How do I fit a whole animated frame.

[Oldspeak]

Think I might be asking something extremely complicated (for me the very least ) So I have this video clip that's about 10 seconds. It basically repeats in the same motion but the frame or the camera moves upward. I want to be able to include the whole frame from the start to the end in one image without the camera moving away. Imagine like a looping gif. but it pans from the bottom to the top. Any tips be a great of help

[zoobie]

Try using a free editor here or there. You'll probably need to zoom in a bit as you're using motion tracking. This way, it will appear that the camera never moved.

[Oldspeak]

I've looked around on "motion tracking" but it doesn't seem to be the thing I need. Let me make another example: a ball bouncing video. It starts from the bottom where the ball hits the ground and it slowly pans up to where the ball reaches it maximum height. How can I edit this to include the whole scene of the ball bouncing from the floor to it's maximum height in one clip. It's a perfectly looping video, just the "camera" moving upwards. Going to play around more when I get back home.

(I'm currently using my friend's Sony Vegas Pro 11)

[poisondeathray]

Do you mean extending the field of view , making it wider? e.g. as if you shot the piece zoomed out in the first place? If so, it's normally not possible with standard tools

You normally cannot combine temporally different information (that's what frames are, they represent different moments in time) to extend the field of view with foreground content like a bouncing ball , unless you're going for a blurred time lapse sort of effect . But you can use motion tracking and matte painting techniques to extend static background content

Is this an animated piece that you made , or is it live action footage ?

Can you provide a better description of your video and goals? maybe even a sample video? - in some cases you can roto out the foreground to composite it with the extended background - but this is getting more complex and requires more than vegas or video editors. It really depends on how the shot was set up and the little details

[Oldspeak]

That's what I'm afraid of but it's not entirely that. The whole animation is the same but with just the camera moving around. There is no zooming or extending the field, just a simple point of view moving upward. Alright, let's put my describing skills to the test.

Here's a "video" sample starting from a scene from bottom to up in numerical order.
-
3
-
2
-
1
-
So what I'm trying to get out of this is to get rid of the camera panning and just have the whole video clip itself:
-
3
2
1
-

This sound impossible as it involves editing the video clip itself but just wondering if it can be done even if it involves the most scrutiny way, I would give it a try. Such as overlapping each scene (since it's a perfectly looping video). It has close to 400 frames, hopefully I can find a easier way.

[poisondeathray]

Ok, but does each frame contain all the data of the foreground object? ie. does frame 1,2,3 -400 all have the object in view? And what perspective is the view ? Are you viewing the object at a slightly different angle? (For example if a ball bounces high, I will be viewing the bottom of the ball, but at eye level, it will look straight on)

Do you just want to stablize the up panning motion of the camera? (you want to elminate it? )

What is the camera motion exactly - What type of camera is it ? - is it free(handheld) or tripod fixed ? . Is it a true pan from a fixed point, or is there Y-translation in the camera movment? For example - If it's only a vertical pan from a tripod, you cannot combine frames, because each frame has the object viewed from a different angle unless the object travels very short distance - does that make sense ? And what about the background? Is there a background at all? These types of details affect how you approach the problem.

Sorry, a more detailed description is required

If it's an animation you've done, why don't you just redo the animation or the camera move?

[Oldspeak]

Every frame is the same and it's just a simple anime clip. Pretend like a particular scene but someone decide to go loco and throw the focus all over the place. No cameras or any angle at all. I want to eliminate the panning motion of the camera and have it focus entirely on the scene.
Here's a example:
https://forum.videohelp.com/images/imgfiles/gallery/zD3eR
Imagine if the person made the focus entirely on the sky at first and then slowly pans down towards the middle(can't see the sky or the ground anymore except Milhouse) and then eventually to the ground. Is like trying to fit a image back together, even though it was a whole one piece at the start but segregated into different views.

Thanks for the help poison!

[Cornucopia]

@Oldspeak,

What pdr said is right on the money. Your original shot "1" mightn't have ANY of the visual background info present in say "3" (and vice-versa) and you WOULD need to "extend the canvas" to be able to handle the whole field of view. Doing so with a concurrent time-displacement calls for In-Painting along with the Motion-tracking of the foreground (as a stabilizer). Your ending DAR would also be very non-standard, requiring letter/pillarboxing. This would be a very complicated process in post-production.
If it's possible, it would be much easier to redo the production at a wider angle.

Scott

[poisondeathray]

Every frame is the same and it's just a simple anime clip. Pretend like a particular scene but someone decide to go loco and throw the focus all over the place. No cameras or any angle at all. I want to eliminate the panning motion of the camera and have it focus entirely on the scene.
Here's a example:
https://forum.videohelp.com/images/imgfiles/gallery/zD3eR
Imagine if the person made the focus entirely on the sky at first and then slowly pans down towards the middle(can't see the sky or the ground anymore except Milhouse) and then eventually to the ground. Is like trying to fit a image back together, even though it was a whole one piece at the start but segregated into different views.

Thanks for the help poison!

