Pro Contrast Moose Peterson

Moose Peterson talks about his signature contrast look which he dubs as ‘Pro Contrast’. Originally he used Nik Color effect now he is using luminar.

Could anyone advice on how to achieve similar results with darktable modules. Thanks!

At least a before/after image would be useful, ideally with a RAW file people can experiment on, but from the description of this module I would try local contrast, if necessary with a luminance-based parametric mask and multiple instances.

1 Like

Boris’ dtpresets ought to be a good start:

Have fun!
Claes in Lund, Sweden

3 Likes

And of course, like most such rather extreme contrast settings, there’s signs of halo formation at sharp edges (look at the horizon in the landscape Moose Peterson posted as example).

Now, on such a small image it’s not too noticable. Full-screen, not so sure…

3 Likes

I have a few of his local contrast presets as well…he leverages reverse multiply quite often…I’ll dig them up

EDIT:
bilat_Clarity-BH.dtpreset (1005 Bytes)
bilat_DNR 1 - BH.dtpreset (1009 Bytes)
bilat_DNR 2 - BH.dtpreset (1003 Bytes)
bilat_DNR 3 -BH.dtpreset (1.0 KB)

4 Likes

I’m guessing here, but it looks like adding local contrast to the shadows and a bit less local contrast to the mid tones.

You could use contrast equaliser , use the 'clarity ’ preset as a start. The mix slider can be used to dial in how much you want . Then use a parametric mask to lessen the effect as the luminosity goes up.

Or the same mask with the local contrast module.

I use the ‘pro contrast’ module of Nik quite often, but mostly for it’s 'color cast removal ’ slider .
It’s contrast slider seems quite normal .
But it’s dynamic contrast slider has the habit of brightening the shadows a bit while adding local contrast . This sounds more like what the local-contrast module does in DT.

Thank you, Todd.

To those planning to experiment with Boris’ presets
it might be wise to remember that
Boris’ Family Motto reads something like
Why use only one instance of a module when you can have seventeen™?
(Which can be seen in most of his darktable tutorials.)

Have fun!
Claes in Lund, Sweden

4 Likes

So true one of my favourite videos was the one where he did a series of edits using only colorbalance…albeit “multiple” times… :slight_smile:

1 Like

AI ProContrast 2000 Platinium.dtstyle (7,2 Ko)

6 Likes

Merci, mon père!

Looks great … but I need to buy a new notebook ;-). Even a i7-7820hq with Nvidia Quadro 1200m is at its limits to deal with this style.

Is there a specific goal for each instance of diffuse…on this recent playraw which happened to be what I tested your style on, the final instance seems to generate too much artifact/grain (seen on the face)…I just took the image added exposure lens correction, profiled denoise filmic auto-selected and then your style…just curious on the though process for each instance of diffuse?? For sure you would have to tweak almost any style…

  • The first instance is for local contrast,
  • The second instance is to smoothen noise and bad things,
  • The third instance is to deblur the lens issues. Of course, this one needs adaptation depending on the glass you are using. I tweaked it for Warm colorful day in fall. If it’s too hard, reduce the number of iterations and possibly the radius span.
4 Likes

Shall I open a new thread for this?

I can confirm that exporting is extremely slow when using this preset (applied to Warm colorful day in fall), due to OpenCL being disabled. I have 6 GB of RAM on the NVidia 1060 card, and 64 GB RAM in the machine; the CPU is an AMD Ryzen 5 5600X. Note the [opencl_diffuse] couldn't enqueue kernel! -4 messages below.

