Return-Path: Received: from mail-yk0-f178.google.com ([209.85.160.178] verified) by media-motion.tv (CommuniGate Pro SMTP 4.2.10) with ESMTP-TLS id 5385988 for AE-List@media-motion.tv; Fri, 21 Feb 2014 17:47:58 +0100 Received: by mail-yk0-f178.google.com with SMTP id 79so6546243ykr.9 for ; Fri, 21 Feb 2014 08:48:04 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :content-type; bh=Q0XQrDzDcf7BBc+UngKl193k1vllFWnU2DTLKGBIbxc=; b=gKPmpK024NQQ3mkPfFi7K/l4J417kwBSO4/Gg4bOU4Br0XLqcArLpqXYyRdPKLqSM3 9j3ZdjnyCgNMuQ7f8d/LQ3k4u8U08jQiK3F2xk009qd7qhc5TMAO/Mi97KLTt4R7ka8V tHQBjqNN+Jjccnn58A1iTnOW22idhbjSlSatJzb49OOOVSE7xjfI1RxxalrbaQZ43kUX 3ARcMpqhe+bfnOt2F1/KEh4WEfubBqh1XIUb4cMK2UgM5Zsl4U07c7/PxDgJEB+Ky/yH TRFA6mBY7X93PNGuDesXZtJvLZUgBGQduR31s5FNb5EJoySYALUERMpaA9rmHaEqxZO5 3P6w== X-Received: by 10.236.7.231 with SMTP id 67mr12672151yhp.30.1393001284043; Fri, 21 Feb 2014 08:48:04 -0800 (PST) MIME-Version: 1.0 Received: by 10.170.160.6 with HTTP; Fri, 21 Feb 2014 08:47:23 -0800 (PST) In-Reply-To: References: From: Teddy Gage Date: Fri, 21 Feb 2014 11:47:23 -0500 Message-ID: Subject: 3D ribbon with Particular To: After Effects Mail List Content-Type: multipart/alternative; boundary=001a11c1fd24b509b104f2ed62fb --001a11c1fd24b509b104f2ed62fb Content-Type: text/plain; charset=ISO-8859-1 >So basically - you have 3 different 3D worlds and you want to composite them together. If your camera is either static or not moving much then you can do it manually by splitting layers, but if you have a dynamic camera move then it will get too hard, too quickly. Yes, this is what's giving me migraines. I do have the (wonderful) simple camera rig script from maaltanon / aescripts, and that allows for a referenced, precomped "global" camera, but I think experimenting with this in the middle of production is not worth the risk. I'm already using it with Sure Target, and adding in that extra 3D space could break the whole rig. As always I will probably just cheat the shit out of it, by limiting the amount of times the ribbon crosses both in front and behind a layer in 3D space. At least I can control that with the layer stacking index. I'm not sure I have time to build a depth pass, because I need to work out the ribbon moves alongside building the 3D layout. For that I am using Mathias' iExpression layer placement package. Definitely recommend that. (I'm not a shill for AEscripts, I promise). sParticular looks interesting, and works for the most part. My problem is it literally doubles the amount of layers within the comp and then actually scrambles the index. I need this stuff meticulously organized so I know what I'm doing and can respond to changes asap. Maybe for another project. On the plus side I re-discovered trapcode Mir for some lighter, gauzy flowing textile elements in the background and it looks great. Some very neat abstract controls, and it plays nice with the comp camera and lights. -TG On Thu, Feb 20, 2014 at 11:54 PM, Chris Zwar wrote: > I haven't used cineware but it sounds like you're heading towards headache > territory. What you're wanting to do is z-depth compositing and that's not > really easy in After Effects. Particular has its own 3D world and outputs > a 2D image, my guess is that's how cineware works too - there's the > chinaware 3D world which is different and separate to the After Effects 3D > world. > > So basically - you have 3 different 3D worlds and you want to composite > them together. If your camera is either static or not moving much then you > can do it manually by splitting layers, but if you have a dynamic camera > move then it will get too hard, too quickly. In theory each 3D world could > spit out a z-depth matte and you would use that to composite everything > together, but that's not simple. I haven't googled around but there are > probably a few tutorials around on how to fudge it - I know I tried a setup > once using the extract plugin to pull luma mattes - so maybe see what you > can find. > > The approach I suggested is a bit of a fudge that is really using > particular as a z-depth compositor. But that works with only 1 other > source, and if you're combining chinaware, after effects and particular > then you have 2. > > (BTW if your camera isn't moving much then it won't be too bad. I'm > thinking of a project I worked on which involved a 3D particular ribbon > being comped into a city flythrough, which was more involved) > > -Chris > > -----Original Message----- > From: Teddy Gage [mailto:teddygage@gmail.com] > Sent: