Community
Arnold for Houdini Forum
Rendering with Arnold in Houdini and Solaris using the HtoA plug-in.
cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Noice filter warning

9 REPLIES 9
Reply
Message 1 of 10
Anonymous
576 Views, 9 Replies

Noice filter warning

I am getting

"Using feature AOV 'denoise_albedo'
    Filters for feature and light AOVs are different: denoising works better with matching filters.
Using feature AOV 'N'
    Filters for feature and light AOVs are different: denoising works better with matching filters.
Working with 5 frames at
I am also seeing this in the documentation
  • The beauty must use the same filter (gaussian_filter) as the other denoiser AOVs (N, Z, variance, and denoise_albedo), which are created by the Arnold plugin for you.


Is beauty not already gaussian? set in the sampling tab?

Or is there another place I need to set the filter?

Thanks

Tags (3)
Labels (3)
9 REPLIES 9
Message 2 of 10
Stephen.Blair
in reply to: Anonymous

What version of Arnold?

Can you post the noice command line?

And the Arnold log for the render (at least the part that shows the list of AOVs) ?

Thanks



// Stephen Blair
// Arnold Renderer Support
Message 3 of 10
Anonymous
in reply to: Stephen.Blair

log_aov.zip


htoa 5.6.1

noice -pr 4 -v 0.55 -ef 2 -l specular -l sss -i /mnt/tmp/main.0651.exr -o /mnt/tmp/main_denoise.0651.exr


Message 4 of 10
Stephen.Blair
in reply to: Anonymous

I have to think it is this:

00:02:08  9968MB         | [aov] registered driver: "/out/rop_ar_main:exr:aov13:renderCamera_MAIN" (driver_exr)
00:02:08  9968MB         | [aov]  * "Z" of type FLOAT filtered by "/out/rop_ar_main:closest_filter:aov13" (closest_filter)


Another thing to check is the metadata in the EXR. That's what noice is actually reading.
For example:

C:\htoa\scripts\bin>oiiotool -v --info  denoiseme.0007.exr


// Stephen Blair
// Arnold Renderer Support
Message 5 of 10
Anonymous
in reply to: Anonymous

Trying to understand the warning, is it saying that the filter for the feature aovs has to be the same as rgba correct?
Metadata:

arnold/aovs/denoise_albedo_noice/filter: "gaussian_filter"
arnold/aovs/denoise_albedo_noice/filter_width: 2
arnold/aovs/denoise_albedo_noice/lpe: 1
arnold/aovs/denoise_albedo_noice/lpe_expression: "((C<TD>A)|(CVA)|(C<RD>A))"
arnold/aovs/denoise_albedo_noice/source: "denoise_albedo"
arnold/aovs/RGBA/filter: "gaussian_filter" 
arnold/aovs/RGBA/filter_width: 2 
arnold/aovs/RGBA/lpe: 1 
arnold/aovs/RGBA/lpe_expression: "C.*" 
arnold/aovs/RGBA/source: "RGBA"

They are the same, is it possibly reporting this regardless? looking around at other posts on here, even ones you've posted it seems like this error pops up everytime

Message 6 of 10
Stephen.Blair
in reply to: Anonymous

I don't see the Z metadata.

Could you post the full output of the noice command? And the full output of the oiiotool command? thanks



// Stephen Blair
// Arnold Renderer Support
Message 7 of 10
Anonymous
in reply to: Anonymous

