[Steve Bentley]

Hey Walter,
Re the RGB blues: I just figured in this day and age the 709 variant in use would be the 10 bit one and not the 8 bit one (i’ve never been sure how this can be the same spec if you have both 8 and 10 bit available inside the same one). So I simply meant the 10bit 709 will always deliver more color variation in color than the 8 bit sRGB. Given the poorly lit, poorly conformed, incorrect codec (add your own malady here) that we get from film crews these days, we work in the highest bit depth/space that makes sense simply to have the head room to color correct things. (as well as take advantage of the non image channels in CGI elements). While the output we’re talking about in this post is for YouTube, where sRGB would be fine, that doesn’t mean the project should necessarily be in lowly sRGB 8bit. I find you end up clipping things when you start pushing the curves around.

The monitor comment of mine was, in retrospect, overly simplified – what I meant was if you have a 10 bit monitor you are missing out when in working in an 8 bit space. And conversely if you have an 8 bit monitor, don’t pull your hair out when all those lovely smooth grads don’t show up that are generated by your 10 bit space. (again sRGB vs the 709 10 bit variant – I’m afraid all my assumptions were based on using the 709 10 bit)

Because we can choose the kind of file we output to in AE (and therefore it’s inherent bit depth and in some cases the inherent color space those formats are restricted to), are you saying that the color space in AE, which is of course defining the envelope you are working in, is also passing that color space data to the output file and affects external representation on another system, another monitor, another environment? (that truly was a question by the way not a rhetorical chest poke). In kinda the same way that photoshop can embed a profile?

You are absolutely correct when you say the sRGB and 8bit 709 are virtually interchangeable in the chroma capability. However (given the gamma questions in this and other recent posts), sRGB is spec’d at 2.2 whereas I always see 709 refereed to as 2.4. Doesn’t that affect a screen delivery difference?

[Steve Bentley]

It may have to do with your audio hardware. 48k audio needs special harware to sync to video. Pretty much any computer can do 41k. This also goes for programs. Some handle decoding 48k better than others.
Try is at both and see what happens.

[Steve Bentley]

In a 3D package everything is relative – you can enter all those dimensions you have and you will be bang on (assuming the data is correct). Export the camera from the 3D package along with nulls for object position and height and you can match those markers in AE when you import all that data.
But if you want to use a 3D model inside AE, you can’t say to the model, be six feet high. As Dave said you have to say “be 359 somethings high” (AE can’t natively use 3d models, there is always a plug in, even a built in one, and that may not translate to pixels).
But it can do renders of models – so once its rendered in a 3D package all the hard stuff is done.
But often eyeballing it is better. Lens distortion can screw up lots of measurements because AE won’t automatically take the distortion into account.

And Mark may have said this – AE isn’t real time so it’s not the best tool if you are live keying a model into a stage shot in front of an audience (it wasn’t totally clear if this was the case). And you might have a killer graphics card that can get close to real time but any type of system activity can interrupt that (like autosave).

[Steve Bentley]

With the Motion 2UI.
Just watching the demo it looked he selected a group of keyframes, changed the value of them all, then selected another set, and changed those. But when he selected a set from the original group the “velocity” as he calls it didn’t reflect the initial value but the value he just set on a the second set of keyframes. So it looked like it was a bit of a one-way street – you can set values but not read the values.

[Steve Bentley]

If the piston and wheel are graphics in the comp you can simplify the whole thing by just parenting one object to the other (with an offset anchor point) and having the animation of one drive the other. Depending on your set up you may have to do an expression that uses only one of the pistons position’s attributes (x or y) so that the rotational motion of the wheel doesn’t make the piston angle.

You could also use an IK chain. You can do a simple script to make this (hint the child is always “looking” back at the parent) or you could use a free resource like DUIK to set that up for you.

[Steve Bentley]

I’m not a PP guy so I’m not sure. but you could render out with an uncompressed or lossless format. AVI uncompressed (big), Quicktime Animation Best (can be crazy small – it’s run length encoded, so things with lots of the same colors are packed up more efficiently – totally uncompressed), a series of PNG or TIFF or Targas, or a series of EXR or Cineon files.

[Steve Bentley]

I think I see where you are going now. The issue is going to be changing the texture images over time.

The simplest way, is to do what we’ve got working already (take the texture from my file I sent and bring that into your project – the one you had in the last project didn’t have the Mograph color shader in the alpha channel so your random effector wasn’t doing anything – the noise needs to move on the random effector so that the alpha changes over time).
But instead of images in the mograph multishader use movies of the images changing – that way you can get the transition effect you are looking for as the images change. A different movie for each texture slot or use the animation features in the texture tag to shift the animation – map the animation to a different set of keyframes for each of the mograph mutlishaders channels.

If you are OK having the images pop when they change then you could use the variation shader (C4dR17 and up I believe). Load the images (not movies) into the variation shader in the color channel just like you do with the Multishader (Add from folder) and then animate the probability setting for each of your images loaded. (one keyframe at the beginning and one at the end of the entire animation). This will cause the images to change over time (among all you have loaded). but they will change instantly or pop.
Keep the alpha of this material set to the mograph color shader just like we’ve been doing and that will look after your random transparency (or you could use the variation shader on the alpha channel and just do a grad from back to white instead of images and animate the grad or the random color slot)

There’s one other way – when you select the texture tag on an object, you can keyframe that material in the attributes manager. With the original material listed in the Material slot of the attributes of a texture tag, hit the keyframe button, go to another keyframe and drop a new material in that slot, hit the keyframe button again. The textures will dissolve from one to the other between those keyframes.
Now that you know how to do that, open each of the textures in the multishader – twirl down the black triangle next to the Texture 1 (2,3,4 etc) and keyframe the file slot – then move to another keyframe and get another file from your drive and keyframe that, Repeat as necessary. Unlike the texture tag on an object that’s keyframed, these will pop as they change over time.