176.436056 [export] creating pixelpipe took 0.053 secs (0.159 CPU)
176.436073 [pixelpipe_process] [export] using device 0
176.436095 [dev_pixelpipe] took 0.000 secs (0.000 CPU) initing base buffer [export]
176.446887 [dev_pixelpipe] took 0.011 secs (0.007 CPU) processed `raw black/white point' on GPU, blended on GPU [export]
176.451013 [dev_pixelpipe] took 0.004 secs (0.004 CPU) processed `white balance' on GPU, blended on GPU [export]
176.458158 [dev_pixelpipe] took 0.007 secs (0.003 CPU) processed `highlight reconstruction' on GPU, blended on GPU [export]
176.546673 [dev_pixelpipe] took 0.089 secs (0.048 CPU) processed `demosaic' on GPU, blended on GPU [export]
176.560071 [dev_pixelpipe] took 0.013 secs (0.009 CPU) processed `lens correction' on GPU, blended on GPU [export]
176.576557 [dev_pixelpipe] took 0.016 secs (0.013 CPU) processed `exposure' on GPU, blended on GPU [export]
176.862751 [dev_pixelpipe] took 0.286 secs (1.330 CPU) processed `tone equalizer' on CPU, blended on CPU [export]
176.938311 [dev_pixelpipe] took 0.076 secs (0.764 CPU) processed `tone equalizer 1' on CPU, blended on CPU [export]
176.995765 [dev_pixelpipe] took 0.057 secs (0.057 CPU) processed `input color profile' on GPU, blended on GPU [export]
image colorspace transform Lab-->RGB took 0.015 secs (0.008 GPU) [channelmixerrgb ]
177.041929 [dev_pixelpipe] took 0.046 secs (0.027 CPU) processed `color calibration' on GPU, blended on GPU [export]
177.087537 [default_process_tiling_cl_ptp] use tiling on module 'diffuse' for image with full size 7374 x 4924
177.087542 [default_process_tiling_cl_ptp] (3 x 1) tiles with max dimensions 4648 x 4924 and overlap 1024
177.087544 [default_process_tiling_cl_ptp] tile (0, 0) with 4648 x 4924 at origin [0, 0]
187.978417 [opencl_diffuse] couldn't enqueue kernel! -4
187.982514 [default_process_tiling_opencl_ptp] couldn't run process_cl() for module 'diffuse' in tiling mode: 0
187.982519 [opencl_pixelpipe] could not run module 'diffuse' on gpu. falling back to cpu path
283.749282 [dev_pixelpipe] took 106.707 secs (1127.128 CPU) processed `diffuse or sharpen' on CPU, blended on CPU [export]
283.749305 [default_process_tiling_cl_ptp] use tiling on module 'diffuse' for image with full size 7374 x 4924
283.749308 [default_process_tiling_cl_ptp] (2 x 1) tiles with max dimensions 7368 x 4924 and overlap 16
283.749309 [default_process_tiling_cl_ptp] tile (0, 0) with 7368 x 4924 at origin [0, 0]
291.479139 [opencl_diffuse] couldn't enqueue kernel! -4
291.483153 [default_process_tiling_opencl_ptp] couldn't run process_cl() for module 'diffuse' in tiling mode: 0
291.483158 [opencl_pixelpipe] could not run module 'diffuse' on gpu. falling back to cpu path
330.331615 [dev_pixelpipe] took 46.582 secs (446.771 CPU) processed `diffuse or sharpen 1' on CPU, blended on CPU [export]
330.331641 [default_process_tiling_cl_ptp] use tiling on module 'diffuse' for image with full size 7374 x 4924
330.331644 [default_process_tiling_cl_ptp] (2 x 1) tiles with max dimensions 6168 x 4924 and overlap 64
330.331646 [default_process_tiling_cl_ptp] tile (0, 0) with 6168 x 4924 at origin [0, 0]
350.124052 [opencl_diffuse] couldn't enqueue kernel! -4
350.128109 [default_process_tiling_opencl_ptp] couldn't run process_cl() for module 'diffuse' in tiling mode: 0
350.128115 [opencl_pixelpipe] could not run module 'diffuse' on gpu. falling back to cpu path
467.724100 [dev_pixelpipe] took 137.392 secs (1337.772 CPU) processed `diffuse or sharpen 2' on CPU, blended on CPU [export]
467.802900 [dev_pixelpipe] took 0.079 secs (0.065 CPU) processed `color balance rgb' on GPU, blended on GPU [export]
467.833439 [dev_pixelpipe] took 0.031 secs (0.014 CPU) processed `filmic rgb' on GPU, blended on GPU [export]
image colorspace transform RGB-->Lab took 0.016 secs (0.008 GPU) [colorout ]
467.892337 [dev_pixelpipe] took 0.059 secs (0.035 CPU) processed `output color profile' on GPU, blended on GPU [export]
468.245263 [dev_pixelpipe] took 0.353 secs (0.353 CPU) processed `dithering' on CPU, blended on CPU [export]
468.342838 [dev_pixelpipe] took 0.098 secs (1.093 CPU) processed `display encoding' on CPU, blended on CPU [export]
468.342945 [opencl_profiling] profiling device 0 ('NVIDIA GeForce GTX 1060 6GB'):
468.342948 [opencl_profiling] spent  0.2184 seconds in [Write Image (from host to device)]
468.342950 [opencl_profiling] spent  0.0019 seconds in rawprepare_1f
468.342953 [opencl_profiling] spent  0.0021 seconds in whitebalance_1f
468.342955 [opencl_profiling] spent  0.0036 seconds in highlights_1f_lch_bayer
468.342957 [opencl_profiling] spent  0.0009 seconds in border_interpolate
468.342958 [opencl_profiling] spent  0.0038 seconds in rcd_border_green
468.342960 [opencl_profiling] spent  0.0053 seconds in rcd_border_redblue
468.342961 [opencl_profiling] spent  0.0049 seconds in rcd_populate
468.342963 [opencl_profiling] spent  0.0038 seconds in rcd_step_1_1
468.342965 [opencl_profiling] spent  0.0029 seconds in rcd_step_1_2
468.342966 [opencl_profiling] spent  0.0018 seconds in rcd_step_2_1
468.342968 [opencl_profiling] spent  0.0050 seconds in rcd_step_3_1
468.342969 [opencl_profiling] spent  0.0027 seconds in rcd_step_4_1
468.342970 [opencl_profiling] spent  0.0015 seconds in rcd_step_4_2
468.342972 [opencl_profiling] spent  0.0043 seconds in rcd_step_5_1
468.342973 [opencl_profiling] spent  0.0070 seconds in rcd_step_5_2
468.342974 [opencl_profiling] spent  0.0070 seconds in rcd_write_output
468.342976 [opencl_profiling] spent  0.0309 seconds in [Copy Image (on device)]
468.342978 [opencl_profiling] spent  0.0117 seconds in exposure
468.342979 [opencl_profiling] spent  0.2296 seconds in [Read Image (from device to host)]
468.342981 [opencl_profiling] spent  0.0094 seconds in colorin_unbound
468.342982 [opencl_profiling] spent  0.0085 seconds in colorspaces_transform_lab_to_rgb_matrix
468.342983 [opencl_profiling] spent  0.0108 seconds in channelmixerrgb_CAT16
468.342985 [opencl_profiling] spent 18.1803 seconds in diffuse_blur_bspline
468.342986 [opencl_profiling] spent 19.9586 seconds in diffuse_pde
468.342988 [opencl_profiling] spent  0.0177 seconds in colorbalancergb
468.342990 [opencl_profiling] spent  0.0064 seconds in filmic_mask_clipped_pixels
468.342991 [opencl_profiling] spent  0.0088 seconds in filmicrgb_chroma
468.342993 [opencl_profiling] spent  0.0087 seconds in colorspaces_transform_rgb_matrix_to_lab
468.342995 [opencl_profiling] spent  0.0166 seconds in colorout
468.342997 [opencl_profiling] spent 38.7750 seconds totally in command queue (with 3 events missing)
468.343011 [dev_process_export] pixel pipeline processing took 291.907 secs (2915.494 CPU)

