % | 10.00 % | 20.00 % | — |
| 5:20 | Shape switched, width opened | Overwrite | Ellipse | 50.00 % | 50.00 % | 41.55 % | 10.00 % | 20.00 % | — |
| 5:50 | Band dragged down over the traffic | Overwrite | Ellipse | 50.00 % | 81.74 % | 41.55 % | 15.98 % | 20.00 % | — |
| 6:30 | Height trimmed — geometry final | Overwrite | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 20.00 % | — |
| 6:50 | Blur filter added | Overwrite | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 20.00 % | 20.0 |
| 7:00 | Blur pushed to show the problem | Overwrite | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 20.00 % | 26.7 |
| 7:10 | Blur zeroed before flipping the mask | Overwrite | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 20.00 % | 0.0 |
| 7:30 | Operation flipped — the effect | Subtract | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 20.00 % | 0.0 |
| 8:10 | Black layer hidden, blur raised | Subtract | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 20.00 % | 71.1 |
| 8:20 | Settling | Subtract | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 20.00 % | 67.2 |
| 8:30 | Driven to the top to expose the hard rim | Subtract | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 20.00 % | 92.2 |
| 8:40 | Softness opened to smooth the rim | Subtract | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 45.21 % | 92.2 |
| 9:00 | Blur dropped back | Subtract | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 45.21 % | 53.0 |
| 9:10 | Final | Subtract | Ellipse | 50.00 % | 81.74 % | 41.55 % | 14.61 % | 45.21 % | 44.8 |
The two dials worth playing with are Amount and Softness. Ben’s own reasoning for pulling the blur back down from 92.2 is on the record: “that’s not how our eyeballs see toys”. Too much blur reads as a lens fault; the miniature illusion wants shallow, not smeared.
Machine transcribed from the video audio. Spoken phrasing is left as-is; only the product and filter names Whisper mangled have been corrected. Every timestamp plays the video in the hero from that second.
In this tutorial I’m going to show you how to replicate the miniature tilt shift effect in post using Shotcut.
Of course this isn’t the real tilt shift effect because in order to do that you would need to do that in camera with a tilt shift lens.
But you can fake that effect in post using Shotcut and I’m going to show you how.
So stay tuned.
I’m just a normal person with no video editing background.
I wanted to start making YouTube videos and maybe cool transitions and effects.
I don’t really plan on being a professional video editor so I was looking for a free easy to learn video editing software.
Luckily I stumbled on Shotcut.
A free open source video editing program that can do many of the tricks you can do on more enterprise video editors like Adobe Premiere or DaVinci Resolve.
But with a much simpler and leaner interface thus dramatically shortening the learning curve.
It just takes using your imagination.
So let’s learn together.
When I was practicing this effect I was actually surprised how easy this was to accomplish just by using a few steps.
And I’m going to be showing you this today.
This tutorial is done on Shotcut version 20.09.27.
If you’re filming your own file you need to make sure that you are looking down on your subject.
This helps minimize your subject and helps with the miniaturized perspective.
Imagine you being a giant looking down on the cars and people below.
That’s the perspective your video needs to have.
If you’re using recorded videos already that’s the kind of perspective it needs.
I happen to be using stock video so that’s the kind of video I was looking for and I found this one here.
If you notice the perspective even before I do anything to it it already somewhat looks miniaturized because I’m looking down on the subject.
And in this instance I think I’m going to be focusing on this thing here.
This area around the middle bottom part.
That’s the part that I’m going to make look like toys or miniaturize.
So with that video chosen I’m going to drag it right into the timeline.
Like so.
I’m also going to make a copy of this particular clip by clicking control C.
I’m going to right click this and I’m going to insert another track right on top of it.
And I’m going to paste an identical copy of it right on top.
So that they’re exactly on top of each other.
All I’m going to try to do is maintain the sharpness of my subject which is this car for example.
And then I’m going to try to blur everything else around it.