Friday, February 21, 2014 02:50 PM > To: 'After Effects Mail List' > Subject: Re: [AE] 3D ribbon with Particular > > Nice Chris, I'm gonna try that. Will definitely give sParticular a look as > well. Would there be a way to integrate a 3D element from cineware into 3d > AE layers? dealing with hundreds of photos mixed with video is much slower > in cinema > -TG > > > On Thu, Feb 20, 2014 at 7:59 PM, Chris Zwar wrote: > > > Forget the obscuration layer, you need to create and use a Z depth map. > I > > do this a fair bit with particular within AE. > > > > There are 2 ways - you can buy a plugin (which is accurate but very slow) > > or you can use an expression, which is fast and free but a bit of a pain > > and not always 100% perfect (because the corners of a layer are a > different > > distance to camera than the centre). > > > > To do it in AE, you need to duplicate your main composition with all your > > 3D layers and make a Z-depth version. Strip out everything that isn't a > 3D > > layer and isn't going to interact with the ribbon. Add the fill effect > to > > each layer, apply an expression to the colour to fill it with a shade of > > grey based on the distance to the camera: > > > > > > camp=thisComp.layer("Camera 1").toWorld([0,0,0]); // camera position // > > lp=thisLayer.toWorld(thisLayer.anchorPoint); // layer position // > > d=length(camp,lp); // distance from camera to layer // > > > > min=0; // this this number for white in particular // > > max=20000; // use this number for black in particular // > > > > v=linear(d,min,max,1,0); // convert distance to value 0 - 1 // > > > > [v,v,v,1] > > > > > > To keep things speedy I would pre-render the z depth map, import it and > > use it with particular -use the same camera distance value in particular > as > > the min and max values in the expression. > > > > > > > > -Chris > > -----Original Message----- > > From: Teddy Gage [mailto:teddygage@gmail.com] > > Sent: Friday, February 21, 2014 10:58 AM > > To: 'After Effects Mail List' > > Subject: [AE] 3D ribbon with Particular > > > > So I am working for a fashion brand asking for a 3D ribbon effect > that > > twists through fields of photos and video, going in front of and behind > the > > images. My first thought was 3D stroke, but it didn't work, as I need it > to > > have some shading from the comp lights. There's also the issue of > painfully > > retiming the stroke moves using percentages of completion, and not > actually > > having full "3D" control. (I can't draw a path that goes in all > dimensions, > > just postcard transforms on the 2D mask paths, AFAIK.) I may however use > it > > for some simple bending and repeating elements. > > > > My next idea is to use a particular 2.0 layer with 0 emit velocity, > > and having the emitter tied to a null that I can then fully animate > through > > 3D space. So far this is working pretty well. Comp lights work, it is > easy > > to control. However I am having some of the classic issues with > Particular > > where it crosses other 3D objects in the scene and doesn't respect the > > stacking order. The ribbon is either completely in front of or completely > > behind the images. > > > > - now I could add 1 or 2 "obscuration layers" but it is not reasonable to > > do this for over a hundred layers individually > > > > - Would I need to precomp the entire layout, collapse transformations, > and > > THEN use the precomp as an obscuring layer? Is this even possible? It > would > > be a difficult way to work, I'd need to have every element layed out in > 3D > > first. Doable, but I'd prefer to create both at once. > > > > - is there another solution I'm not thinking of? A good script for > > Particular maybe? > > > > Thanks, > > -TG > > > > > > > > -- > > _____________________________ > > VFX & Motion Graphic Artist > > teddygage dot com > > > > > > > > +---End of message---+ > > To unsubscribe send any message to > > > > > > -- > _____________________________ > VFX & Motion Graphic Artist > teddygage dot com > > > > +---End of message---+ > To unsubscribe send any message to > -- _____________________________ VFX & Motion Graphic Artist teddygage dot com --001a11c1fd24b509b104f2ed62fb Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable
>So basically - you have 3 different 3D worlds and you want to composi= te them together. =A0If your camera is either static or not moving much the= n you can do it manually by splitting layers, but if you have a dynamic cam= era move then it will get too hard, too quickly.

