Hi all,

hope this dont sound silly.

I understand users define what 1 houdini unit is, can be in metres, centimeters and so on. Say i have a prop with a diameter of 5cm, a sphere have a radius of 0.025 which makes it insanely tiny.

Is this going affect the lighting, texturing or rendering process, is it be advisable to treat 1 houdini unit as 1cm instead to work with bigger props within the viewport?

Another doubt is shown inside the viewport, i highlighted 2 boxes on the construction grid is equals to 1 houdini unit - am I wrong?

Please! :?

eitht.

1Houdini unit = 1meter. That is the default if you want to use something else remember to adjust your simulation values in Pops , Dops etc

Rob

As long as you don't have to deal with matching real world's focal lengths/disctances, or exchange data with other apps were it could matter, the only thing I know of that's more or less scale-dependend in H is: DOPs and POPs

This doesn't mean, you can't work with other scales, but it's much easier, because standard values (e.g. gravity), parts of the tools shelf, … are optimized for 1 unit = 1 meter

Appreciate both your replies!

i'm matching a still photo with local length.
There's no DOP/POP involved.

Is it safe to say i will not encounter any lighting/texturing problems if i use the default of 1 unit = 1 metre?

eitht.

remember to strip the gamma from your textures !

Rob

Rob, if you are willing to answer another question…
All along I de-gamma by setting 0.455 gamma to my texture images in Photoshop, is this the proper way or do one need other editing softwares to perform this task?

Thanks in advance!

eitht.

I think the best you can do is convert your files to RAT files and you can specify gamma of the texture while converting it, therefore you will have correct linear RAT files you can directly use in all shaders, so no need for other software

other thing you can do instead of converting is append Power VOP after each sRGB texture in shader and set it to 2.2 which has the same effect as setting gamma to 0.45 (RAT files might be better though)
and of course don't forget to have linear color swatches as well, but this is handled automatically if you have Edit/Color Settings/Color Correction set to 2.2 and checked everything

I strip out the gamma in the shader. ive found it will do 99% of the time

Rob

it will do, but converting to RAT has many advantages (like MIP mapping and random access from disk), so if you decide to do this conversion you may as well ensure they are linear and you don't need to modify shader in any way, if you don't need this optimization however, you can simply just modify the shader

tamte
of course don't forget to have linear color swatches as well, but this is handled automatically if you have Edit/Color Settings/Color Correction set to 2.2 and checked everything

You referring to the color swatches in Houdini? By default, Color Correction/Gamma was set to 1.0 and that makes it non-linear?

Trying to wrap my head around the whole LWF, here's a screen-cap of what I did to the de-gamma. Does the renders look like they were done correctly?

This is what I would work in the future,

1) De-gamma textures
2) Render stuff out with linear color space in my mantra rop
3) Render out my final composite with sRGB color space in NUKE

Anybody see a problem or am I on the right track now?
Apologize for being long-winded!



eitht.

eitht
…
You referring to the color swatches in Houdini? By default, Color Correction/Gamma was set to 1.0 and that makes it non-linear?
…

well in a sense of WYSIWYG yes, it's non-linear
since you set your color swatch to color you see on monitor and you want to have the same color in final picture
if that setting is 1, then the color you see on monitor is not the color represented by numerical values set on the color parameter, since monitor has gamma 2.2 and therefore changes that color, basically it's the same as for textures.

so if you set Color Settings to correct all outputs to 2.2, then you will see the swatch with Gamma 2.2 applied on top to compensate monitor, which keeps the value of the color parameter linear, because you see directly the color which is represented by numerical values

EDIT: so if you are going to use linear workflow in Houdini, set that setting to 2.2 (or whatewer gamma is your monitor) and keep all checkboxes checked, then the only concern is to make sure all your textures are linear or make them linear in the shader.

And if you are not rendering to linear formats (like EXR or other floating point data), you can bake in the gamma 2.2 to your renders by adding vm_gamma property to mantra ROP, in that case you would need to make sure you are not having gamma set on Mplay or render window since it will be already baked in

I assume the color settings and swatches greatly affects how the virtual lights behave too huh? :?

eitht.

they don't change rendering behaviour at all, it actually doesn't matter how you set that setting, mantra will still compute the same result given the same input values

it just changes how you see the values, so setting all to 2.2 gets rid of monitor gamma correction so you see directly linear values and therefore set them correctly for mantra, which can give you much realistic result compared to feeding mantra with false values and expect correct result.

but it's all the same principle, just search for linear workflow on the internet, plenty of info there, it's not specific to houdini

btw, saving an image from Mplay which has correction 2.2 doesn't automatically bake that correction into the files (there is an other option for gamma when you chose to save the frame) so you probably saved images (in your updated post) without baking the gamma in, because neither of them looks correct, but if you apply gamma 2.2 on the second one, it may be correct.

I am understanding this much better now!

Baking the gamma in is this checkbox in the screen-cap?
That's what i did when i saved both of the images above.

I was told simply typing 2.2 into mplay doesn't bake it but setting this when saving out images bakes it.

eitht.

it just changes how you see the values, so setting all to 2.2 gets rid of monitor gamma correction so you see directly linear values and therefore set them correctly for mantra, which can give you much realistic result compared to feeding mantra with false values and expect correct result.

I leave all that at default and strip the gamma from textures in the shader refer to my site (http://www.circusmonkey.com.au/circusmonkey/Shops.html) [circusmonkey.com.au] . I then render and in the mplay display and set my gamma to 2.2. What I see in mplay on my monitor it whats important and as long as there's lots of range. All my renders are linear and then in comp when you output your image you can set the gamma to 2.2 if you wish. As a standard though I always deliver linear for the render out of Nuke.

The next question I swear we will get is why are lights in Houdini set using intensity and not stops !

rob

I cant thank you guys enough for the explanations given so far!
And Rob's prediction was spot on… i started a topic related to lights. :shock:
Will visit your webby, cheers!

eitht.

circusmonkey
…
I leave all that at default and strip the gamma from textures in the shader refer to my site (http://www.circusmonkey.com.au/circusmonkey/Shops.html) [circusmonkey.com.au] . I then render and in the mplay display and set my gamma to 2.2.
…

in this case be very careful about your color swatches, because untill you set correct colors blindly, they are not linear, having this setting at 2.2 is beneficial if you use LW, since it will even set gamma in Mplay to 2.2 automatically and in render view, COPs, simply anywhere you need it, it still doesn't bake anything in. (If we are of course talking about the same setting Color Settings/Color Correction)

I don't encourage baking the gamma in to renders as well, it was just an option if you want to save jpeg which can be seen directly. For production it's always beneficial to save in linear space to floating point format so you have full range to process in comp

With the screen shot below, am I understanding Houdini grids, is each grid 1 meter ?

If I wanted to model something that is 0.28194 width in meters how can, this may be really simple question create a cube which is of that width ?

Also when creating large bodies of water as a example how do you any of you get the measurements of a body of a water ?