Histogram and STF readout buttons (please respond)

[ngc1535]

Hi Again,

This is a followup to my last post.
I would expect the  behavior of the three readout buttons in the Histogram Transformation tool to be like the buttons in the STF tool by setting the blackpoint, midtones, and white point.
However, in HT this does not appear to work as the values in the fields only go to "0" by activating one of the buttons and clicking the image.
Does this currently work properly for other users?

On another note- I am currently below a 50% response rate on my questions to this forum. I would enjoy hearing some suggestions on how to garner more answers to questions.

Thanks,
Adam

[drmikevt]

I think the problem is that you are asking questions that you can answer yourself by looking at the associated documentation within in the program.  Check out chapter 3.11 of the Histogram Transformation documentation.  You can see it by clicking the 'browse documentation' button on the bottom bar of the process window. 

[OldSkyEyes]

It seems to work for me, or I don't understand your problem. I see the lines move when I go to different positions in the image.
You can also drag the STF triangle (New Instance) to the bottom bar of the Histogram Transformation, this will copy the settings.
(after applying disable STF, otherwise you see it twice)

[ngc1535]

Section 3.11 reads:

"In the black point histogram readout mode, pixel readout values are used to set the shadows clipping points for individual RGB channels, when the selected view is a color image, or for the RGB/K combined channel, if the view is a grayscale image."

The value is NOT set in the fields by using the buttons on my version of PI in HT. No one yet is addressing this behavior.
It could be that the behavior of the button is to inform the user of the value- which is fine. However, this is a different behavior than the same buttons in the STF tool which sets the value.

[ngc1535]

Here is a video of "my issue"

https://youtu.be/ZSgfdGoMYQU

[pfile]

yeah - that seems like a bug.

but to be honest in almost 10 years i've never tried to use those things; i generally move the midtones and shadows slider by hand, and while adjusting the shadows, see how much i'm clipping. if my experience is the norm, this may be why no one understood your question.

rob

[ngc1535]

Thank you Rob. I really appreciate you looking at it.

I frequently explain stuff to others and I will inevitably go through things with a fine tooth comb. These are the kinds of small bugs- that of course really don't matter and users likely do not use often- but can otherwise add to the appearance of complexity where none exists. Novice/Intermediate users might click on buttons that are clearly there for a reason and find that the behavior doesn't make sense (and chalk it up to the "advanced" nature of the program). It was in that vein that I needed to know if *I* was missing something.

From a programming point of view- if it is not used (as perhaps demonstrated by the fact a bug report has not shown up)- a decision can be made to either fix it- or simply remove it in order to streamline the UI.

If I can claim a credit for not having a really dumb question- perhaps I can repost another recent question I had that I think is more significant with the hope of some attention to it as well?

-adam

[pfile]

no problem - although it turns out the code for the HistogramTransformation module is in github, it's not clear to me what's going on. if i could step thru the code i might be able to figure it out but i think the binaries have been stripped of all their symbols.

i did try this with a RGB image and what's interesting is that although the HT seems to update the realtime preview in a way that seems more sane in response to clicking the image in the different readout modes, the UI of HT is never updated - neither the sliders move in response to a click with the readout cursors, nor do the numeric displays update. worse, it seems like the clicks with the readout cursors are updating a "shadow" state of the HT process such that the sliders superpose with whatever was calculated by the readout cursor click.

Juan is going to have to take a look at this. i suspect more than one funny thing is happening here.

rob

[Juan Conejero]

Yes, this is a bug, thank you for detecting it.

The black point, midtones and white point readout modes work correctly for color images, and also for alpha channels of color and grayscale images. However, these modes don't work for the nominal channel of a grayscale image (they are simply ignored when you click on a grayscale image).

This bug is now fixed in the HistogramTransformation tool. The fix will be available in version 1.8.5 of PixInsight.