Yes, this is what's giving me migraines. I do have the (= wonderful) simple camera rig script from maaltanon / aescripts, and that al= lows for a referenced, precomped "global" camera, but I think exp= erimenting with this in the middle of production is not worth the risk. I&#= 39;m already using it with Sure Target, and adding in that extra 3D space c= ould break the whole rig. As always I will probably just cheat the shit out= of it, by limiting the amount of times the ribbon crosses both in front an= d behind a layer in 3D space. At least I can control that with the layer st= acking index. I'm not sure I have time to build a depth pass, because I= need to work out the ribbon moves alongside building the 3D layout. For th= at I am using Mathias' iExpression layer placement package. Definitely = recommend that. (I'm not a shill for AEscripts, I promise).=A0

sParticular looks interesting, and works for the most p= art. My problem is it literally doubles the amount of layers within the com= p and then actually scrambles the index. I need this stuff meticulously org= anized so I know what I'm doing and can respond to changes asap. Maybe = for another project.

On the plus side I re-discovered trapcode Mir for some = lighter, gauzy flowing textile elements in the background and it looks grea= t. Some very neat abstract controls, and it plays nice with the comp camera= and lights.=A0
-TG


On Thu, Feb 20, 2014 at 11:54 PM, Chris Zwar <chris= @chriszwar.com> wrote:
I haven't used cineware but it sounds like you're heading towards h= eadache territory. =A0What you're wanting to do is z-depth compositing = and that's not really easy in After Effects. =A0Particular has its own = 3D world and outputs a 2D image, my guess is that's how cineware works = too - there's the chinaware 3D world which is different and separate to= the After Effects 3D world.

So basically - you have 3 different 3D worlds and you want to composite the= m together. =A0If your camera is either static or not moving much then you = can do it manually by splitting layers, but if you have a dynamic camera mo= ve then it will get too hard, too quickly. =A0In theory each 3D world could= spit out a z-depth matte and you would use that to composite everything to= gether, but that's not simple. =A0I haven't googled around but ther= e are probably a few tutorials around on how to fudge it - I know I tried a= setup once using the extract plugin to pull luma mattes - so maybe see wha= t you can find.

The approach I suggested is a bit of a fudge that is really using particula= r as a z-depth compositor. =A0But that works with only 1 other source, and = if you're combining chinaware, after effects and particular then you ha= ve 2.

(BTW if your camera isn't moving much then it won't be too bad. =A0= I'm thinking of a project I worked on which involved a 3D particular ri= bbon being comped into a city flythrough, which was more involved)

-Chris

-----Original Message-----
From: Teddy Gage [mailto:teddygage@gmail.com]
Sent: Friday, February 21, 2014 02:50 PM
To: 'After Effects Mail List'
Subject: Re: [AE] 3D ribbon with Particular

Nice Chris, I'm gonna try that. Will definitely give sParticular a look= as
well. Would there be a way to integrate a 3D element from cineware into 3d<= br> AE layers? dealing with hundreds of photos mixed with video is much slower<= br> in cinema
-TG


