sauron, I’m having trouble replicating your process. I have to quit now for the evening but it looks like the top and bottom tracks description are reversed. Track V1 would be on the bottom (which you have labeled 16:9) and it should have the Rotate Scale = 135.0 degrees, isn’t that right. Will check in tomorrow.
-=Ken=-
I also see the use of this “effect” when TV stations use 4:3 archive footage and need to present it with 16:9 aspect ratio. The alternative would be show the black vertical bars or zoom in to fill the screen. From all this options I don’t know which is worse when trying to fill the screen. 
This is true, but my take on the OP was that ‘portrait mode’ referred to the proclivity of smartphone users shooting everything in portrait mode.
Yes that is correct. Track V1 (16:9) is the bottom track and track V3 (4:3) is the top track. Color in the middle. My bad.
It is indeed.
I changed the topic heading to make it more useful.
I went to change my project and found sauron had already updated his post. When corrected my project worked properly - thank you!
As Steve said the Blur filter has a problem if you set a high blur - the edges seem to fold in. It is interesting that the Glow filter has a Blur slider and its blur does not have edge problems. And combining them seems to mitigate the problem somewhat.
Fiddling with the filter settings - it is difficult for me as a casual (and older) user to remember all the filter functions. Glow has a blur slider. Rotate has Scale slider. I get confused with this. And once one of you tells me what obscure feature is hidden in a filter with a name that claims to do something else - will I remember it two years from now when I need it again? 
I’m not sure how much further to take this topic now. I could futz with it for days I guess. I like the “picture-in-picture” effect without the colored frame. But then, the visual impression in my monitor is that the track1 sits below track3 - and I would prefer the opposite. See what I mean? I could spend hours on this.
BTW, is everyone agreed that this is called a “Pull Focus” effect? -=Ken=-
That is why I added the Glow filter. The Blur filter did not look good, maybe because of the clip I used, so I added a small Glow to add a little enhancement to the blur.
Pull Focus it is.
I tried a little experiment. Disabled the Glow filter and added a Color Grading filter. You can get some interesting results when you fiddle with the sliders not the color wheels. You can fine tune the blur.
Well no, not really. I can’t agree that the purpose of the ‘effect’ is to pull the viewer’s focus to the central overlaid video [as expressed by the author at the link you posted], an empty black background will to that much better, out-of-focus background is actually a distraction as it pulls your focus away from the central content.
More often than not it is simply expressed as ‘filling the sides of vertical (or portrait) video’. Or ‘Adding background blur to portrait video’.
But at least I will know what you mean if you use the term again.
Heh, heh. If you will indulge me further I will keep going.
I have been combing through videos and found this one using Vegas.
At the 3:30 mark he adds a “2D Shadow” or drop-shadow effect that gives the foreground video a depth as though it is floating above the background. Is there anything like that in Shotcut?
-=Ken=-
I guess you could add a black rectangle to a track between the BG and FG then blur it.
Ken, I had another thought about the problem of having someone else editing a slideshow of pictures, and then adding the Pull Focus effect.
If someone else (or you!) put together a Shotcut project with a series of images (some or all of which were in Portrait orientation), with or without transitions, you can save that MLT file… and then start a new project, create two tracks (or three if you want a slick mask like Sauron’s example), and “Open MLT as Clip” to add the MLT to both tracks, then apply:
Here’s the Shotcut version of “How to make a portrait video look better”.
Used the same technique that I used for the “Pull Focus” except I had to cheat a little. Couldn’t figure put how to stretch the portrait video for the background in Shotcut so had to Avidemux to resize the video.
This is something that might interest @Steve_Ledger. If you make the display area larger and offset it the Blur filter will not have fuzzy edges.
Yes that is a work-around.
Though I have found my PC chokes when I distort/stretch and blur ‘video’ in Shotcut.
Anyone here know of a way to stretch the portrait video to fill a 16:9 screen?
Hey! Success! I tried John’s method adding photos, saving the file and opening the mlt as a clip. Size and Position Distort fills the screen and grabbing the “Handles” I could stretch it even further so the Blur effect errors are off screen. Opacity didn’t work for me but the Brightness filter did. I haven’t tried this with cross-dissolves yet but intend to. Thanks John! This may be the way to go since using this method the foreground clips now seem to float on top of the background without even having to do something more.
sauron, your demo video is real action! And the video floats! I think you can just use Shotcut to stretch the background so the Blur problem is masked. Grab the center dot and drag it off center temporarily so you can use the Handles, stretch it some and move it back in the center. I used Blur Width and Height of 25 px.
Nice going guys!
-=Ken=-
I know this method but it won’t work with the portrait video I used that I found here.
https://www.youtube.com/watch?v=ikO91fQBsTQ
Really? I don’t know why that would be so. I am working with SD NTSC 720x480 but Shotcut will handle 4K video. 
I tried a new project with cross-dissolves and Shotcut will not pass the lower layer through. I may have to be content with back to back dissolves.
Edit: I tried using Alpha Channel: Adjust and it just didn’t work right. So - back-to-back dissolves it is.
-=Ken=-
Using the Size and position filter I run out of screen space and what I am able to fit in the screen does not look good. Tried every filter I could think of but finally had to resize the video with Avidemux to get the Sony Vegas effect that you posted.
Oh. It must be because you are using a bigger pixel space. You just don’t have room on your video monitor to stretch it out!
I tried using Fade Video Out and Fade Video In and I get the same problem with the background track not working right. This stuff is starting to drive me nuts! Perhaps another editor would not have so much trouble. Keep it simple and Shotcut is fast and reliable.
So, I think I am done. Thank you all for your participation. It has been fun!
-=Ken=-
For some reason I can’t wrap my head around what this means unless you’re 1) only using the handles to resize instead of typing in values, or 2) not using the “distort” setting on the S&P filter. At first I thought it might have something to do with the filter working differently on images vs videos, but I tried it out on a video (took me a while to find a portrait video in my archive!) and it worked the same (i can post a screenshot if you want).
I did notice that when I tried mixing filters together, including trying to use the Rotate filter and S&P, weird things happened, especially with the Blur effect. I found that applying the Blur effect first was important, so that its edges could be moved offscreen, and that the resolution of the image needed to be high enough that it wasn’t excessively pixelated when the blurred image was enlarged.
If we’ve been having the same issue, I haven’t been able to re-create this every time. In one of my projects I find that as long as I leave specific tracks empty, such as V1, I can crossfade in other tracks. But as soon as something is in V1, the crossfade fails. In my previous dual-MLT example, as the screenshot shows (barely), V2 and V1 are both displaying MLTs that are cross-dissolving. But when I just tried it, in a different project, with two different images, neither separate cross-dissolves nor using dual MLTs let me see “below” the top cross-dissolve. IN THEORY, though I haven’t tested this, if you can cross-dissolve on a single track, and if compositing is working for you when not cross-dissolving, you should still be able to export the entire thing as a video, and then overlap the videos for the pull-focus effect without needing to do any cross-dissolves in the pull-focused project.