Apologies,


channel list: R, G, B, A, N.R, N.G, N.B, P.R, P.G, P.B, RGBA_LIGHT_PRP_01.R, RGBA_LIGHT_PRP_01.G, RGBA_LIGHT_PRP_01.B, RGBA_LIGHT_PRP_01_variance.R, RGBA_LIGHT_PRP_01_variance.G, RGBA_LIGHT_PRP_01_variance.B, denoise_albedo_noice.R, denoise_albedo_noice.G, denoise_albedo_noice.B, emission.R, emission.G, emission.B, motionvector.R, motionvector.G, motionvector.B, specular.R, specular.G, specular.B, specular_variance.R, specular_variance.G, specular_variance.B, sss.R, sss.G, sss.B, sss_variance.R, sss_variance.G, sss_variance.B, transmission.R, transmission.G, transmission.B, uv.R, uv.G, uv.B, variance.R, variance.G, variance.B, variance.A
    arnold/AA_samples: 3
    arnold/AA_seed: 1650
    arnold/aovs/denoise_albedo_noice/filter: "gaussian_filter"
    arnold/aovs/denoise_albedo_noice/filter_width: 2
    arnold/aovs/denoise_albedo_noice/lpe: 1
    arnold/aovs/denoise_albedo_noice/lpe_expression: "((C<TD>A)|(CVA)|(C<RD>A))"
    arnold/aovs/denoise_albedo_noice/source: "denoise_albedo"
    arnold/aovs/emission/filter: "gaussian_filter"
    arnold/aovs/emission/filter_width: 2
    arnold/aovs/emission/lpe: 1
    arnold/aovs/emission/lpe_expression: "C[LO]"
    arnold/aovs/emission/source: "emission"
    arnold/aovs/motionvector/filter: "gaussian_filter"
    arnold/aovs/motionvector/filter_width: 2
    arnold/aovs/motionvector/source: "motionvector"
    arnold/aovs/N/filter: "gaussian_filter"
    arnold/aovs/N/filter_width: 2
    arnold/aovs/N/source: "N"
    arnold/aovs/P/filter: "gaussian_filter"
    arnold/aovs/P/filter_width: 2
    arnold/aovs/P/source: "P"
    arnold/aovs/RGBA/filter: "gaussian_filter"
    arnold/aovs/RGBA/filter_width: 2
    arnold/aovs/RGBA/lpe: 1
    arnold/aovs/RGBA/lpe_expression: "C.*"
    arnold/aovs/RGBA/source: "RGBA"
    arnold/aovs/RGBA_LIGHT_PRP_01/filter: "gaussian_filter"
    arnold/aovs/RGBA_LIGHT_PRP_01/filter_width: 2
    arnold/aovs/RGBA_LIGHT_PRP_01/source: "RGBA_LIGHT_PRP_01"
    arnold/aovs/RGBA_LIGHT_PRP_01_variance/filter: "variance_filter"
    arnold/aovs/RGBA_LIGHT_PRP_01_variance/filter_width: 2
    arnold/aovs/RGBA_LIGHT_PRP_01_variance/source: "RGBA_LIGHT_PRP_01"
    arnold/aovs/specular/filter: "gaussian_filter"
    arnold/aovs/specular/filter_width: 2
    arnold/aovs/specular/lpe: 1
    arnold/aovs/specular/lpe_expression: "C<RS[^'coat''sheen']>.*"
    arnold/aovs/specular/source: "specular"
    arnold/aovs/specular_variance/filter: "variance_filter"
    arnold/aovs/specular_variance/filter_width: 2
    arnold/aovs/specular_variance/lpe: 1
    arnold/aovs/specular_variance/lpe_expression: "C<RS[^'coat''sheen']>.*"
    arnold/aovs/specular_variance/source: "specular"
    arnold/aovs/sss/filter: "gaussian_filter"
    arnold/aovs/sss/filter_width: 2
    arnold/aovs/sss/lpe: 1
    arnold/aovs/sss/lpe_expression: "C<TD>.*"
    arnold/aovs/sss/source: "sss"
    arnold/aovs/sss_variance/filter: "variance_filter"
    arnold/aovs/sss_variance/filter_width: 2
    arnold/aovs/sss_variance/lpe: 1
    arnold/aovs/sss_variance/lpe_expression: "C<TD>.*"
    arnold/aovs/sss_variance/source: "sss"
    arnold/aovs/transmission/filter: "gaussian_filter"
    arnold/aovs/transmission/filter_width: 2
    arnold/aovs/transmission/lpe: 1
    arnold/aovs/transmission/lpe_expression: "C<TS>.*"
    arnold/aovs/transmission/source: "transmission"
    arnold/aovs/uv/filter: "gaussian_filter"
    arnold/aovs/uv/filter_width: 2
    arnold/aovs/uv/source: "uv"
    arnold/aovs/variance/filter: "variance_filter"
    arnold/aovs/variance/filter_width: 2
    arnold/aovs/variance/lpe: 1
    arnold/aovs/variance/lpe_expression: "C.*"
    arnold/aovs/variance/source: "RGBA"
    arnold/auto_transparency_depth: 4
    arnold/bounds_max_x: 1023
    arnold/bounds_max_y: 575
    arnold/bounds_min_x: 0
    arnold/bounds_min_y: 0
    arnold/camera/far_clip: 1e+06
    arnold/camera/near_clip: 0.01
    arnold/color_manager: "color_manager_ocio"
    arnold/color_space: "ACES - ACEScg"
    arnold/diffuse_depth: 1
    arnold/diffuse_samples: 1
    arnold/fps: 24
    arnold/frame: 1650
    arnold/host/hw: "1 x Intel(R) Core(TM) i7-9700 CPU @ 3.00GHz (8 cores, 8 logical) with 64162MB"
    arnold/host/name: ""
    arnold/host/os: "Ubuntu 18.04.5 LTS (Bionic Beaver), Linux kernel 5.4.0-70-generic"
    arnold/license_state: "valid"
    arnold/specular_depth: 1
    arnold/specular_samples: 1
    arnold/sss_samples: 1
    arnold/stats/date: "Tue Oct  5 17:26:30 2021"
    arnold/stats/geo/curve_segments: 1.74172e+07
    arnold/stats/geo/triangles: 1.47171e+08
    arnold/stats/memory/peak: 22332.7
    arnold/stats/memory/start: 75.6016
    arnold/stats/rays/all/pixel: 28780
    arnold/stats/rays/all/total: 1.69751e+10
    arnold/stats/rays/camera/pixel: 38.3232
    arnold/stats/rays/camera/total: 2.26039e+07
    arnold/stats/time/render: 2442.23
    arnold/stats/time/setup: 0.598896
    arnold/texture_max_memory_MB: 4096
    arnold/threads: 8
    arnold/total_depth: 4
    arnold/transmission_depth: 4
    arnold/transmission_samples: 1
    arnold/version: "Arnold 6.2.1.0 [359aaa66] linux clang-10.0.1 oiio-2.3.2 osl-1.12.0 vdb-7.1.1 clm-2.0.0.235 rlm-14.1.3 optix-6.6.0 2021/04/21 13:54:09"
    arnold/volume_depth: 0
    arnold/volume_samples: 0
    CameraFilmApertureHorizontal: 1
    CameraFilmApertureVertical: 0.5625
    CameraFocalLength: 1.48008
    CameraFov: 37.332
    compression: "zip"
    DateTime: "2021:10:05 17:26:30"
    PixelAspectRatio: 1
    screenWindowCenter: 0, 0
    screenWindowWidth: 1
    worldToCamera: 0.830845, 0.0286764, -0.555764, 0, 0.00524006, 0.998224, 0.0593402, 0, -0.556479, 0.0522147, -0.82922, 0, 22.6598, -2.05679, 16.2795, 1
    worldToScreen: 2.45943, 0.15091, -0.555764, -0.555764, 0.0155114, 5.25315, 0.0593402, 0.0593402, -1.64726, 0.27478, -0.82922, -0.82922, 67.0765, -10.8239, 16.2695, 16.2795
    oiio:ColorSpace: "Linear"
    oiio:subimages: 1

