PreComp transformations collapse & cache by approach

Similar Questions

• Mix of framework (optical flow) Composition, so I did not collapse Transformations?

  Hello

  So I think I get that if you want to time a model, you must define pixel motion for images in the preComp then collapse Transformations and assign Pixel Motion, because you can not FrameBlend a model more.

  But what happens if I don't want to do that, because it would break other animations?

  Is the only way to use the time warp effect?

  Is there an easy way to link with time-remapping, so it can be applied to a existing built, comp time remapped with pre nested compositions using effects, I don't want to collapse?

  Or is it a new way of thinking.

  Is there a tutorial somewhere on this subject?

  TRIS

  because you can not FrameBlend a model more.

  to my knowledge, there has never been an option to use the native frame of mixture for a demo. only for images.

  If you want to time stretch a model, you must define displacement of the pixels for images in the preComp then collapse Transformations and set it to the displacement of the pixels

  If you want to use the native mix frame (temporal, time remapping) then YES

  But what happens if I don't want to do that, because it would break other animations?

  Is the only way to use the time warp effect?

  You can always make you composition and re - import (also called prerendu) and now it's pictures you can use native on interpolation. If you do not use Timewarp, you can also use a plugin of 3rd party like the critically acclaimed Twixtor - RE: Vision Effects . If you need control actually more, even if you have a merge option setting, you would use Timewarp or 3rd party plugins.

  Is there a tutorial somewhere on this subject?

  TimeWarp and Twixtor - more than you can count. on TW-, it's my personal workflow:

  1. Firstly - I put the timeline to display pictures and not in seconds

  

  2. then I add layer time remapping

  

  3. Add the effect of TW and set it to executives, not speed

• Collapse a masks of the transformation parameters

  So I have an adjustment with a mask layer and I adjusted it's position/scale/rotation with the transformation parameters and I want to copy the mask to another layer, and of course when I do this the mask goes to form, as it was before the transformation parameters have been applied, if I copy the mask as well as the parameters of transformation , the transformation parameters to ruin the layer they are copied and while I know there are ways around that long, I wanted to know if there was a way to simply reduce the current transform parameters of a mask so that the mask would stay as it is with the transformation parameters reset.

  Masks are based on the size of the layer not scale so the behavior you describe is planned.

  There are a few things you can do.

  First of all, I would modify the workflow and instead of scaling of your adjustment with his mask layer I want to animate the mask path or adjust the mask path. If this is not practical, you can duplicate the adjustment layer, remove effects, turn the adjustment layer switch, then use the copy as a matte for your another layer.

  The other option would be to use the layer > Auto Trace to generate a new layer with a mask solid. This mask can be used as cache by approach or the new mask copied and pasted to another layer.

• collapsed borders CS6 vs CC precomped layer

  Hello.

  Just noticed something about the precomped layer in earlier versions.

  in CS6 - when you precomp a collapse of the layer and press you can actually see

  the borders of the layers inside, directly in the layout of the master.

  as follows:

  

  If the precomp elements you see this:

  

  but in CC and CC 2014, the result is the same:

  Anyway - borders of composition and not layer borders i.e is collapsed and items remains the same.

  It seems that in CC, the only way to reproduce this behavior is to the two precomped layer and

  the precomp himself 3d:

  

  Anyone got any information on this subject?

  Hey Roei. It's been almost 2 years, and now finally in CC 2015.3 aka 13.8.1 they fixed. You can watch it here at 08:20.

• Add the precomp major issue mask

  I guess it's yet another bug in AE 13.6.1.6?

  I took has several layers with overlays, precomp... nothing new here.  Click on the "for layer Comp: Transformation of the collapse" button to keep my overlays... again, no problem. Ahh, now add another layer as a 'matte' and all recoveries precomp disappear and do not work.  Really making me nervous, tried 6 ways to Sunday to work around that and I cannot.

