configuring linear workflow

for linear workflow i am setting the loader gamma space to itu-r_bt.709 and ticking remove curve

then in the display view globals i am enabling the Buffer LUT and loading the OCIO File Transform ViewLUT edited to use rec709.spi1d file and setting 'direction' to reverse

then before the saver i add a OCIOFileTransform node configured the same as the global veiw LUT

is that correct so far? everything looks ok

next i am enabling 'remove curve'->srgb from all the image creator nodes?

is that correct?
is there a way to do that globally?
also, what about 3d lights. do i need to run those through a gamma nodes?

thanks

Let me ask you a question? Are you grading the content in Fusion or you are just compositing stuff and then plugging it back into conform pipeline like Resolve or Scratch?

make sure you are talking the same language.
Color industry refers to linear as flat gamma, and compositors as linear light.

Let me ask you a question? Are you grading the content in Fusion or you are just compositing stuff and then plugging it back into conform pipeline like Resolve or Scratch?

Doesn't matter as far as compositing is concerned
proper linear workflow is required for DOF, blending, motion blur and compositing CG elements together.

It should be used within compositing regardless of output...not mandatory of course, but recommended to remove artifacts from blending operations, and realistic lighting.


Stephan has a pretty good write up on the workflow,
http://www.comp-fu.com/2012/02/linear-g ... in-fusion/

As far as the 3d...it can output linear, set the render to 32bit float.

Fusion's 3D renderer is linear regardless of output format.

Fusion's 3D renderer is linear regardless of output format.

one common mistake is to light your scene without a lut. It will come out linear, but completely overexposed for view luts.

There will only be compositing but my concerns are the ones Mike Harrington mentioned.

Thanks for the link Mike.

Stefan Ihringer's writup was helpful (thank you Stefan) but it sortof deepened my confusion. Should i use his shaders? Are they different than using the saver node's 'remove curve' with itu-r_bt.709 (scene) selected and then using the rec709.spi1d file with the OCIO File Transform in reverse?

If i add the view transform will swatches be color managed?

Thanks Pieter and Rony. Sorry for my clumsy use of compositing terminology.

I use his shaders as viewer LUT's

I use the remove gamma like you are in the loader, view through the LUT, and then at the end of the chain just add a gamut tool and go into the saver.

There is probably quite a few ways to do it properly, just remember: remove gamma(if needed) use a LUT to view, re-apply gamma at the end...or not depending on output.

It's more important to understand the end goal, rather then a specific workflow.. as the process can hit some snags

For instance you run into a whole bunch of issues with LUT's and superbrights....they cannot be mapped properly.

I posted some links to R3D conversion LUT's and a description of that workflow here
http://www.steakunderwater.com/wesuckle ... f=16&t=258

Basically in this workflow....if you want to preserve the RED gammas.... you are linearizing the data, doing your comp work, converting to LOG, apply a LOG to target gamma LUT, then saver

The LOG conversion does a normalization....so the LUT can then be mapped properly.

Color swatches and others have some issues....you will have to defer to Stephan's long list of complaints/suggestions here
viewtopic.php?f=22&t=29507&p=184694&hilit=linear#p179404

ok Mike. Thanks for the workflow tips, red luts and the heads-up about clamping. Working my way through Stefan's lwf post now. should have read that first. I'll be sure to search this forum and steakunderwater before posting.

just learning my way around the world of compositing/fusion and its community now. hopefully it will not be too long before i can make a contribution

Thank you all very much:)

Stefan Ihringer's writup was helpful (thank you Stefan) but it sortof deepened my confusion. Should i use his shaders? Are they different than using the saver node's 'remove curve' with itu-r_bt.709 (scene) selected and then using the rec709.spi1d file with the OCIO File Transform in reverse?

My blog post is rather old, some things were changed in Fusion 7. But I still think my shaders make things easier because you get sRGB/rec709 entries directly in the LUT menu.

I've never used OCIO transforms for the "simple" things like sRGB and rec709. I think the curve is identical to Nuke's which is probably identical to Fusion's or my view shaders.

You've mentioned "remove curve" in connection with the Saver node... just to be sure: did you mean Loader?

The basic linear workflow would be this:

• add Loader with footage
• select appropriate gamma space or trust "auto" setting
• check "remove curve" option
• add viewer LUT and comp away
• add Saver
• select appropriate gamma space for output footage (the "keep" option relies on metadata from your Loader and is only valid if your output format is the same as your input format. If you load exr files but want to render jpegs, for example, you need to select your sRGB output space manually unless I have missed some kind of "auto" setting for Savers...)
• check "apply curve" option

If i add the view transform will swatches be color managed?

