69 replies to this topic

[SecondMan]

dts74, on Apr 21 2006, 04:05 PM, said:

Thanks SecondMan!

Thoght I would chip in with some smallish macros..
mDiv
mMult

And her's a pic showing it on use i fusion, Left without (and without pre divide/postmultiply options, they would have to be used on bothcolor nodes to provide correct result) and Right the Shake way...take your pick the fg element is premultiplied semi transp fantastic spheare

Oh and they both have the exactly the same cc corrections, and that black edgeon the spheare is from the render..default light

<{POST_SNAPBACK}>

Great, thanks!

Could you add a bit more functionality to these tools, like the ability to use an effect mask and a blend slider?

Cheers.

[dts74]

Sure...here you go.

Attached Files

•  mDiv.txt   1.74K   90 downloads
•  mMult.txt   1.01K   67 downloads

[SecondMan]

Cool. Here's my last one for today. Got some more lined up but can't get them to work in a macro. Waiting for tech to get back at me for those.

Cheers.

Attached Files

•  SM_Tile.setting.txt   3.29K   62 downloads

[dts74]

And as with all preMult Img with sharp highlight it runs the risk of getting clipped, and will get clipped in both using the fusion way or Shake way, solution promote premult img to float 16 or 32 do your cc then unpromte it back to 8 or 16 int if you like or keep it. Or better render highlights in separate pass

I missed it on the previus img...oh well

[force_redo]

Wow. You guys are great. Your faster than I can update my original post! (Thanks god it's weekend...)

SecondMan, on Apr 21 2006, 01:25 PM, said:

LookupHLS etc...

[W01]You could simply convert the Color Space and use the CC tool in between them to do that...

Yes, that's what I actually meant by "sandwitching". Put a curve tool between two colour space conversions...

Quote

MDiv

[W01]Custom Tool

MMult

[W01] Custom Tool

ISubA

[W01] Custom Tool

Hmmm. You're right technically. Thing is that the Custom Tool is so powerful that I could virtually put it in as a workaround for everything. So I didn't do it, since I thought that was obvious. But then, if somebody is desperate and wants to look it up here, it might be a hint. I don't know, what do you think?

Quote

ContrastLum

Read your description and that's exactly what my macro should do actually... could you verify that?

I double checked it, it's almost there. Only it seems to contrast the blue channel (saturation component) instead of the green (luma)

Quote

Xor

[===] Merge -> Xor

Doh! Missed that one. Too obvious for me

Some comments about the macros:
First of all: Thanks a lot, once again. When you're finished we can do some Gui hacks and call it "Fake" (or "Shusion")

Rand:
The colour distribution seems a bit low. I.e. if I blur the heck out of it, I get an average colour around 0.3. I would expect it to be around 0.5, just to verify the noise is evenly spread. Furthermore it would be great if it would generate some noise in the alpha, too. I know I could copy it there, but...

PercentBlur:
There's something strange there. It works great in integer, but with float it gets "pixelated"

IBlur:
Can't really understand what it's doing, sorry. Seems to be a Glow in there?! The shake IBlur is really quite close to the DeepBlur, only the DeepBlur seems to have some issues. The Falloff doesn't seem right and at extrem blurs it creates a hole in the matte. (How can I upload files here, I could probably upload an example comp)

Color Replace:
Was great! Is now even better!

Clamp:
Does the right thing, only in the gui it doesn't say which slider is for which channel. I believe the order is ABGR now?!

Checker Plus:
Just got better, too. (Only the alpha doesn't work anymore)

Tile:
Yet another bull's eye...

Wow! and again: WOW!

FR

[force_redo]

dts74, on Apr 21 2006, 05:18 PM, said:

Sure...here you go.

<{POST_SNAPBACK}>

Thanks a lot for this. I was missing this desperately. Call me old fashioned or stubborn, but I think it's an improvement when you're putting together CG passes.