  

  If there is only one fusion mode then just grab the blending modes in the layer (pre-comp) nested and edit it to add. Which is available even if it is not visible. Everything that you CT will work. If you're just hide or animate the lower third with a cache by approach then CT should not be necessary. Don't get me wrong. I use ways of merging all the time. I just know that if I want to embed a layer that has a bunch of mixed modes and I need to hide this layer nested with a matte or mask that I have to design so that I have not to condense the transformations or I'll have to animate the mask or add the mask in the nested model. I have this all the time. Unless I'm completely wrong, understand the use of the matte, I see no reason that it could not be included in the demo.

  Give the blending mode to ADD a test on your computer at the title bar of switching. I could offer you a better suggestion if I knew what was the structure of the model.

  If it was me and the place with 13 inside and the bar with the necessary addition mode comes alive with a matte that I'd keep probably the two layers in the main comp, parent from one to the other and use Set Matte to wipe it off with a demo as a source. Another approach would be to implement the expressions to animate the size of the shape layer, you use the bar based on the in and out point of the layer, perhaps put the 13 in a box on the same shape layer, use blending modes internal and trains facilitators to make the box blade in as needed without keyframes at all. Just setting the in and out point would do it. The shape layer would then be saved as a preset. This is another thing that I do all the time.

• Pixelation continues with confined comps, 3D layers and caches

  I have a project consisting only of objects HAVE contained in comps, used with 3D and I have a similar problem to many I've seen on this forum, but I tried all the solutions provided without result.

  The lowest model level are a few layers to HAVE that are animated to the comp. appeal he comp A.

  Two copies of model A are in comp B, each track stuck with another layer of HAVE.

  Model B five copies are in C comp.

  Model C is placed repeatedly in 3D space in the main computer where a camera moves beyond.

  All vector layers are 3D and set its CR

  All the layers of the model are in 3D and NOT worth CT

  Without the appropriate parameters for CR/CT, caches everywhere drag (but rasterization of things well).

  WITH the settings above, nothing not rasterized correctly (but the caches work). The compositions appear returned to the original format of the model A, which is highly pixelated (or blurred with quality facing up) that the camera is moving beyond.

  I tried precomposition caches and their objects contained in comp B in their own model (also according to settings above) and got no change.

  Of course, with hindsight, I could have climbed A comps objects, B and C, but at this time the project is in the final stage and I am wary of all the settings needed to go this route (it also seems like a very bad way to go about working with vector graphics).

  Can anyone help, or it's just the way that is AE? (I'm experienced enough, but 1 year <.)

  Thank you for this forum.

  Here's what you're probably missing. We're going to implement something very simple. You have a hand Comp and a PreComp. The main computer contains the PreComp and a camera. The PreComp contains only single-layer - TestShape which is a 20 X 20 pixel solid red and he has a circular mask applied so that it looks like a point. When you add a mask to a solid you've essentially turned this vector layer in a layer.

  Let's go through a certain number of things in the PreComp. With the fist, TestShape 600% globally and look at the edges. They are a mess. Turn on continuous rasterization and they are perfect.

  Look at the model of the hand and intensify the layer PreComp 600%, and the terrible aspect of the edges. Turn on the Transformations of the collapse and the edge is sharp.

  Note that I used two different terms for the same key. Collapse Transformations only applies to layers that are compositions. This means that elements inside this composition are treated as if they were in the main computer. In other words when the collapse Transformations is applied we are scaling of the TestShape layer 600%, then who still 600% effectively making the layer 3600% greater. Who do you follow? It is important.

  Now back to the PreComp and add a solid second blue circular 20 X 20 with a mask. Call the Blue Solid 2 TestShape and make it 3D and turn on CR. Add a camera to the comp. Adjust the z value of the layer 2 TestShape so that the layer is about the same size as the 2D TestLayer. If the layer of TestShape is proportional to 600%, then the position of the Z TestLayer 2 will be about 80% of the value of the camera Zoom. Now the position of the camera X just a little to the right in order to have a blue circle overlapping a red circle in your COMP. Turning CR on and off the layer of TestShape 2 to see how it works. Pass the TestShape layer 2 more near the camera is exactly as the scaling of the layer.