interesting there is no Z, I am making my own Z aov for this rop. Does noice want to make its own?

noice output:

noice 6.2.1.0 [359aaa66] - the Arnold denoiser
Using 8 threads.
Loading images...
Loading file "/mnt/I/tmp/test.0655.exr".
Loading file "/mnt/I/tmp/test.0653.exr".
Loading file "/mnt/I/tmp/test.0654.exr".
Loading file "/mnt/I/tmp/test.0656.exr".
Loading file "/mnt/I/tmp/test.0657.exr".
Using feature AOV 'denoise_albedo'
    Filters for feature and light AOVs are different: denoising works better with matching filters.
Using feature AOV 'N'
    Filters for feature and light AOVs are different: denoising works better with matching filters.
Working with 5 frames at 1024x576
Will denoise AOV "RGBA", using associated variance
   Output file will be "/mnt/I/tmp/test_denoise.0655.exr"
Will denoise AOV "specular", using associated variance
   Output file will be "/mnt/I/tmp/test_denoise.0655.exr"
Will denoise AOV "sss", using associated variance
   Output file will be "/mnt/I/tmp/test_denoise.0655.exr"
Start denoising (patch radius 4, search radius 9, variance 0.55)
Denoising RGBA
Denoising specular
Denoising sss
Finished denoising
Saving image /mnt/I/tmp/test_denoise.0655.exr (1024 x 576 x 10)




Message 8 of 10
Anonymous
in reply to: Anonymous

looking again, I forgot that I am writing Z out to a seperate exr
Message 9 of 10
Stephen.Blair
in reply to: Anonymous

Ah, separate EXRs for everything. You're right, that may always trigger the warning (eg it's a bug). I was testing with a merged EXR,



// Stephen Blair
// Arnold Renderer Support
Message 10 of 10
Anonymous
in reply to: Stephen.Blair

Ok cool thank you for the help! One less warning to worry about 🙂

Can't find what you're looking for? Ask the community or share your knowledge.

Post to forums