And the way I’m going to do that is I’m going to mask the top video and basically punch a hole through it.
And this is how I’m going to do it.
In order for me to illustrate this technique I’m going to have to add an extra step.
Now mind you this extra step has nothing to do with this technique.
All it’s going to do is help me show you what I’m trying to do.
So first and foremost what I’m going to do is as you can see I’ve added a third track on top of the previous track that I’ve created before.
And let’s zoom in here.
And what I’m going to do is I’m going to choose this video here and I’m going to move it up to that top track.
And then in this middle track I’m going to click open other right there.
I’m going to choose color.
And in this color box I’m going to choose the color black.
Like so.
I don’t know why that keeps choosing transparent but yeah.
So that’s black.
And then I’m going to drag it into here to fully cover that bottom part.
Right there.
Stretch it all the way to the end.
And so now the order is the original video and then if I make that layer invisible it’s then a black layer and then underneath it is a copy of the video.
And I’m going to show you why I did that.
Okay so while we’re on this top layer what I’m going to need to do is we’re going to need to go to filters here.
Like that and we’re going to choose mask.
And in this instance I’m going to choose Mask: Simple Shape.
So in this particular mask we want the ellipse shape.
And then I’m going to increase the width of it.
Like that.
And then I’m going to move it down and I’m going to find that car that we want to focus on.
Right there.
And so that’s going to be the center of our focal point.
I’m going to change the height a little bit so that it can kind of focus on more things maybe.
Right there.
And so now I’m going to show you why I chose a black background in this middle point here.
Because if this black layer did not exist, if I made it invisible there’s no way for you to tell where the mask is.
Because I have two identical videos on top of each other.
And so you’re not going to see what I’m doing but if I include that black mask or the black background in the middle you know exactly what I’m trying to do.
Okay.
So with that chosen I’m going to go back to the original video on the top layer which is on v3.
I’m then going to choose another filter, click the plus button and we’re going to choose Gaussian blur.
Now what’s happened, so if I increase the blur you’re seeing that what’s happening is it’s blurring the ellipse in the middle.
But in fact what I want it to do is I want it to blur everything else except for this middle part.
So we actually want to go back to the Mask: Simple Shape and instead of this operation set to overwrite we want to go to subtract.
Right.
And so now what it’s doing is it is showing everything else except for the part that I want to focus on.
And so now that the selection has been created with the help of that black background I don’t need that black background anymore.
So I can actually just make that invisible by clicking this eyeball here.
And then with that selected knowing that everything else is chosen I can go back to the Gaussian blur here.
And then I can then increase the blur.
As you can see in my image I’m blurring everything else except for that oval shape in the middle.
Now what we notice here if I drive it all the way to here what I notice is the edge of the oval is a little too harsh.
And so what we want to do is we actually want to go back to the Mask: Simple Shape and go into the softness setting.
And we actually want to make it even softer so that the oval looks a little bit more gradual.
And then let’s go back to the Gaussian blur.
And we want to drop that blur by a lot because this is that’s not how our eyeballs see toys.
So this is probably about the right setting which is it looks like 44.8.
And so let’s go ahead and play that and see what that looks like.
And that’s basically it really believe it or not this is the whole this is the whole technique.
So if I uncheck that top layer and I play this video the original video that’s what it looks like.
But if I add that additional layer on top of it this is where that toy it actually helps that my computer is stuttering a little bit.
So it creates kind of that herky jerky movement when in actuality it should actually be playing in full speed.
So let me just set the preview scaling to 360 so that you can see it in full speed.
And so that right there is the tilt shift technique replicated in post editing using Shotcut.
Now let me render this so that you see what this whole thing looks like after it’s been exported.
So there it is creating the tilt shift effect in post editing using Shotcut.
Pretty easy right?
So once again thank you for coming to my channel watching my tutorial.
And once again last but not least just asking you guys if you like what you saw please click the subscribe button.
I appreciate your subscription.
I’ll see you in the next tutorial.