  Now return to your hand Comp. With CT turned on, you should see only the blue circle. This is because CT is ignorant of the camera in the nested model and using only the Active camera in the MainComp. Disable the CT and you will see what you see in the demo, but the edges will be soft. The only way to get the same point across in two compositions is to copy the camera in the PreComp and paste it into the nested model.

  Here is one more interesting thing happens. If you do the PreComp one 3D layer it will multiply the value of the position of all 3D layers in the PreComp making camera movements to an interesting but impossible to deal with. In this example if you were to make 3D PreComp it would look like the blue layer has disappeared, but simply move the camera back and it will appear again.

  Now let's add a matte by creating a new rectangular solid that is 1000 pixels wide and 200 pixels high and add a mask elliptical. This 600% increase scale use as a matte and turn on or off the CR. The results should be predictable. If you make a layer of mast track 3D move the Z position will not affect what is frosted and outside in the nested model. If you had 2 3D layers in the nested model you cannot move the cache by approach between the two 3D layers and mask only one of them.

  I hope this helps. This is a good exercise to help you understand how CT and CR. Beginning throw in effects and learn more.

• Need help to transform a shape layer in a layer mask blur adjustment

  I understand how to do a blur mask on the image adjustment layer, so it blurs the images below, which is essentially what I'm trying to do, but in this case, I already have the shape layer (or rather imported from Illustrator > converted to a shape layer) and is in a nested composition.

  First of all, is it possible to blur the images of a composition if the adjustment layer is nested in a composition of the child, or the adjustment layer must be in the parent publication; the composition is blurry?

  Second of all, how we simply take a form that they have already created and use that for a mask, in this case, a mask that will blur everything below?

  Forum thanks!

  There are a few things that are not effective in your workflow and your approach.

  Lets talk first about the Illustrator files in After Effects. Firstly, absolutely the only reason you would ever convert a layer to illustrator in a form would be to use animator tools found in the add menu at the top of the shape layer properties in the timeline panel or to animate a Bézier path on this layer. There is no other advantage in converting a vector layer to a shape. NONE.

  Second, if you want to hide an adjustment layer then you can draw the mask directly on the adjustment layer or any layer with transparency, including an illustrator layer, layer photoshop, any hidden layer or layer shape directly on top of the adjustment layer and use this layer with transparency as an Alpha or Alpha reverse cache by approach to the adjustment layer.

  You can use the luminance values of any layer that has no transparency that a luma or reversed hides by luma approach. The white areas offer transparency with luma standard. IOW, if you have a radial with white in the middle and black gradient on the edge and you use a luma matte and white will be transparent on the layer below.

  Standard caches only affects the layer immediately below the source.

  Adjustment effect layers each layer directly under the adjustment layer. Masks and caches can be used to control the effects of an adjustment layer.

  You can also use masks on any layer to control the effect added to this layer if you have a recent version of AE. You will find the option under options to Compositing.

  If you have any layer with transparency or the strong luminance values, you can use Auto Trace to generate a cache for this layer based on transparency or luminance. Any way shape mask or Bézier path can be copied and pasted to another layer.

  We will take your last question now. You should be able to answer. To use any shape you created as a mask, you use this layer as a track matte.

  We will change that approach a bit and assume that you have created a shape layer by using the pen tool or converted an Illustrator layer to a shape. You have a path that you can be found by selecting the shape layer and by pressing twice the key U. find the path you want to copy, rotate options path to reveal animation controls and to set a single keyframe and then copy (Ctrl/Cmnd + c). Even of the forms created with parametric tools like star, ellipse, rectangle can be converted into a path by right-clicking the shape and selecting convert to paths. When you have copied the path put you on the layer where you want to paste the path, select the pen tool to begin to draw a mask, click anywhere on the layer and paste it (ctrl/cmnd + v) and that you have created the mask. This works for forms and masks on any layer.

