How to create this beautiful effect?

[stephan80]

Hi!

I would be really interested in knowing how this beautiful effect have been created: it's a movie of a guy kiteusrfing. while all the image is moving normally all suddenly ONLY THE KITESURFER starts to go in slow motion while all the rest keeps going at normal speed.
You can udnerstand better watching it:
http://it.youtube.com/watch?v=uCWBfW9lbDI&feature=PlayList&p=C3F2588306C4178E&playnext=1&index=15

Go to 0:55 and you will see what I mean.

It would be simply amazing if someone could tell me how this effect have been created. A tutorial on how to do it would be great!

thanks!!

[Snakebyte1]

Hi.. I can't see the video here from work so I may be way off, but what I am visualizing would be to use two layers. The background would be one layer and would run at normal speed. The surfer would be a second layer composited on top of the background layer, and its playback speed can be modified as desired.

[stephan80]

Hi, thanks for your reply.

Your idea was what came up to my midn as well. But then after I started thinking that maybe that's not the solution since the camera si moving even while the slow/normal motion is going on.
So taping 2 times the SAME IDENTICAL spot moving the camera the same identical way, one time with the kiter and one more time without him, seems pretty impossible to me. Plus you shoudl consider the water: it's not like taping a car on the asphalt that remains always the same (the asphalt). The water is always changing the shape depening on a lot of factors and one of the factors is the surfer int he water. So to me seems kinda imposible to tape 2 times the same spot one with the surfer and one without him and then merge the 2 layers...
If you could watch the video I am sure you would understand better what I mean.
thanks for your help!!

[Snakebyte1]

You would not have to have two separate shots necessarily. Place the shot on one layer, then place a copy of the same shot on an above layer. Using a mask, mask out everything except the subject (ie surfer) on the upper layer. Then, slow down that layer and use motion paths to move the subject so that it hides the original subject on the background layer. Since you are working with the same clip the camera angles and motion would be identical. You just have to adjust the positioning of the subject to compensate for the fact that it was slowed.

This is just something that comes to mind. I'm sure there are several ways to do it based on what software and equipment you may have at your disposal.

[johns0]

Snakebyte1 is right,you can tell the slowed down parts are a second layer because the wake is going in slow motion as well as the surfer but the rest of the scenery is moving in real time.

[stephan80]

hey thanks a lot for your help.
Things starts to become more clear even if hiding the original subject seems pretty hard to me. He is moving quickly from one part to the other of the screen. Plus he is producing alot of water "splashes"...
Let me know if you can watch at the video that would probably help a lot...

[stephan80]

Anyone knows where I could find a tutorial on how to reproduce this effect?

[redwudz]

It looks like this is the production company: http://fatsand.com/
From their site, they mention they are using Apple's final cut studio software. That probably doesn't help unless you have a upscale Mac, but you might check out the software to see the capabilities and find similar with a PC solution. http://www.apple.com/finalcutstudio/

[PuzZLeR]

Adobe's latest (expensive) line has such effects.

For example, a film of vehicles in traffic on the Golden Gate Bridge was displayed.

Then the same clip, with the same bridge, was shown in another clip - the bridge was empty. No vehicles.

What the software did was calculate, and re-create the background, that the vehicles were driving over. It can even show the vehicles separately in another clip, in the same motion sequence, in another clip with a different background.

Something like this would help in your case - recreating the backgrounds as separate layers.

[edDV]

Its more technique than canned effect. You need to separate layers by creating or extracting an alpha (aka: mask, key, matte) either manually frame by frame or find info in the shot (e.g. color, motion, luminance value, edge characteristic) that allows automating the separation.

Some of the techniques:

- rotoscope (manually paint a mask)
- luminance key
- chroma key
- motion detection to mask

Don't expect this to be easy.

[Snakebyte1]

I finally got to see the clip and it was a bit different that what I had imagined, but its obvious that they used multiple layers and masks.

