32 bit exr's

[instinct]

Assuming you got VRay set as your current renderer just execute the following line in the maxscript listener:

renderers.current.output_rawExrUseHalf=false

You got to do that on a per-scene basis.

Regards,
Thorsten

P.S. i still wonder why you are aiming for fullfloat tho...

[TimHawker]

do you mean the vray vfb ?
As for the vrayVFB you can set this via maxscrtipt:


.output_rawExrUseHalf : boolean

defaults to on saving half-float. Setting it to false will write 32bit full float.

Regards,
Thorsten

Really interesting, didn't know that V-Ray defaulted to 16-bit.

Thanks for the script, but what would I write in order to change it to full float? If I run that line it just creates an error.

Tim

EDIT: Sorry this has already been answered, thanks instinct. (didn't notice it went on to page 2)

[instinct]

Keep in mind that this is NOT Integer 16bit, but half-float and sufficient in most cases (haven't ever saved to 32bit full float actually)

Regards,
Thorsten

[TimHawker]

Thanks for the info instinct.

A simple way of enabling/disabling the full float output is to run this maxscript before render. It's very easy to do as shown in the picture below. Just choose the file you want to run in Vray Render Settings.

It is true though that 16bit floats are usually more than enough and use far less memory.

Tim

[jonahhawk]

Keep in mind that this is NOT Integer 16bit, but half-float and sufficient in most cases (haven't ever saved to 32bit full float actually)

Regards,
Thorsten

So integer 16bit is something like saving a 16bit tiff or a 48bpp png correct?

[jstrob]

I am really surprised by this thread. I thought vray was outputing 32 bits exr by default. Like written in the help file about VFB:

One of the feature is:
"Keeps the image in full 32-bit floating-point format;"

What does that means if it save to 16 bits by default?

If we want to know if 16 bits is enough or not, we first have to know if we render 32 or 16 bits...

[instinct]

So integer 16bit is something like saving a 16bit tiff or a 48bpp png correct?

No it isn't.16bit half-float is a floating point format.16bit Tiff is integer.
And vray internally calculates at even higher precision so if you set it to 32bit it doesn't convert to16bit just to upconvert for saving I'd say.
If it is enough ? For image layers I'd say defiantly. For audio channels it depends.

[jstrob]

@Instinct

Can I ask you wher did you learn about that maxscript command? Is it documented somewhere?

[instinct]

I've got no idea tbh. Been like that since ages.

[jonahhawk]

No it isn't.16bit half-float is a floating point format.16bit Tiff is integer.
And vray internally calculates at even higher precision so if you set it to 32bit it doesn't convert to16bit just to upconvert for saving I'd say.
If it is enough ? For image layers I'd say defiantly. For audio channels it depends.

That is what I said/meant/asked: integer 16bit == 16bit tif. Just wanted to clarify. PNG is same (color depth) as 16bit Tif. I assume 16bit color mode in PS is integer as well? 0-65k colors per channel.

[instinct]

Yes it is. And sorry for the typo above (sent from my phone) I meant AUX channels (read Position, Normals etc)

Regards,
Thorsten

[Brunelleschi]

can someone please confirm what #7 of this thread states that if i save out a 32-bit .exr from max frame buffer it is really full float 32-bit file. Also if this is this case what is the advantage of saving out from vray frame buffer and running that pre-script?

Thanks.

[wxyz]

It might be needed to produce a VRayAtmosphere Pass.

[Dariusz Makowski (Dadal)]

Heya

A lil bit off question, EXR can it save to 300 DPI ? and if yes then how ?

[WobiNich]

EXR does not care about DPI. Actually this is only a print value, noone really needs it apart from the printer. You can set that when sending something to printer or in photoshop... This does not have to do anything with rendering images where you render to a given pixel size. It just says how big a pixel will be when it gets printed. You can always go into photoshop image size dialog, turn off resample and set whatever dpi you want, it will not in any way alter your image