  Here's a bet to keep everything by copying and pasting masks. The size and position of a path is calculated from the center of the layer and has no relationship to scaling, rotating, or the position of the layer. If you drew a mask on a layer that is resized to 50% and moved to the top right of the comp and rotated 45 ° and then paste this path to any other layer you will NOT get a path in the same position and the same size as the layer sampled in the model. You end up with a path that match if every sampled layer transformation properties have been reset and the layer was at the center of the model and the rotation and scale were normal. You must pay attention to this issue by copying and pasting traces of Bézier.

  I missed a question type. If you nest or call a layer previously and there is an adjustment layer in the main computer it will affect all layers below. If you dial before a layer and a top adjustment layer then only layers in the nested model will be influenced by the adjustment layer.  I replied before but wanted to make sure.

  BE very careful when you watch YouTube tutorials on workflow because a growing number of them is bad teaching techniques and pointing you to the bottom of the dead-end alleys. Make sure you vet your training.

  If you have any other questions the best place to find good community resources is to type a word in the search help field in the upper right of the AE. For example, if you typed layer adjustment or masks, you would find a lot of very good advice on how to use these features.

• mask changes size when overlay cache track is used to link a clip to it

  Hello

  I'm new to the Prime Minister and made only very basic clip front mount.

  I want to apply a mask to a fixed element and animate both of the scale in and rotation.

  Everything goes well with the scale and rotation of the identical key frames.

  Until I have add overlay track matte for the clip and the layer mask.

  The mask reduced in size, but still indicate the values of the appropriate scale associated with keyframes.

  If I increase the masks scaling, it behaves as if it is to be cut by an invisible box and cultures on all sides

  Can anyone help because I am confused?

  If I removed the link cache by approach, everything behaves as expected?

  Masks and hides by approach work with alpha and probably get in way mutually.

  By nesting eliminate you conflicts.

• disable the cache

  Hi, I found there are 2 ways to disable the cache.
  
  (1) edit \OracleBI\server\Config\NQSConfig.INI, change of
  
  ===================================
  [CACHE]
  ACTIVATE = NO;
  ===================================
  
  OR
  
  (2) uncheck the box 'Cacheable' on the dialog box properties of each physical tables and views
  
  WILL BE THE METHOD 1 AND METHOD 2 BACK EVEN EFFECT?
  
  Thank you!

  the difference between the two approaches is the scope.

  Activating/Deactivating the cache in the file NQSConfig.ini, you will have a global effect, all by checking / unchecking the box "cacheable" at the level of the RPD table allows you to choose the tables whose data can be cached. both approaches will have the same effect.

  the choice depends on your specific requirement of the project.

• Effects WE Walker prémultiplier

  I have CS4, but I was reading an old book by the Meyers, create Motion Graphics with AE flight 1. the essential (© 2004).  On the DVD, they have a few free Plugins - by various manufacturers.  In the chapter projects that they mention that some of the plugins work better than those provided in the EI program (i.e. AE version 6.5) because I like the idea of having more toys to play with, of course, I want to have them at my disposal in AE CS4

  However when I placed the where the factory designated (AE Plug - in the folder) and open lessons AE of the book, I get the message that the plugin is missing.

  I was wonder 2 things.

  It is likely that the plugins are so old that they are simply not compatible and therefore do not appear in CS4;  and

  Since CS4 is more recent than when the book was written and those 5 years made plugins, it is also highly likely that AE now comes with plug-ins that are better than one of those old freebies and/or the features of the AE has improved so that those plug - are obsolete.

  For example WE premultiply was supposed to be good by knocking out a solid black color when you use it for a stray light or other lighting effect.  Is it possible that now in CS4 the "Multiply" blending mode is much better and that's why a plug-in like prémultiplier US is useless.

