Helicon Soft Forum

Hi,

I am learning to use Helicon Focus. I have been working with stacks shot with a f4 aperture, but yesterday tried a large stack shot at f16. I tried all three methods but I prefered the results of Method="C". I found that the dust on the sensor caused dark streaking on the Helicon Render Output, so today I tried making a Dust Map after reading the Helicon Focus help files.

To make a long story short, I have tried several different Dust Maps with the source files that I used yesterday and I am getting mixed results at the render output.

1) I tried a properly exposed photo of a white card as a dust map. It was nearly impossible to view any dust specs in this dust map as the white was "white" and the dust specs were, I assume, very light as well.

2) I tried a Dust Map that was photographed without any exposure compensation so the "white card" appeared as neutral grey in Photoshop. The dust specs were somewhat visible.

3) I used the Dust Map that was photographed without any exposure compensation but I also made a Levels adjustment in Photoshop so that the field was brighter and the dust spots were darker before saving it as a copy.

When I used the first example in Helicon Focus I barely noticed any reduction of dust spot streaking.

When I used the second example in Helicon Focus I may have noticed some slight reduction of the dust spot streaking.

When I used the third example, with the exaggerated dust speck visibility, there was a remarkable, nearly complete, reduction in dust spot streaking, but the whole image had a weird cross hatch effect that was especially noticeable at high contrast borders. The cross hatching made the results seem coarse and unusable.

I am finding that I get the best results with Method = "C" for the work I am doing but I did try Method ="B" with the second example dust map. The dust spot streaking was reduced more so than with the "C" method but there was also a weird addition of "banding splotches", for lack of a better description, in the out of focus areas. The pleasant Bokeh was ruined, which was not the case when I did not use any Dust Map for Method = "B" processing on these source files.

It seems to me that there must be some best way to prepare a dustmap so that Helicon can use it effectively.

Has anyone made similar experiments and established a best practice recommendation?

Thanks for sharing your experience. You are welcome to upload the stack with dust maps and we test it ourselves.

Please make sure that you are shooting dust map with the same exposure as the stack itself because dust points are rendered differently depending on aperture. It is also important that white surface you are photographing is not in focus. Otherwise the texture of the white surface might be also considered dust spots. Anyway, you'd better send us at least the dust maps and we will try to help you.

Hi Stas,
Thanks for the offer to look at the dust maps. I made sure that I used a smooth white card that was out of focus and my initial tests used a properly matched aperture.

I gave up on the idea of using an automated process after trying your built in options as well as the process offered in the Canon's Digital Photo Professional utility.

What I have found is that neither process seems adequate and that the photos still require touch up For my work flow it just seems easier to see an obvious dust streak rather than looking for streaks which need touch up but are somewhat hidden by an interim process.

I suspect that I was too optimistic to think that an automated process would provide the results I was hoping for so I have moved on and am happy to do the touch up work by hand.

Thank You.

We still would like to see your stack. I think automatic process might be much better for the following reasons.

The program tries to make the dust spot transparent. That means that after aligning layers each "dust hole" is filled with information from other layers.

If you try to remove the dust spot from stacked result, you have to remove the whole trace.

I suspect that your dust map should be a bit preprocessed to be efficiently used in the program. Maybe some sharpening or increase of contrast should be applied. I will be able to tell more after seeing your stack.

Hi Stas,
If I get some free time I'll try to gather what I was using and send it over.

I am getting a bit over whelmed with the vast amount of data that accumulates with focus stacking, and when I was doing the dust map tests I wasn't spending a lot of effort to archive and collate data that I didn't think I would be using later.

Thank You.

Ok, we just need an image with a dust map and one of the source files.