[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/3/ - 3DCG

View post   
File: 684 KB, 960x540, texture.png [View same] [iqdb] [saucenao] [google]
599020 No.599020 [Reply] [Original]

/3/,

Why am I getting so much noise in my render using Blender?

I'm using a standard HDR image with no lightning.

Using 1024 samples i'm getting the attached result. Could someone assist?

>> No.599021

>>599020
>Blender

>> No.599023
File: 2.04 MB, 3264x1836, 20180101_135810.jpg [View same] [iqdb] [saucenao] [google]
599023

>>599020
Try bumping up the clamp direct and indirect, that might fix it.

>> No.599026

use denoiser, clamp indirect to 3, clamp direct to 1, disable caustics,don't use small lamps, reduce bounces to 4

>> No.599030

>>599023
>>599026
>clamping
>less bounces
how to suck at rendering 101
keep it up guys

>> No.599042

>>599020
under "Render Layers", turn on Denoising. won't work with Progressive Refine, but final renders shouldn't be done with prog. refine anyway

>> No.599050
File: 15 KB, 388x193, Capture.png [View same] [iqdb] [saucenao] [google]
599050

>>599030
shh.
>>599023
is correct.
>pic related

>> No.599098

if you're using cycles, increase the size/radius of the lamp. it's really the only good way i've found to reduce noise

>> No.599099

>>599098
Not same anon, but what would be a good size/radius range to work with?

I've tried it before and I honestly never really got a good result from it.

>> No.599104

>>599099
you'll just have to experiment

>> No.599106
File: 436 KB, 200x150, 1486771071100.gif [View same] [iqdb] [saucenao] [google]
599106

>>599098
>>599099
>>599104
blender can't even render smaller sizes of lights without noise? holy fuck that's hilarious

>> No.599111

>>599106
cycles definitely can; it's just that the amount of samples required seems rise sharply the smaller your lamp becomes. it really depends on how the lamp is used, though. some uses will be more noise prone than others

>> No.599121

>>599106
It's not that it can't, it's just that the results are usually undesirable. It's more situational, than anything.

Though of course, my computer is also utter shit, so that may be playing a part in it.

>> No.599212
File: 6 KB, 252x240, 1423017713462.jpg [View same] [iqdb] [saucenao] [google]
599212

>>599121
don't you think your software should account for all desirable (at least physically correct) situations

>> No.599216

>>599020
Just like increase samples? Are you retarded?

>> No.599219

>>599216
probably is desu

>> No.599221

>Using Cycles when Renderman is LITERALLY free

Top qeq

>> No.599224

>>599212
It's not the software I'm worried about.

It's the hardware running the software. or rather, the lack of.

>> No.599233

>not using redshift
it's like op really is a faggot

>> No.599293

>>599020
You have multiple importances checked, right?

>> No.599509

>>599221
>Top qeq
i read this as "top check".

>> No.599539

>>599023
no no you should only change clamp direct as a last resort. it changes the brightness of the render.

>> No.599540

>>599042
YES denoising is magic.

>> No.599667

use the new eevee engine, almost just as good