Because the length of the object that is spline-wrapped will change over time the textures are going to stretch. Because you are spinning the spline you will see the textures edge-on sometimes – this may or may not show due to your spiral. I’m not sure why this is needed. You could have the swept objects along the traces always face camera. Remember much of the new tech in C4D is just old tech in a new wrapper – tracers and objects are just paths swept along a spline. Mograph is just a wrapper on what we could always do in expresso. So many of the old techniques apply to the new features.

Finally you might want to tap Adam Trachtenberg and Brian Jones in this forum as they have done a bunch in this area as well. The issue is the changing textures over time – at least I think that’s what you are after.

I thought perhaps we could have two shader effectors on the cloner- one generating a black and white noise and one generating a color noise – then we could have the transparency be affected by the B&W noise and the image chosen driven by the hue of the other noise. I haven’t been able to make that work yet.

There’s another way altogether that might be more organic. Have your points emit a few TP particles every 10 frames or so, short lifespans and with no velocity so they are just left in the characters wake. Then randomly, you could join those particles with a spline – like the plexus effect and sweep an object along those connect lines – You might even be able to do proper lightning if you made a connection algorithm that used the emission order or particle ID so you could branch the connecting lines.

[Steve Bentley]

First, in this day and age, sRGB is a horrible color space. As I’ve complained before in this forum, there aren’t even enough blues to do a smooth grad from one side of an HD frame to the other.
Its best to work in a color space that is at least equivalent to the one you will output to or preferably one that is more spacious.
You can always work in a “none” color space in AE (good) or the ACES space (better, but might be overkill) and choose a 16 or 32 bit composition and then choose your output file format. But it depends on the monitor you have. You don’t want to building in a correction that might fight with user’s monitor brand X.

I’ve been watching the back and forth with 709 conversation. I’m pretty sure that technically HD is 2.2 gamma not 2.4. (Even though some broadcasters are requesting 2.4). HD was supposed to fix all this but instead of one format we have, what, 19 now?

The problem these days (as in non CRT days) is that everyone has their screen tuned up to… well they aren’t tuned up to anything and depending on the manufacturer, the light pack in behind can produce a different set of blacks – even a different set of blacks in different areas of the screen. And don’t even get me started on OLEDs.
Its best to talk to your broadcaster or your post house and see what their specs are. I’ve got three networks that we send to regularly and while their specs should be identical, they aren’t.

DSC labs makes some terrific HD bars files that are tuned to make a scope sit up as straight as a well trained guard dog. (And they are industry standard)
It used to be you could use a scope and rely on that. But I find now that you have to use a scope to make sure you are technically correct and accurate and then use a high quality bars file (with pluge please) and match what your monitor is putting out, by eye if you can’t afford a light sensor puck.
Now you might be saying – “but we’re compressing to H264”. That’s great for output but when creating you need more headspace to color correct or do effects with multichannel elements that need a 16 or 32 bit space to render the colors correctly (for things like depth or blur passes).

When it gets right down to it you need to make sure your screen is WYSIWYG regardless of LUT. With a pluge or elements you make on your own (with lots of detail in the toe or the shoulder of the color space curve), the difference between a 2.4 gamma and a 2.2 becomes glaringly obvious.

Remember, much of the color space talk has nothing to do with the files themselves, it more about how the file is delivered to the screen.
I personally wouldn’t want to embed a color space in a file because that may alter the things you want to do with it later, or restrict what a user can see who’s screen is set up for another color space.

Unfortunately color space is a deep subject (wait till you get into Logarithmic space and LUTS!) and there are so many myths out there now due to the web being a wild west of formats. And I guarantee that 90 percent of the screens used for watching youtube are not set up properly. So we are all kinda trying to craft a Faberge egg for people who don’t care.

But take heart, even with my firm’s OCD approach to color, color space and output formats (EXR) and the TV’s I have at home (one plasma and one LCD) that I’ve tried very hard to tune perfectly to standards, I am always horrified when I see stuff we have created being broadcast (over air or via chromecast). It is a shadow of its former self after recompression, low compression quality or Pedestal that got added by some over eager tape jockey. You are not alone in your frustration.

So for web stuff here’s what we’ve done.
Create some content with colors that really speak to you (so you know them when you see them) and lots of details in the 0-5 percent grays and 95 – 100 percent whites. Put that up on the web in the compression you normally would. Go to a local computer store and check out that link on lots of different machines and monitors – make sure to include some Mac retina screens. Then tweak the content or your work screen as necessary so that it looks in the store like you saw it on your screen.

[Steve Bentley]

I can take a look in an hour or so. In the meantime, I would switch to the Shader effector – it can do everything the Random can do but with more flexibility. (in fact since the shader effector came along the random effector is redundant).
Second, you can use multiple shader effectors on the same cloner to get at different attributes. Changing the order in the list of effectors in the cloner determines which one works “first” or is “on top” and you can increase or decrease the intensity of the effectors there as well so you can blend them.

[Steve Bentley]

Ok I’ll bite. If we can’t access the CR state from an expression, can we really control the CR state from a check box? Wouldn’t that by its very nature require an expression?
Is there a trick?