- This topic has 26 replies, 8 voices, and was last updated 9 years, 1 month ago by
.
-
Posts
-
Added to RFC @yves !
Add a warning box / Check+info step /suggestions just before the atual stacking process to inform the user about any settings that are sub-optimal. Like “No Bias frames added for Flatframe calibration (Flatframe ISO/Gain = XXXX, Lightframe ISO/Gain= XXXXX)”
Good suggestion Kees, so a window showing all settings and possibly unadavised settings when you click on stack. Will add this to the RFC ;-)
I move masters to a different directory as the work directory, it would be nice if I can tell APP to change the location of the master which is in the list.
I move masters to a different directory as the work directory, it would be nice if I can tell APP to change the location of the master which is in the list.
@supernov, Would it be an idea, for APP to automatically maken a Masters folder, where it saves the results?Would it be possible to auto-populate the LP boxes with settings I can change?
@supernov, can you clarify Auto-populate the LP boxes? I don’t think I understand what you are referrring to.1. Yes it would be nice to have an automatic masters folder, but it isn’t possible to let the user choose the location?
2. I’m actually refering to how PI is populating the measurement point for Dark Background Extraction. APP knows the value of the background I guess, so could it use that to draw the boxes in area’s with that or lower values? It would save a great deal in time to manually draw them. And it would also be very nice to be able to save such a model to disk, so when I restart APP I can load in that drawn model.
1) @supernov, well it could be done, but then you will need to select this before you start the process, otherwise APP will need to wait for your location confirmation before it can continue. I’ll think about this.
2) Letting the software choose where to place the area select boxes is the same as setting you up for a bad result. The sample problem for good gradient correction is extremely!!! difficult and fraud with danger. I have discussed this recently with Wei-Hao Wang (professional astrophysicist). Currently there is no robust answer to do this using an automated process. And since it is fraud with danger, I will not implement this, because chances are much bigger that the results will be worse.
Personally, placing 30 select boxes only takes a couple of minutes, so I don’t consider this a real problem. And it is virtually impossible to place the select boxes correctly all at once in images with a lof nebulosity, let alone automating it…. You need to study the gradients and work towards an optimal solution.
The gradient/light-pollution correction is usually done on a stack of multiple frames. The gradients in the data are extremely complex due to this and simply can’t be perfectly described by simpel models. So letting the software place select boxes in areas with similar values like for instance your first select box is really ill advised. It won’t help at all in my opinion (i have tested this thoroughly). It will only contribute to more errors and a less optimal model.
Saving such a model is really no solution, since no model will be the same. If you make a light frame of 5 minutes exposure time. Then another one of 5 minutes exposure time on any given night, the gradients will already be different. Combining the 2 lights will have another model, and so one…
I do understand your ideas and suggestions, but the problem of getting a good gradient correction is so complex that automating this, like you propose, will only give rise to less optimal solutions, so I don’t see much value for it.
Personally, I wish for the best solution, and this can be achieved with some experience using the interactive remove LP tool in APP. I will try to make a movie with sound showing you how to use this tool and how you can work towards an optimal solution.
Haha, veeeery clear Mabula, thanks. I totally get it’s too difficult and since I can just put the boxes over star-areas as well (which I didn’t know) helps a lot. I also notice I just have to do some trial and error to get the best result and most of the times it’s amazing. On the other hand, I’m also just placing the boxes pretty randomly (over the gradients ofcourse, but basically across the entire image), in that sense letting APP place the boxes in a grid and then letting me move them might also be nice. But it’s fine as-is, a good point to make in the manual.
Thank you @supernov ?
There is off course a big difference between
- images with lots of background and little or no nebulosity, like your recent M81/M82 data
- or data full of nebulosity
A correction is simple and quickly done in the first case, and that could be more or less automated.
But in the second case, you will need to work step by step, I’ll make a video showing this on difficult data with lots of nebulosity ?
One more tip, placing the select boxes evenly over the image helps, since each select box has a weight in the model calculation.
- You must be logged in to reply to this topic.