On Thu, Feb 20, 2014 at 7:59 PM, Chris Zwar <chris@chriszwar.com&= gt; wrote:

> Forget the obscuration layer, you need to create and use a Z depth map= . =A0I
> do this a fair bit with particular within AE.
>
> There are 2 ways - you can buy a plugin (which is accurate but very sl= ow)
> or you can use an expression, which is fast and free but a bit of a pa= in
> and not always 100% perfect (because the corners of a layer are a diff= erent
> distance to camera than the centre).
>
> To do it in AE, you need to duplicate your main composition with all y= our
> 3D layers and make a Z-depth version. =A0Strip out everything that isn= 't a 3D
> layer and isn't going to interact with the ribbon. =A0Add the fill= effect to
> each layer, apply an expression to the colour to fill it with a shade = of
> grey based on the distance to the camera:
>
>
> camp=3DthisComp.layer("Camera 1").toWorld([0,0,0]); // camer= a position //
> lp=3DthisLayer.toWorld(thisLayer.anchorPoint); // layer position // > d=3Dlength(camp,lp); // distance from camera to layer //
>
> min=3D0; // this this number for white in particular //
> max=3D20000; // use this number for black in particular //
>
> v=3Dlinear(d,min,max,1,0); // convert distance to value 0 - 1 //
>
> [v,v,v,1]
>
>
> To keep things speedy I would pre-render the z depth map, import it an= d
> use it with particular -use the same camera distance value in particul= ar as
> the min and max values in the expression.
>
>
>
> -Chris
> -----Original Message-----
> From: Teddy Gage [mailto:teddygage@gmail.com]
> Sent: Friday, February 21, 2014 10:58 AM
> To: 'After Effects Mail List'
> Subject: [AE] 3D ribbon with Particular
>
> =A0 =A0 =A0So I am working for a fashion brand asking for a 3D ribbon = effect that
> twists through fields of photos and video, going in front of and behin= d the
> images. My first thought was 3D stroke, but it didn't work, as I n= eed it to
> have some shading from the comp lights. There's also the issue of = painfully
> retiming the stroke moves using percentages of completion, and not act= ually
> having full "3D" control. (I can't draw a path that goes= in all dimensions,
> just postcard transforms on the 2D mask paths, AFAIK.) I may however u= se it
> for some simple bending and repeating elements.
>
> =A0 =A0 =A0My next idea is to use a particular 2.0 layer with 0 emit v= elocity,
> and having the emitter tied to a null that I can then fully animate th= rough
> 3D space. So far this is working pretty well. Comp lights work, it is = easy
> to control. =A0However I am having some of the classic issues with Par= ticular
> where it crosses other 3D objects in the scene and doesn't respect= the
> stacking order. The ribbon is either completely in front of or complet= ely
> behind the images.
>
> - now I could add 1 or 2 "obscuration layers" but it is not = reasonable to
> do this for over a hundred layers individually
>
> - Would I need to precomp the entire layout, collapse transformations,= and
> THEN use the precomp as an obscuring layer? Is this even possible? It = would
> be a difficult way to work, I'd need to have every element layed o= ut in 3D
> first. Doable, but I'd prefer to create both at once.
>
> - is there another solution I'm not thinking of? A good script for=
> Particular maybe?
>
> Thanks,
> -TG
>
>
>
> --
> _____________________________
> VFX & Motion Graphic Artist
> teddygage dot com
>
>
>
> +---End of message---+
> To unsubscribe send any message to <ae-list-off@media-motion.tv<= /a>>
>



--
_____________________________
VFX & Motion Graphic Artist
teddygage dot com



+---End of message---+
To unsubscribe send any message to <ae-list-off@media-motion.tv&g= t;



--
=
_____________________________=
VFX &= amp; Motion Graphic Artist
teddygage dot com
--001a11c1fd24b509b104f2ed62fb--