Card info:

0.051513 [opencl_init] device 0 `NVIDIA GeForce GTX 1060 6GB' has sm_20 support.
0.051716 [opencl_init] device 0 `NVIDIA GeForce GTX 1060 6GB' supports image sizes of 16384 x 32768
0.051720 [opencl_init] device 0 `NVIDIA GeForce GTX 1060 6GB' allows GPU memory allocations of up to 1519MB
[opencl_init] device 0: NVIDIA GeForce GTX 1060 6GB 
     CANONICAL_NAME:           nvidiag
     GLOBAL_MEM_SIZE:          6077MB
     MAX_WORK_GROUP_SIZE:      1024
     MAX_WORK_ITEM_DIMENSIONS: 3
     MAX_WORK_ITEM_SIZES:      [ 1024 1024 64 ]
     DRIVER_VERSION:           470.82.00
     DEVICE_VERSION:           OpenCL 3.0 CUDA

According to nvida-smi, with the browser open and darktable idling in the background, I have 465MiB in use of 6077MiB.

Is there some setting I could change to allow the operation to run on the GPU? In the manual, the only setting I found mentioning OpenCL and tiling is only concerned with the speed of memory copies: opencl_use_pinned_memory, but nothing about OpenCL and tiling memory sizes. Starting with an empty darktable config did not help.

more than 2 minutes, ouch…

Yep, ouch. The steps of same style, applied to a 16 MPixel Nikon D7000 image:

82.015155 [dev_pixelpipe] took 8.002 secs (8.036 CPU) processed `diffuse or sharpen' on GPU, blended on GPU [export]
84.511067 [dev_pixelpipe] took 2.496 secs (2.439 CPU) processed `diffuse or sharpen 1' on GPU, blended on GPU [export]
93.126881 [dev_pixelpipe] took 8.616 secs (8.136 CPU) processed `diffuse or sharpen 2' on GPU, blended on GPU [export]

@kofa
How did you arrive at those clockings?
My machine if just a little bit better than yours,
and that same style took

  • 21 seconds with openCL
  • 33 seconds without openCL

when I transformed a RAF into a JPG.

Have fun!
Claes in Lund, Sweden

It’s not a bug, it’s hardware. Your GPU doesn’t have enough memory available.

CL_MEM_OBJECT_ALLOCATION_FAILURE -4

more than 2 minutes, ouch…

Still 1/8th of a blind deconvolution runtime.

Have you tried using the image from Warm colorful day in fall ?

@aurelienpierre - The card has 6gb, but

allows GPU memory allocations of up to 1519MB

That could be a reason.