Sorry , it's still a poor description (or maybe I'm just too dense to get it)

OK, "focus" indicates something completely different... I'm sure you didn't mean camera focus in this context - you just meant pointed the camera somewhere else

Even if there is no camera - it's important to describe the type of simulated camera used (just pretend this was shot with a camera - describe what happens with the camera motion). The reason is the type of motion determines what software and techiques you would use to "fix" it. Details are important. Objects farther away are supposed to be smaller, objects closer are larger - also parallax effects.

Depending on the type of motion, you can use software stabilization e.g. deshaker, prodad mercalli, new versions of vegas - but I have a feeling this is not what you need

If there is no substantial change in angle that the foreground object is viewed at, and the simulated camera move was as if it was from a fixed tripod - then you could use the rotoscope and composite background approach

[Oldspeak]

Nah, I'm rather new to this so the apology is all mine. But what Corn have said pretty much nail the fear I had at the beginning. The whole thing of the first clip not containing any visual background of the next clip. Just hoping if there's a way I can fit/merge the first and the next scene together and etc, to make it into a one whole clip.

Although what you said at the last line sounds pretty intriguing. What I'm trying to edit is pretty much like that link I've shown (along with the camera angles moving ofc).

[poisondeathray]

Do you have a sample clip ? A picture is worth a thousand words, but a video of the issue would be worth a million

[jagabo]

If the background is static it's not impossible to do what you want. As I understand it, you have something like this:



And you want to change it so the frame no longer pans down. You can use the first frame to create the missing top part of the rest of the frames, and the last frame to create the missing bottom part of the earlier frames. Then you have to coordinate the motion of the animated clip over the background.

First construct a background from the first and last frames. I've darkened the last frame here so it's obvious what I did:



Then you overlay the original animation onto that background and add motion tracking to undo the panning.

[Oldspeak]

Nice, I was in the middle of making exactly what you just did and you beat me to it. That's exactly the problem and what I'm trying to resolve. Now I just need some guidance on Sony Vegas how to pull this off, my friend knows about as much as I do :/

[poisondeathray]

Nice, I was in the middle of making exactly what you just did and you beat me to it. That's exactly the problem and what I'm trying to resolve. Now I just need some guidance on Sony Vegas how to pull this off, my friend knows about as much as I do :/

If jagabo's gif is exactly simulating your problem , there is no pan or vertical tilt. The camera itself is moving vertically downward. This causes the picture to look like it's moving upwards. ie. it's a "pedestal shot".

Vegas doesn't have a motion tracker, so you would have to manually position each frame using keyframes. With a motion tracker , it would apply the opposite movement. e.g. if a frame moves 10 pixels vertically up in relationship to the previous frame, you move it down 10 pixels. So you would be left with a 10pixel black border at the top - that is why you "borrow" information from other frames as illustrated in jagabo's example, to fill in the black borders

A strict pedestal shot doesn't contain additional horizontal information in adjacent frames - you cannot reuse horizontal data, so a zoom out, or wider field of view is out of the question . If this is what you wanted to do, you would have to reconstruct the horizontal edges from other material.

[Oldspeak]

So let's say if I manage to piece each frame together and create a fluid animation in the end, wouldn't the new video I made actually be bigger than the normal size? i.e: The height of the frame will be twice as long as it fit the entire person into the shot when it was normally shown only in parts. Seems like this will create a odd video resolution as the height would be abnormally longer, causing more black border on the sides.

Garggabhll, I'm going to buy a study book on this!

[poisondeathray]

So let's say if I manage to piece each frame together and create a fluid animation in the end, wouldn't the new video I made actually be bigger than the normal size? i.e: The height of the frame will be twice as long as it fit the entire person into the shot when it was normally shown only in parts. Seems like this will create a odd video resolution as the height would be abnormally longer, causing more black border on the sides.

Not if is exactly like jagabo's gif . (And if it's not EXACTLY like jagabo's gif, then explain more clearly or provide a video sample.)

Is your foreground object moving in the air? like your bouncing ball example? Because the seesaw isn't. The base of the seesaw is level on the ground. This means the seesaw should retain the same physical relationship to the ground and background objects in all frames - despite the camera move. When you undo the camera move, the seesaw will be in the center of the frame in every frame. So the frame will be the same size and dimensions. But the bouncing ball case will not be. When the ball is high up in the air, it will be farther away from the trees and ground. Do you see the differences?

If jagabo's gif isn't representative: e.g maybe there is a lot more movement than depicted there, or objects go out of the frame etc.. then there might be other things you have to do

Normally if you track a bouncing ball into the air, you would use a vertical tilt, not a pedestal shot where the camera moves vertically up. The shots are different and the way you treat them is different.

[Cornucopia]

Yeah, in a PAN up, the distance/size relationships between objects in foreground, middle and background change, In a PED up, they DON'T - only the whole scene changes.

But I think what the OP wants is more like the 2nd GIF of jagabo's, and that DOES have a change in AR compared to the original, so it will need some type of padding/in-filling or cropping after-the-fact (or both) to retain it's original AR. If the OP is ok with a changed AR, then one less thing to worry about...

Scott