No, the color swatches are not affected by viewer LUTs (neither their appearance nor their inner workings).

yes i misspoke when i said saver.

i will use your shaders and the gamut node for output. That combination works better because when i use the OCIO as a global view transform and the OCIOFileTransform node as the output transform they lose their respective file connections after the comp has been close. Your method is faster and works across sessions.

thank you:)

just learning my way around the world of compositing/fusion and its community now. hopefully it will not be too long before i can make a contribution

Thank you all very much:)

Steakunderwater is a good source...I've learned a lot from Stephan, Pieter, Chad and others....
Some really clever guys there.

Clever yes, but they are here too!

One thing I wish would work right with linear work flow would be to have the Shadow, Midtone, and Hilite regions in the CC have a setting for working correctly with linear work flow. Those regions are designed for gamma 2.2.

You can either adjust the ranges yourself, kind of guessing where they ought to be (and save as default), or you can stick a BC before and after the CC node. This seems like a lot of extra adjusting, though, and may eventually posterize a 16bit float image.

One thing I wish would work right with linear work flow would be to have the Shadow, Midtone, and Hilite regions in the CC have a setting for working correctly with linear work flow. Those regions are designed for gamma 2.2.

You can either adjust the ranges yourself, kind of guessing where they ought to be (and save as default), or you can stick a BC before and after the CC node. This seems like a lot of extra adjusting, though, and may eventually posterize a 16bit float image.

Adding to this: while you can set up ranges that work better in linear, one thing you can't do is move the highlight ranges beyond 1.0. Being able to do so would be really important for adjusting highlights in linear images.

When you need to dull specular highlights that reach values of 5 or higher, it's no good treating everything above 1.0 the same way.

We typically use a BC to scale things down, or a Curve. But, yes, the CC really needs to be updated to work in linear space more transparently.

We typically use a BC to scale things down, or a Curve. But, yes, the CC really needs to be updated to work in linear space more transparently.

Right, that's what I do too, but normalizing to a 0-1 range is a workaround at best. Having a real unclamped workflow would be so much nicer.

Which is odd since it's really just a UI problem. It's not like Fusion doesn't handle the data, it's just that the tools don't. And yes, I understand that things like Gamma are what they are, but the ranges in CC should definitely work better.

And yes, I understand that things like Gamma are what they are, but the ranges in CC should definitely work better.

LOL. Chad, to say that you understand gamma is like saying Wernher Von Braun understood physics! You have more than a basic grasp of it

Good to see familiar faces.

for linear workflow i am setting the loader gamma space to itu-r_bt.709 and ticking remove curve

then in the display view globals i am enabling the Buffer LUT and loading the OCIO File Transform ViewLUT edited to use rec709.spi1d file and setting 'direction' to reverse

then before the saver i add a OCIOFileTransform node configured the same as the global veiw LUT ...

In our workflow all shots go to a grading suite after VFX has done their work.

Our working practice is to preserve the camera colour space all the way through the pipeline until the grade so.. savers match the loader Not the viewerLUT

Loader and saver nodes are set to the same colour space. Log->lin for the loader means Lin-> log for the saver. sRGB -> lin for the loader would mean lin-> sRGB for the saver.

When working with Log footage it does not make sense to set the savers to rec709 since dynamic range is lost and options in the grade become limited.

The viewer lut is normally set to the monitor colour space. A sRGB monitor needs a sRGB viewer LUT. A calibrated rec709 monitor needs a rec709. Most off the shelf monitor are sRGB so unless I know its been calibrated I'll use a sRGB viewer LUT

If possible we'll try and get a LUT from the grading suite so we know where going. In which case there a few techniques for applying that to the viewer LUT , but the loaders and savers would still be set to the original camera colour space.

Our work flow is written up on my site. Its based on the sony image works pipeline.

arg I want post a link but the board permissions are preventing from doing so

And yes, I understand that things like Gamma are what they are, but the ranges in CC should definitely work better.

LOL. Chad, to say that you understand gamma is like saying Wernher Von Braun understood physics! You have more than a basic grasp of it

Good to see familiar faces.

No, I just mean that 1^ɣ is still 1. Beyond 1, gamma does the opposite of what it does between 0 and 1. Whereas gain, brightness, saturation, contrast, etc. behave more comfortably through those unclipped areas.

There are other tools that are also unfriendly, not just CC and CCv. SGlo, surprisingly, has it's threshold clipped at 1. I don't have a complete list of tools that would also fit this category.

The basic linear workflow would be this:

• add Loader with footage
• select appropriate gamma space or trust "auto" setting
• check "remove curve" option
• add viewer LUT and comp away
• add Saver
• select appropriate gamma space for output footage (the "keep" option relies on metadata from your Loader and is only valid if your output format is the same as your input format. If you load exr files but want to render jpegs, for example, you need to select your sRGB output space manually unless I have missed some kind of "auto" setting for Savers...)
• check "apply curve" option

if we load 8bit jpg's and tiff's into Fusion is ''auto'' the same as setting it to sRGB? it would be great because than we would only need to check ''remove curve''.