Histogram issue

[barcokid]

I could use a little help with an issue I'm running into. This is nothing new to me but I thought since you're closing in on your first release, you may want to address this soon. However there is a good chance it is something I am doing wrong since I have/had this same issue dating back to the freeware version. Here's what I see.

I open up the histogram transform window, select my image and perform my adjustments as needed and than click apply. Virtually everytime I select apply, its ends up clipping well into my  data even though my original placement was nowhere closed to clipping. Usually on the 2nd or 3rd try it will stay where I set it. As I mentioned above this same thing happens in the pixinsight freeware version. Anyone else seeing this or any suggestions to help me. It's no huge deal but it is irritating me just a bit.

Other than this issue, I love your program. You all have done a great job.

Thanks!
Brad[/list]

[LD]

Brad,
I too have noticed this since first using the LE version. It is quite different than doing a stretch in other programs. If you scroll down to a post of mine entitled "Histogram Question?" from July you'll see an explanation.

As I understand it, the histogram is "live" even after you apply the stretch to a preview so you are now seeing the endpoints you set imposed on the "new" histogram for the stretched preview. If you store the preview and then hit reset, you should see the histrogram more as you intended and you should see that the image really isn't clipped as the blackpoint "jumps" left. (Althought I do find myself "predicting" where the blackpoint will end up and doing a few iterations to get it right.)

I could have this all wrong, of course (not an unusual occurrence!) and I'm sure the authors will be quick to correct me!

Hope this clarifies more than confuses,
Larry

[Carlos Milovic]

Histogram transforms are applied following this order: Red, Green, Blue and later the RGB composite (same adjustments to all channels). So, if you have changed the black or white points (shadows or highlights clipping) the RGB composite channel will not show the raw histogram, but the predicted one after all the prior adjustments. I think that this is issue that is happening to you. So far, none of our tests have found that the applied changes were different from what the user specified. As Larry said, try reseting the HistogramTransform instance after applying it, so you'll see the true raw data displayed at the RGB channel.

[barcokid]

Guys - thanks that clears up the mystery.  I am able to hit the reset and see it as I expected .

Clear skies