I have done a couple of presentations/demonstrations of KStars and Ekos for the Astrophotography Special Interest Group in my Astronomy Club. Last night I did a full demonstration at my house for a small group of 8 astrophotographers with plate solving, guiding, and imaging various targets the others wanted to see or recommended just for fun. We imaged a whole bunch of targets during the night and they asked a number of great questions. They were very impressed with the software and we had a very informative and fun session. There were a few ideas/questions that I did not have an answer to and I promised I would make a comment about it. I am going to list the ones I remember here. One of the things we were trying to do was look at my guiding errors because I have had troubles with guiding and they were trying to help me with it, so there are a number of comments and questions about that.
Questions/ideas about Guiding
1. In the guiding tab, the RA and Dec guiding errors are plotted on one graph, one in green and one in blue. Which is which? Can there be a legend? (to my knowledge RA is green and I did a test to verify it and we agreed we think it is green--but a legend is good)
2. They asked me to change the zoom/scaling on the plot and the boxes at the bottom worked well for that, but the farthest I could zoom in on the plot was a scale of 5 arc seconds vertically. That might be enough, but is there a reason it cannot go further? I don't know the answer.
3. One member asked if there was a way/option to separate the RA and DEC graphs so you could look at them individually because he was having trouble seeing each. I don't think I would like having them always on separate graphs because it would take up more screen space, it would require the graphs to shrink vertically, and it would not be as easy to get the information in one glance, but I can see how having separate graphs might help you analyze one of the axes carefully. Maybe leave it the way it is by default and provide an option to see them separately or one at a time if you click a button? I don't know.
4. Another idea for improving guiding was to set one (or both) of the axes to make corrections only in one direction. I know how to do that in PHD2 but I don't know how to do that in Ekos. Is there a way?
Questions/ideas about plate solving
1. In the iterative plate solving routine, when does Ekos send the command to sync/align to the mount, after every iteration, or after it is done all the iterations?
2. The most important information printed at the bottom as the plate solve is progressing is how far off each iteration was and how many iterations have been done. I think you were talking about making a target-like display for that, maybe with a small plot of the sky behind it showing where it was pointing and where it should have been pointing? When I was demonstrating it, I had the thought that would be really nice and visual
Questions/ideas about the Fits Viewer
1. Somebody asked how saturated an image I had just taken was. It had been auto-stretched. I moused over it and believe the number of counts for the spot I moused over was at the bottom right, but we decided to take a look at the histogram. When we opened that up, the red graph was all the way at the left and we couldn't see where the "white pixels" were as one member said. I'm not sure exactly what he was looking for in terms of the graph, but I think he was trying to figure out the max counts for the brightest pixels. (I don't know how you would make sure that such a value doesn't include hot pixels though!) So I don't think I understand.
2. When you zoom in it always goes to the top left corner. Is there a way to zoom in on the middle?
3. For the file that opens in the fits viewer, if you apply one of the filters, I know this doesn't affect the file saved to the disk, but if you wanted to adjust the stretch to one of the other options in the drop down menu, so that you could analyze the image further, would you be changing the original data or the stretched version? Another words is the displayed image computed from the original data with modifications from the filter or is the displayed image a new set of data that has been altered by the filter?
There were no questions about the Focus routine because I told them you were working on that and we couldn't give it a good/fair test if it can't detect the stars.
I answered a few questions about the Capture tab and Sequences and about the planetarium side of KStars, but nothing major and no suggestions were made there.
I did not demonstrate the Scheduler.
I think that was most of the questions. If I remember any more, I will post them here. Mostly folks were very impressed with the software and they just had a few constructive ideas on how it could be improved. Great Work!!!
Thanks for the demonstration and questions Rob, great feedback as always!
Guiding: I am in fact re-designing the whole guiding GUI but it would be a while.