Cheers,
FR

[SecondMan]

force_redo, on Apr 21 2006, 11:03 PM, said:

IBlur:
Can't really understand what it's doing, sorry. Seems to be a Glow in there?! The shake IBlur is really quite close to the DeepBlur, only the DeepBlur seems to have some issues. The Falloff doesn't seem right and at extrem blurs it creates a hole in the matte. (How can I upload files here, I could probably upload an example comp)

<{POST_SNAPBACK}>

Well, I realised that in Fusion the Glow Tool with the Glow parameter set to zero does exactly the same as a Blur. With the crucial difference that a Glow actually has a Glow Mask input on top of an Effect Mask input, which should allow it to do exactly what you were describing as what the IBlur does. I guess the only way to check this is for you to play around with it and see whether the visual result is the same...

Will check into the others on Monday and post another update. We'll get there

[SalaTar]

After reading this thread more we really do need a way keep track of macros/scripts and variations on them on the web

[SalaTar]

You think a wiki with a bugtracker/cvs and some sort of document tracking would do?

[SecondMan]

SalaTar, on Apr 22 2006, 03:32 AM, said:

You think a wiki with a bugtracker/cvs and some sort of document tracking would do?

<{POST_SNAPBACK}>

That would be a very nice start!

No idea how to set that up though...

[Daniel Koch]

Nice to see some useful macros being made :-)

In the same spirit, I've done some refinements to the Checker and Tile macros, attached below.

Checker uses a lot less tools (and is faster), and uses a few simple InTool scripts instead of Publish & Expression (more friendly for animating values). It also has Global In/Out controls, is a bit cleaner, and doesn't have the 1831 x 1153 size limit anymore. Another trick I used was to rearrange the macro's inputs in a text editor, for more logical grouping.

Tile actually uses more tools because I added a "mirror" option, though a couple of InTool scripts made the original tools a little simpler. It has more transformation options too.

I'm also attaching the source comps, for those that would like to see how things work. I've added comments to most tools.

Attached Files

•  CheckerPlusPlus.setting.txt   7.68K   68 downloads
•  TilePlusPlus.setting.txt   7.35K   57 downloads
•  CheckerPlusPlus.comp.txt   3.74K   55 downloads
•  TilePlusPlus.comp.txt   6.65K   55 downloads

[Daniel Koch]

Here's mMult and mDiv too, done with a single Channel Boolean (which is much faster than the Custom tool) - you just set the operation to Multiply or Divide, and the RGB channels to be operated on by the Alpha channel. I added a couple more masking & channel options as well.

For simple pixel math, it's almost always faster to use one or a series of Channel Boolean tools than a single Custom tool (even if you have to convert to/from float32 as well).

Attached Files

•  mMult.setting.txt   1.83K   156 downloads
•  mDiv.setting.txt   1.81K   122 downloads

[mrrafs]

dts74, on Apr 21 2006, 10:05 AM, said:

Thoght I would chip in with some smallish macros..
mDiv
mMult

<{POST_SNAPBACK}>

-why do the renders look different? as long as one presses the pre-div button in fusion they should look identical?

-also why do it the shake way in fusion, when it makes for a longer render?

--
Rafal

[mrrafs]

Whats shake equivilent to a subtractive merge? or a blend of an addiditve merge(over) and a sub one?

whats the maths for a sub merge?

anyone know?
ta
--
Rafal

[force_redo]

mrrafs, on May 2 2006, 10:01 AM, said:

Whats  shake equivilent to a subtractive merge?
or a blend of an addiditve merge(over) and a sub one?
whats the maths for a sub merge?

I believe (because the Fusion Manual doesn't say, as far as I know) the additive merge (default setting) is a "normal" over of a premult'd FG.
i.e. fgRGB + (1-bgA)*bgRGB
Furthermore, I think the "subtractive method is the same for unpremult'd FGs
i.e. fgRGB*fgA + (1-bgA)*bgRGB

But this is really all just guesswork. As soon as I'm back home, I can try it out...

FR