  I am new to AE with CS4 (2009) so I have no reference to compare to.

  The gifts are: Boris Continuum Complete, DigiEffects Delirium (fireworks and OF Glower), Digital Anarchy text display, digital tools (WarmCool), red giant (IT Alpha ramp), The Foundry (T LensBlur) and effects of Walker (prémultiplier US)

  Finally, install one of these old effects eventually would lead to conflicts and accidents.

  Thank you in advance, I will wait for your response before you try to install one of the other Freebies

  You are right that the old AE 6.5 plugins are just too old for CS4 compatibility.  Note that even now we all get around the fact that CS4 plugins do not work in CS5!

  There are a number of ways to unmultiply / prémultipliez layers.  I usually precompose the layer in question, then duplicate the layer of precomp-ed and assign it as Luminance cache by approach to the first.

  The most common third eruptions optical Lens Flare Plugins Red Giant Knoll Lens Flare and video of the first officer, both have unmult capabilities built into them, that's why many plugins to standalone unmult have not been updated in recent years.

  But Red Giant Software still offer a free version of UnMult.  It is labeled as AE 7 compatible - you could try and see if it works in 6.5.

• Animation of zoom/dolly in the map of the world, the best workflow?

  Hi all, I'm animating maps for a travel agency and they want me to zoom to a vision of the world in smaller places, like Costa Rica for example. Then animate a few lines and icons on the map in 3D. I've attached a video link of a that I already have. I work with an illustrator maptorian of the world card and import the illustrator file in AE. Can I use a camera to dolly in and making all the moves. My problem is that AE runs VERY slowly when you work with such a giant illustrator file (there must be millions of vectors). I tried to work with PNG files exported, but resolutions should be too big for them not to Pixelize (200 million pixels is not enough!) and then I met the same problems of slowness.

  A second problem I have is with the shadows. because my illustrator file imports as 2800 x 1600 when I Zoom so far (although I am rastoring) my elements are only 1 pixel above the map (or at least this is the unit of measure AE uses even with the base vector files) so as you can see in the video the shadows move around annoyingly.

  Can someone advise me on a path that I could work with a low card-based pixel resolution of the whole world and then zoom in on a small card vector space, that I need to work. Or any other ideas are welcome.

  I'm using a Mac pro running 10.11.6 with 6 core and 16 GB of ram and AMD graphics cards.

  AE is 13.8.1

  Thank you

  The illustrator file is highly textured and takes too much time to calculate, you should cut your chart into segments. I made maps of the world for many years of news and documentaries, to show the world to zoom in on a country and a city. You should separate you graphic into segments. map of the world say 4000 x 4000 pixels for example. and costa rica to 4000 x 4000. now composite costa rica on your map of the world. My guess is that you have a few places to zoom in on so just make these large sizes and place them on your map. If you have a high resolution on a large file low resolution.

  For example, this card I did. I have to get all the way zoomed between the Morocco and the Spain, then completely out to part of the Middle East.

  

  So I have this model master 1605x905px

  

  You can see the limits of layers. I carefully place one above the other and resize it down. so I have a map of the area I zoom in to 1313x922px reduced to 60%. now, because I had yet another card on top of these 2 and he had to be reduced to very low values that are more difficult to place, I created a precomp with collapse and inside, there is another card to 2356x1493px 7% nationally.

  

  This way I can cover on the other and do not have too many pixels at the same time.

  Another example more complex because it contains more elements which must be organized differently to support a fine composition on both zoomed in and zoomed in on the map.

  

  same idea. This type of card is dissolved in as I crash zoom

  

  This guy a precomp with many items I can carefully place if the composition does not suit me perfectly.

  

  Fun, right?

• Hair dye in After effects?

  Hei,

  My question is above.

  How can I dye my hair in darkest legacy?
  Thanks for your replies

  You can isolate any detail in a picture in AE, but it requires a fairly high level of competence and pre-production planning. The better pre-production planning and more the quality of the original film the easier than it is to do. It's all about creating mattes (masks). There is no magic click button problem.