1. Legend will be added
2. Will see if it can be increased, this was imported from lin_guider so I'll check if it can go down further.
3. Maybe a checbox to show which to plot? I'll experiment with that
4. Yes under Control you can see "Enable Directions" where you can check/uncheck RA/DEC
1. Yes, after each iteration
2. It is already supported (now in my development version it is not working) so I'll check it out.
1. Click "Statistics" to find the min/max. You can also zoom/pan within the histogram
2. Right, you can double-click to get crosshair now, and I'll make it zoom to that point (TODO)
3. The default autostretch does not modify original data, but from the view if you select a filter, it modifies the data. It is not saved unless you explicitly save.
For guiding question 4, I am not sure I explained it well. I think your checkboxes turn off RA or DEC guiding completely. He did not mean turning off the RA or DEC completely, but just in one direction. For example, you could make guider only make DEC corrections to the North if you know there is a steady southern drift and thus avoid a situation where it oscillates back and forth. He said that in his software you there were buttons for each of the four directions so you could turn each on and off individually.
For alignment question 1, I have a followup question. If you are using the plate solve commands to help the mount build a pointing model like I am doing with my Gemini, then I am wondering how the multiple align commands in close proximity would affect the pointing model. Since it is a plate solve, I am sure it is an accurate location you are sending the mount, but then you have a bunch of alignment points really close. But maybe this is really a question for Losmandy or Gemini folks now. What do you think?
For Fits viewer question 3, I have another followup question. If you do autostretch and then you want to try a different filter because autostretch was too much would you then need to reopen the file, or would it go back to the original data and then apply the filter?
Great to hear someone else is using the Losmandy Gemini...
Concerning the pointing model: as I understand it, Ekos syncs the model each time it solves a plate, but does not affect the point model as suxh. I've been wondering about that, and was planning to have my eldest daughter, who studies to be an engineer, to dig into the code, but she found a boyfriend instead.....
As opposed to with the EQMOD, the Gemini seems not to build a pointing model, as I understand it. But then: do you really need a pointing model with the plate solving of EKOS? I normally do not even build a pointing model any more, just turn on the scope, slew to south to do polar alignment and plate solve and run from that. But maybe there are benefits with a pointing model that I haven't seen?
When we are talking Gemini - what I miss most is startup and park sequences. I need to start the mount with the handcontrol, and the unpark command does not work. So if Ekos has parked it, I need disconnect and reconnect and maybe manually slew somewhere...
I'd be very interested in assisting with any development of the Gemini supprt!
Celestron C11, Skywatcher 100 ED Pro
Atik 383L+, ASI294
Mlarsson, I have some good news for you! When I first started playing with plate solving this summer I was concerned that it would not update the pointing model. The reason is at the gemini has two commands, a sync command and an align command. The one you want ekos to give is the align command to build a pointing model, but since the gemini is using the lx200 indi driver and doesnt have one specifically written for it, it can only send the sync command. One solution would be to write a driver, which is not too hard, but I realized there was an easier way!
All you need to do is change a setting on your gemini so that every sync command sent by the computer is interpreted as an align command. Check your manual, I believe it was in the setup menu, it is the sync vs align setting. It works perfectly! I have been doing that for months now. After just a couple of plate solves, I get a really good pointing model. After I found that I lost interest in writing a driver and moved on to other challenges, but don't get me wrong I would still appreciate a driver that knows to send the align command.
I think what would be great functionality on the guiding tab is:
-Saving the calibration so it will be used always on any session without the need for new calibration.
-After a meridian flip, having the option to just flip the RA+DEC so a new calibration would not be needed. Maybe using some stored variable that tracks imaging E of the meridian and W of the meridian.
-The option that if no guidestar is found, the sequence continues without the autoguiding and just blind tracking of the mount, and the possibility to not abort the sequence.
-Maybe the option to search automatically a suitable guidestar, using the SNR ratio of the image, recursive searching for a suitable exposure time until a guidestar is found within some upper and lower limits.
This functionality would be absolutely fantastic. But it's sure not easy to implement such a thing.
Pretty much! My current TODO list would take an estimated 3-6 months to finish, only after which I will have time to look at additional features. Btw, it already automatically searches for a new star after the flip and calibration is a good test to see if everything is OK, if it fails, then guiding would probably fail too.