Diffuse module is now in the master.....

Processing
73

I don’t know what you’ve done here but this looks like you’ve checked out the release-3.7.0 tag, where you should just checkout master. At the risk of repeating myself, version 3.7.0 is identical to version 3.6.0 and so does not contain diffuse. The current development version is 3.7.0+485, which means it has 485 additional changes on top of version 3.7.0 (some of which are the diffuse module).

If you’re on Manjaro/ArchLinux and you want to track the development version you would probably be best by installing the darktable-git package from the AUR.

74

thanks so much for the pointer about the checkout. I incorrectly assumed that the highest number was the master. I scrolled up on the list and saw the master and did the git checkout on that and it worked perfect. Now I can spent the remainder of the weekend exploring diffuse. Thanks again. I’ll pay it forward someday for someone.

1 Like
75

Same here. 3 or 5 iterations may work, anything more is just endless calculation …
What would be a good “minimum requirement” to have this working in a reasonable time?

CPU:
  Topology: Quad Core model: Intel Core i5-3570 bits: 64 type: MCP 
  arch: Ivy Bridge rev: 9 L2 cache: 6144 KiB 
  flags: lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 27280 
  Speed: 1605 MHz min/max: 1600/3800 MHz Core speeds (MHz): 1: 1920 2: 1940 
  3: 1991 4: 2182 
Graphics:
  Device-1: Intel Xeon E3-1200 v2/3rd Gen Core processor Graphics 
  vendor: ASUSTeK P8H77-I driver: i915 v: kernel bus ID: 00:02.0 
  Display: x11 server: X.Org 1.20.4 driver: modesetting unloaded: fbdev,vesa 
  resolution: 1920x1200~60Hz 
  OpenGL: renderer: Mesa DRI Intel Ivybridge Desktop v: 4.2 Mesa 18.3.6 
  direct render: Yes

I also have a Nvidia 1050 but openCL is just crashing every time the machine is waking up from suspend-to-ram :frowning:

Btw: debian Buster

76

It really needs OpenCL to run in any reasonable speed. There are some new “fast” presets in latest master (that use just one iteration) which might work better without OpenCL.

1 Like
77

ok, thanks. So I’ll have to see if I can find a mainboard with some UEFI that’s not faulty (I have two of those that are) so I can get that openCL working without those crashes.
(and suspend-to-ram is welcomed here in Germany as we have world’s highest electricity prices).

78

PA111122_01
PA111122_012000×1493 222 KB

PA111122
PA1111222000×1493 621 KB

This is what I was able to get out of this photo with diffuse/sharpen. The lens was foggy (it was cold outside the conservatory).

36 Likes
79

Good example of its power: must make @anon41087856 a happy person. :medal_sports:

1 Like
80

It does, thanks

2 Likes
81

Would you care to share your settings for your example.
I am finding that presets are particularly important with this module.

82

The settings are in the jpg file.

83

Thomas beat me to it …load the JPG as a sidecar

1 Like
84

oh wow guys, 21 favs in 6 hours, thank you all!

1 Like
85

You get triple because

  • it highlights the power of a relatively new module
  • it is a good picture (besides the fog before processing)
  • you are popular :sunglasses::stuck_out_tongue:
3 Likes
86

One thing I love about this module (along with Color Balance RGB) is how they preserve the colors. Somehow, to my perception, the colors in both pictures are “the same”, except they are foggy in one and not in the other. It’s mind-blowing.

2 Likes
87

On that, it would be a good exercise to use it on mountainous landscapes with a strong haze (see Play Raws or sample images from other threads). Usually haze removal comes with saturated colour casts. Newer algorithms have improved but it would be nice to see how this module does.

88

Nah, it’s just accurate physics. Rutherford said “science is either physics or collecting stamps”. He was right.

89

I simply never expected such a brilliant result (figureatively and literally) and was quite amazed :slight_smile:

90

I am really glad you are more into physics than stamps :slight_smile:

1 Like
91

I know, but the way I see it, something can be accurate and physical, and still mind-blowing :smile: Don’t get jaded! It’s almost as bad as believing it’s magic.

92

Blending dehaze in lightness or some other mode will usually avoid the cast…though I think some people enjoy the boost in saturation…I know the dehaze in RT is a very strong effect but I think it does also offer to use a luminance mode…Funny I like the dehaze preset in diffuse module but to me it does not give the traditional appearance of dehaze…it adds a lot of detail and actually seems to brighten the image a bit as opposed to darkening it …at least in my experience…