  According to the pictures, it may require that you:

  • Follow in Mocha, stabilize Track Motion, Motion, track, track camera mask
  • Rotoscoping, Rotobrush
  • Adjust the color channels
  • Use the blend modes and pre and roto comps to combine multiple caches in a usable hand pole
  • Matte and adjustment layer with the (possibly dyed) color effect applied)

  To learn more about these types of techniques, none of the items in the field find help in the upper right of the AE and learn about the resources of the community.

  The basic procedure is to isolate the hair from the rest of the image. It is usually a 2-step process involving a rough selection (rotoscoping, mocha track) on a copy of a copy of your film layer, then isolate and separate the hair of the immediate background by manipulating channels color and luminance, so you can create a cache in order to limit a color changing effect. That's why pre production, planning and high quality is essential to the process. Blond hair blowing in the wind with your market through a wheat field model is going to be very difficult. Blond hair to the wind with your model on a dark background which is quite monotonous just is going to be difficult.

  Open the pictures that you plan to use and take a screenshot and drag it to the field of response on this forum so that we can see what kind of images you have to deal with. Yet these frameworks are all available on Adobe Stock video clips. If you have not drawn the images still find something similar and we are pointing to it.

  4 hours really hard work because of the soft edges and soft in the foreground - perhaps for 10 second tour:

  

  Easier - perhaps a few hours for 10 seconds:

  

  Well planned and shot - maybe call 45 minutes for 10 seconds:

  

  The most difficult part of this shot is separating the hair out of the face.

  Lastly but not least - of good luck with this one:

  

  It would be possible but blur of movement and the uneven background would be a bear to deal with.

  Here is a start on the "easy" one - the workflow is:

  1. Duplicate images
  2. Mask to follow from the outside of the hair and face
  3. Start working on the details of the outline of the hair using Colorama to get some dark borders outside the hair

  It's that I received this morning. Took me about 15 minutes. Here's what it looks like:

  

  The procedure will be repeated for other color values using additional hide until you have enough layers to combine in order to get a black mask and decent white generated when the hair is black and everything else is white. I guess about 4 or 5 more duplicates. The girl walks so I'd probably motion stabilize the whole thing so the rotation of the hand would be easier. I would use just a solid black to fill all the holes between it the edges. All duplicates in grayscale would be composed pre and which would be used as cache by approach for one layer setting I would use to tint hair. The second set of 15 shot probably would take two to three hours to perfect. Here's what it looks like with a layer of black shape quickly drawn to fill in the hair. The hard edges around the face and shirt can be fixed with some pre mattes more. The blonde is now a red head:

  

  Good idea for a tutorial. I should write a book - hint hint

• Make files hidden media?

  What is the point of this folder? Whenever I start PP a few minutes while working on a project of my SSD is full with more than 30 concerts of this shit that makes it so I can't save or export because it fills the reader. And every time I remove my project is fine its exactly how I left it, but the files keep rebuild no matter how many times I remove them.

  So, what are they do other then to be boring?

  Hi DennisJa,

  Matter what Premiere Pro video and audio in some formats, she transforms and caches versions of these items that he

  easily accessible when generating previews (these are the media Cache files). Imported audio files are each followed a new .cfa file and MPEG files are indexed in a new .mpgindex file.

  The media cache greatly improves performance for previews, because the audio and video elements do not need to be reprocessed for each preview.

  To clean the media Cache:

  Edit > preferences > Media (Windows) or Premiere Pro > Preferences > Media (Mac OS)

  Thank you

  Ilyes Singh

• Keys on a white background and track 3d movement

  It is actually two questions, so if I need to split it into 2 posts, I can do.

