[Ben Rollason]

If it’s in 2D, which I guess it is, ‘cos you said circle not sphere…. how about

on the anchorPoint

circleRadius = 100; //the radius of your circle
seedRandom(this.index, true);
[random(circleRadius),value[1]];
//[random(circleRadius), value[1]]; //uncomment this for in a circle, instead of inside a circle

then on the z rotation:

seedRandom(this.index, true);
random(360);

then set the position to the centre of your circle.

It’s a fairly simplistic method, but it should work. If you don’t want to move the anchorPoint, you need a bit more math. You could put…

circleCentre = [100,100]; //your circle centre
seedRandom(this.index,true);
circleRadius = random(100); //your circle radius
//circleRadius = 100; //uncomment this if they should be in a circle rather than inside a circle!
tRnd = degreesToRadians(random(360));
circleCentre + circleRadius*[Math.cos(tRnd), Math.sin(tRnd)];

on the position property.

If it is in 3D space, just put the z rotation expression on the y and x too.

-Ben.

vfx.benrollason.com

[Ben Rollason]

I’ve tried most things. At higher blur values, the blur moves in discreet steps… it will hold the same frame for four or five frames and then change. And this regardless of whether you use keyframes, linear, eased, continuous or use a smoothly changing depth matte layer in 8, 16 or 32 bit.

For instance, animating the Blur Focal Distance without using a depth matte in 16 bit, the property’s value on frames 2,3,4,5 are 20580, 20797, 21015, 21232 respectively – a uniform change of 218 per frame. Yet frames 2,3,4 are identical and 5 is suddenly different.

I think there’s something kooky with the maths inside the plug-in.

-Ben.

http://www.benrollason.com

[Ben Rollason]

You probably need to be inputing the positional data in world space

instead of

tC = thisComp.activeCamera;
tL = thisComp.layer("a child layer"); //this layer is parented to something.
tC.toComp(tL.position);

maybe something like this.

tC = thisComp.activeCamera;
tL = thisComp.layer("a child layer"); //this layer is parented to something.
tC.toComp(tC.fromWorld(tL.toWorld(tL.anchorPoint)));

I think that’s more universal as it makes sure your original point is in worldspace, then brings that into the camera’s layer space and from there into the camera’s 2D comp space.

-Ben.

http://www.benrollason.com

[Ben Rollason]

Parent all of the faces to a null.

Then this in the scale.

tOrigWidth = 300;//width of original source
tOrigHeight = 100;//height of original source
L = thisLayer;
[tOrigWidth*100/L.source.width, tOrigHeight*100/L.source.height, 100];

Wouldn’t it be easier to use precomps though and just get your user to import their cover into the precomp and resize it?

-Ben.

http://www.benrollason.com

[Ben Rollason]

[Arvin Bautista] “Now to figure out how to only handle integers…”

Math.round(value); 🙂

-Ben.

http://www.benrollason.com

[Ben Rollason]

This issue has existed for years and it is a problem regardless of how many million of satisfied iPhone customers there are. As a video professional, this problem has cost me weeks and months of time.

I assume Apple is trying to compensate for some legacy gamma issue. Or that it assumes that everything Quicktime was originated on a PC and played back on a Mac. It is color management that simply doesn’t work.

When I compress a piece of footage into a Quicktime format, I don’t expect it to raise the gamma to 1.2. The original and compressed file look completely different side by side on the same desktop. Different original codecs come out differently although the original files look identical.

I stopped delivering QTs to facilities for laying off to tape years ago, because it’s too risky.

Quicktime X is much better. I haven’t fully crash tested it yet, but I haven’t noticed the gamma problem nearly so much. Quicktime X and Quicktime 7 actually display the same video differently, but I don’t think it’s available for PC yet.

-Ben.

http://www.benrollason.com

[Ben Rollason]

Thanks Randy,

But did you mean to post a link to it? I can’t seem to find one!

-Ben.

http://www.benrollason.com

[Ben Rollason]

Thanks Alan. Worked perfectly!

ben.

http://www.benrollason.com

[Ben Rollason]

Thanks Xinlai!

-Ben.

http://www.benrollason.com

[Ben Rollason]

I don’t think it’s a fields issue. The artefacts in your video look to me like the perfectly normal side effects of trying to slow down a piece of video 20 times. Twixtor is very clever – as are the built in frame interpolation methods in After Effects – but these technologies can’t perform miracles.If you’ve slowed down your footage 20 times, the software is having to create 19 frames out of thin air for every one that you shot. The interpolation between frames is linear, which is why you get this strange cadence.

The more you slow footage down, the more obvious the artefacts become.

In my experience, using these effects are no match for shooting with a high speed camera and we can’t really expect the results to look the same.

As a rule of thumb, reduce the shutter time as you shoot by the same factor as you wish to slow the footage down. i.e 4x slow-mo, shutter time = 1/200th. This will make the motion blur match correctly. Apart from that, it’s simply a matter of tweaking the settings until you get it as good as you can.

https://www.benrollason.com