opacity and layers?

Posted by Anonymous User 
Anonymous User
opacity and layers?
March 28, 2013 12:25AM
i've never dealt with this issue and need some guidance. if i have the same
image stacked on itself ten times occupying tracks 1-10 and want to give each
track equal weight, to be able to have each instance of the clip show through
just as much as the other 9, how do i set the opacity? do i set each to 10%?
tried that and the overall was dark. set the track 1 clip to 100% and the
others to 10%? what's the formula for this?

thanks,
BabaG
Re: opacity and layers?
March 28, 2013 06:55AM
There is no formula. We don't know what effect or matting you did on each layer. You need to do the planning yourself -- if Layer 8 on Track 8 has an element that is different from the rest and needs to be seen, you need to make sure Layers 9 and 10 don't have elements that can cover that part up.


www.derekmok.com
Anonymous User
Re: opacity and layers?
March 28, 2013 12:09PM
this seems to work right:

track 10 1/10 10
track 09 1/9 11.11
track 08 1/8 12.5
track 07 1/7 14.29
track 06 1/6 16.66
track 05 1/5 20
track 04 1/4 25
track 03 1/3 33.33
track 02 1/2 50
track 01 1/1 100

to keep things even seems to require a graduated scale rather than setting all to equal values.

BabaG
Re: opacity and layers?
March 28, 2013 08:54PM
But it's still going to be influenced by the content of the vision, so maths is not really the answer. If you have a large bright object in one layer covering a small dark object in the same part of the frame on another layer, your mileage will vary.

Re: opacity and layers?
March 30, 2013 03:16AM
I have never stacked ten copies of the same image on itself. Curious what you are trying to accomplish, but I did stack at least two identical images. Then I offset them by about 1 second and used composite/blending mode to allow me to see through the top layer. I ended up with echo of the action with exactly the same image vibrance/opacity/look whatever on both instances. I did it this way because reducing opacity reduces color saturation on the opacity lowered layer and makes things all pastel/washed out looking. I did not want that. I have also stacked two identical versions of the same clip and used this composite mode change technique without time offset of the image just to give the image more depth/less flat and it will be more vibrant. I suggest experimenting with composite modes in addition to or instead of opacity to see if you like it. Only certain of the modes will make any diff depending on the type of images you have so you have to try them all on a couple of layers first to see which does anything. You might even try different blend modes on different layers of the same clip (on at least 3 layers/stacked copies). I can't remember what combination I ended up with but I was very happy with the results. It was an amazing diff from using opacity. If you want to see through something/transparency as an ethereal effect without the reduction in color saturation it is the way to go.
Anonymous User
Re: opacity and layers?
April 05, 2013 01:29PM
thanks for the input. the additional ideas are useful, generally, but i'd point out that the original question was about the SAME image, not the same image with different treatments. this makes the math more relevant. ganymede comes closest to the issue. would like to know how to do this with composite modes. it's a technique i'm familiar with using in film effects where the math is very relevant. and, by the way, these are b&w images.

thanks again,
BabaG
Re: opacity and layers?
April 06, 2013 03:09AM
My layers of the same image were b & w too but then I sepia tinted them before compositing. Even for b & w, I'm sure reducing opacity will reduce saturation and you will lose your rich blacks and get all shades of gray instead. As for how to do it, I recollect you just experiment with different composite modes applied to the top layer(s) to see which one(s) you like. Some will do something and some won't make any difference. I still am curious what specifically you are trying to accomplish since I am having trouble visualizing what you are doing from your description. Perhaps could be more helpful if I knew.
Re: opacity and layers?
April 06, 2013 04:18AM
Oh, I missed that it was the same image over and over, sorry. To use composite modes, right click on each clip and choose modes from the dropdown menu. You'll need to experiment a lot because result vary depending on your image. 'Overlay' might be the ticket though.

Re: opacity and layers?
April 08, 2013 09:58AM
BabaG: your question is important. Yes, the math is very relevant in film effects and it should be even more relevant in digital video effects because these are the results purely of calculations within the applications.

"Opacity" has no analogue in film. It is derived from the graphic arts where the opacity of layers of paint or of paper is physical and the effect is complex. In our virtual graphic arts "opacity" is an abstraction, just hinting at physical opacity, and this abstraction was too much for little brains of the writers of the FCP7 User Manual who write both:
"White represents 100% opaque and black represents 100% transparent."
"White is equal to 100 percent transparency and black is equal to 100 percent opacity (solid)."

The FCP opacity parameter and its operation in all the various FCP composite modes needs serious documentation. Here's how "Normal" mode works:

If at some pixel, N stacked images have the values (top to bottom) V1, V2, ... VN, and each image has its own opacity Q1, Q2, ... QN, then the composite has this value at that pixel:

Q1*V1 + (1-Q1)*Q2*V2 + (1-Q1)*(1-Q2)*Q3*V3 + (1-Q1)*(1-Q2)*(1-Q3)*Q4*V4 + ... to the N'th term

I verified this by making a big stack of random grey tones and applying random opacities to each.

Your N is 10 and you want all 10 coefficients to be equal so the 10 images get equal weights, and you also want the composite of identical images to look just like a single image, so you can solve for the Q's and get:
Q1 = 1/10
Q2 = 1/9
Q3 = 1/8
...
Q10 = 1/1

This is exactly the solution you found. Bravo!

The long sum reveals how opacity works in FCP Normal composite mode. The values of the separate images are added with weights corresponding both to their opacities and to what all the opacities of the images above them have missed. This is much simpler than how physical opacity works, and even a bit cuckoo, but it is our algorithm.

I expect that the FCP opacity parameter will have different meanings in the different composite modes.

Dennis Couzin
Berlin, Germany
Sorry, only registered users may post in this forum.

Click here to login

 


Google
  Web lafcpug.org

Web Hosting by HermosawaveHermosawave Internet


Recycle computers and electronics