As edDV pointed out there are various techniques available to create such mattes and masks using various software to do it however the concepts are pretty much the same no matter what software you use. Higher end packages offer more controls, variables, settings etc used to refine the masking which you would need when trying to mask water spray or hair blowing in the wind etc, but its quite doable if you have a bit of experience and practice. There are all kinds of tutorials out there on how to create mattes and masks in various programs like After Effects just to name one if you wanted to get an idea of what is involved.

[stephan80]

Adobe's latest (expensive) line has such effects.

For example, a film of vehicles in traffic on the Golden Gate Bridge was displayed.

Then the same clip, with the same bridge, was shown in another clip - the bridge was empty. No vehicles.

What the software did was calculate, and re-create the background, that the vehicles were driving over. It can even show the vehicles separately in another clip, in the same motion sequence, in another clip with a different background.

Something like this would help in your case - recreating the backgrounds as separate layers.

Hi,

to wich software are you referring exactly?

[stephan80]

Snakebyte1, edDV, thakns for your advices. Since I am not an expert of editing but just medium level, doing it all manually sounds pretty difficult to me.

It would be nice if like PuzZLeR sayd there would be a software ready to calculate and do it!

[edDV]

Snakebyte1, edDV, thakns for your advices. Since I am not an expert of editing but just medium level, doing it all manually sounds pretty difficult to me.

It would be nice if like PuzZLeR sayd there would be a software ready to calculate and do it!

Unlikely, this is an advanced effect.

[guns1inger]

Even software like Mokey, which can be used to remove object from video require some skill and knowledge of masking and rotoscoping.

Also, if you frame by frame the video you will notice that the background water actually changes at the point where he freezes the first time, and back again when he unfreezes the second time. In between it uses two distinct clips. It is well done, but not overly difficult.

The reason why most newbies find these types of effects difficult or impossible is because they don't plan for them They see them, then try to recreate them with footage they already have. If you are serious about trying to do this type of thing for yourself, then plan for it. Block out how you will approach it, list out the footage you will need and the shots required to get it, then go out an shoot it. You may have to shoot 10 - 20 takes of the background to get one that is finally usable. So what. Tape is cheap.

[edDV]

If you shoot low against a blue sky you may be able to chroma key.

If you shoot early or late in the day, the sky is brighter than the subject allowing luma key separation.

Be sure to shoot second layer at the same time to make the lighting match.

[stephan80]

Amazing advices.
Now things starts to become more clear to me. I should try to tape 2 times the same exact place with as much as possible similar light.
What made me confused was the fact that I was usre it was the main layer only duplicated.
While indeed seems like they taped 2 times the same "spot".
This way things seems more clear.

[Snakebyte1]

You can also create a mask by manually setting points around the object you want to composite. If the object is in motion you would set Key Frames and manually adjust the points of your mask at those key frames. The application would them automatically adjust the mask between the key frames for you. The faster the action the more key frames are needed. Its not overly difficult but it can be tedious manually adjusting the mask points.

Once you get your mask drawn you can then adjust some of the settings like feathering and blurring around the edges. This helps blend your object into the background clip. This is where practice and experience comes into play, knowing what setting to adjust and by how much.

Guns1inger is absolutely right in that the more you pre-plan the easier it will be to do the effect and the better the end results will be. For example, If you wanted to film a person on a beach but you shot the person against a brick wall, when you try to insert them into the beach background you'd see the brick wall through the flowing hair. But, if you filmed the person against the sky or against a background similar to the beach background you wanted to use in your final composite, they would likely blend in much better. Alternatively you could use multiple tools to improve the effect. If the background behind the person was a solid colour, you could attempt to key out that colour just around the hair. If you had identified this as a potential problem before filming you could have your subject wear a hat and thereby avoid the problem. The more you think things through before you film the better off you'll be.

Using tools such as keying on a particular colour, such as Green (ie green screen) requires, in most cases, a lot of planning in order to get lighting right. In my case, I often use a lot of trial and error. I'll try one keying method and if after some effort I still can't get what I think is acceptable, I'll try a different one. As noted, there are many different tools used for different situations.