  First of all, I have my talents on a white background (because of the green in costume) and I want to put a background behind them is a picture of a track or a stage (see the second screen capture). Initially, I did this by going into AE and make two copies of the item. I then used the HUE/SATURATION effect to the top of the page copy b & w, and then the LEVELS to make high-contrast. Then I put it as a cache of luminance (reversed) on the background copy. It worked pretty well, at least better than Ultra key or Luma Key, which left a bit of white on the edges of the subjects. Still, you can see a white border around the topics today, although he the one smaller than with the coding of effects. Does anyone know a better way to remove the white background, so there is no white edges around subjects?

  Second, you can see that the microphone has a green cover. I need to put our logo instead of the green in the football game perspective and rotation. I thought to follow up with an effect of pin corner of movement, but because of the rotation, I don't know if it will work. That talent is the mic autour, the Green Exchange box tilt and rotation, as well as position. What do you recommend that I use to match my logo (seen in the second screenshot) the microphone?

  

  Screenshot of the music video with white background

  

  Capture the screen with white background clip removed (white edge reviews) and logo to replace the green box at the microphone

  The first thing I would do with this project is to convert 32-bit if you have more control over the colors. To create the matte luma that I would start by converting the image in black and white, as you did, but I would go a lot easier on the contrast control. I would probably use curves instead of levels. I would have also used adjustments red green and blue when you turn the copy of black and white images to darken the skin as much as possible.

  To achieve a fully black matte I can duplicate the black-and-white images contrast and the value of the copy to multiply. I would do this several times until most of the gray was gone. Then I would have added layers of black shape to fix everything that was causing problems. All layers of black and white and fill the forms would compound before and the demo would be used for a cache by approach. See the HC demo with 10 copies of the layer of black and white.

  I would then use this HC demo to create a cache of luminance on the original film. I blurs the original film about 8 pixels with fast blur then I would have some edges to work with the final model. See Matt model of track.

  The last step would be to put the matte model of track on the top images in a main model and add a little more blurry fast if necessary and curves and Simple Choker. I watch the Alpha channel of the model, the solo track matte layer, and then set the Alpha channel of the curves to clean any opening in the matte. I may return the RGB model control panel and set the layer of matte as a cache for the images alpha and adjust around simple neck and blur accelerated to even out the edges.

  I hope this helps. Here is a screenshot of a computer creating the matte the way that I would.

  

  The last step would be to follow the bug of mic using mocha and corners allows you to add your logo.

  If this had been my project and there is no blue screen I would have fired on the screen of the project green and changed the composition for the guy with the green face. Then I'd use technical color change to make the green face again after the kay were made. I would also use a mic (cube) bug which was a good amount of details in it. It is easier to get a good track when there are a few details in the shot that is when the surface is flat and ordinary.

  I hope this helps. Pulling a luma key when shooting actors on white background is much more difficult to inlay.

• Mask eyes blinking

  I'm working on a video project and I need to mask blinking on its end upward, is there an easier way to do this rather than the frame by frame?

  What you do are called Rotoscoping. It requires rarely you can set keyframes for each image. The trick is to use as few as possible. If I try to roto someone's eyes, I would use the AE built in Tracker to stabilize images by using the angle of one of the eyes because it would likely stay the same throughout the shooting. I then add a solid color (blue or red would work) and set the blending mode to screen so that I can be solid. I then start the first image and use the pen to create a mask covering the eye. I would then go down the timeline until the mask had to be adjusted and make adjustments. When the mask was exactly how I like it I would add a null value to the composition, then add an expression for the null value between the position of the anchor of the stabilized film null. Return to the first frame of the CTI and parenting then both layer roto and the film stabilized to null removes the stabilization of the passage and add movement to the layer of roto. In this example, I needed to set the exposure on the right eye so the roto layer was used as cache by approach to an adjustment layer. Even if there was a lot of movement of the head of the actors only a few pictures keys are required for roto. The first screenshot shows the only visible solid blue you can see the mask, the second shows the final result. The whole process takes about 6 minutes.

  

  

  As Mylenium says, if we saw your shot we would have a much better chance to propose an effective technique.