×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

Ekos and Indi - Canon DSLR Issues\Improvements

  • Posts: 9
  • Thank you received: 1
I first want to say that Mr. Mutlaq you and your team have created a fine suite of astrophotography tools, a great driver library, and of course the planetarium software as well. This is what drew me to Kstars\Ekos\Indi, the all in one solution. It will be exciting to see it grow and I am dedicated to being a part of that process. I have imaged with Canon DSLRs for many years, first starting with Nebulosity then to Backyard EOS and now to Ekos.

This past Friday and Saturday were my first nights out with Ekos. I let it loose on two scopes, a Meade 127mm Triplet on a Losmandy GM100 and a Meade 14'' LX200ACF on a Losmandy GM200. The Losmandy GM100 and GM200 are push-to mounts from the late 80's still going strong. The GM200 has digital setting circles (monitored by an Argo Navis) and the GM100 I use the mechanical setting circles. I have a Canon T5i on both scopes and both are off-axis guided. The guide cameras are an SBig STi on the Meade 127mm and a Lodestar on the Meade 14'' ACF.

I experienced a few issues I wanted you to know about, maybe they are already known but if not I wanted to make sure you knew about them, so without further delay here is the list:

1. When framing with a DSLR sometimes 20 seconds is not enough to pick up a faint object and all its details, even at the highest of ISO settings, is there a way to increase this? Remember I have push-to mounts so plate solving is not a solution for me, I must center my targets the old fashioned way. :)

2. Sometimes when framing (Focus tab) my image dimensions would get flipped, meaning instead of 5204 x 3678 as it should be the image became 3678 x 5204 in the Fits Viewer. It did this at random but what was really annoying about it was it would stay with the wrong dimension even if I shut down Kstars\Ekos\Indi. As I said before my cameras are Canon T5i's and I have two of them on two different scopes, two different computers and they both exhibited this random odd behavior. I had to reboot the machines to get Ekos to use the proper image dimensions again. I know there is the oddity with the gp-photo driver in that you must capture an exposure (and maybe send it to dcraw??) to determine the resolution of the camera, maybe this is where the bug (accidental dimension flip) lies. It makes centering a target quite fun as your usual directions are now flipped, lol! :)

3. Again when framing, when using the "Auto-Stretch" filter, this would cause Kstars (Ekos and everything else) to lock up almost as if an endless loop condition had occurred. It may not happen right away, sometimes it does though, but give it 3-5 framing exposures and it will happen eventually. The "Equalize" filter also exhibited this same behavior. I was using 3, 10, and 20 second framing exposures both with darks and lights and it would happen with any of them. When this happens the bottom-left corner indicator on the Fits Viewer is yellow, if that is of any help. I have to force quit Kstars as the lockup does not go away.

4. I am a manual focuser, someday I may have motorized focusers but not anytime soon, what I would really like is if the HFR plot graph also worked in manual mode as well. I saw no plots being drawn when in manual mode, basically you just have the number to go off of. Also it would be ideal if in manual mode one can select the star to use to calculate the HFR against. I have no idea what it is calculating that HFR off of in manual mode? Does it pick a star automatically? Maybe one could use the focuser simulator (so you can use auto focus mode) to get the graph but even then you risk the auto focus routine cutting out short on you, maybe setting a high tolerance would prevent this though, just thought about that.

5. I noticed in Indi when I set the ISO in the Image Settings tab and saved the configuration this was not being saved, it would always default to the highest resolution? I have the setting set to always load saved configuration on connect.

6. In regards to changing the ISO, it is quite difficult to have to find the Indi window and change the ISO from when I am framing\focusing to when I am actually imaging. Of course when framing you want a higher ISO so the target shows up in a shorter amount of time but when imaging you will drop it down of course to avoid higher noise. Maybe displaying an ISO dropdown in Ekos would be helpful, both on the framing tab and on the CCD tab, and in the CCD tab maybe it should be one of the properties of your sequence.

7. The guiding graph line colors are quite hard to see with a red filter over the screen, maybe some configuration of their color and maybe even their thickness is in order at some point.

That is it for now, let me know if I can be of any further assistance. I would love to help fix some of these but I am not sure of the process with that, just let me know as I am a software developer.
Best Always,

Daniel
Last edit: 9 years 3 months ago by Daniel.
9 years 3 months ago #2227

Please Log in or Create an account to join the conversation.


Glad you find Ekos useful! Hopefully it would be a decent replacement, it continues to be improved but I'm the only one working on it, so things might be slow at times. :-)

That's indeed a unique setup! I'm curious to know if you managed to connect Argo Navis to KStars as well? I believe it has a "meade" emulation mode and you can use Meade INDI driver to read the DSC coordinates into KStars.
1. When framing with a DSLR sometimes 20 seconds is not enough to pick up a faint object and all its details, even at the highest of ISO settings, is there a way to increase this?  Remember I have push-to mounts so plate solving is not a solution for me, I must center my targets the old fashioned way. :)

I increased maximum to 300s now, so exposure time is the only limiting value?
2. Sometimes when framing (Focus tab) my image dimensions would get flipped, meaning instead of 5204 x 3678 as it should be the image became 3678 x 5204 in the Fits Viewer.  It did this at random but what was really annoying about it was it would stay with the wrong dimension even if I shut down Kstars\Ekos\Indi.  As I said before my cameras are Canon T5i's and I have two of them on two different scopes, two different computers and they both exhibited this random odd behavior.  I had to reboot the machines to get Ekos to use the proper image dimensions again.  I know there is the oddity with the gp-photo driver in that you must capture an exposure (and maybe send it to dcraw??) to determine the resolution of the camera, maybe this is where the bug (accidental dimension flip) lies.  It makes centering a target quite fun as your usual directions are now flipped, lol! :)

You are correct, if you're capturing RAW images, dcraw is used to get the dimensions. If it's jpeg, the dimensions are read using libjpeg. I have Canon 550D and it will try to replicate this today, can you just confirm whether you were capturing JPEG or RAW images?

Will test this as well today.

The absolute position vs. HFR was meant for autofocus, I can add another graph for manual focus, but what would it be? # of iterations vs HFR? timestamp vs HFR? At any rate, the manual mode asks you to select a focus star and the then if subframing is selected and supported by the driver, it subframes to the selected star. In case of gphoto driver, subframing is not supported, so it tries to detect stars in the whole image, measures HFR for all of them, and then it reports HFR for the brightest star in the field, or average of all stars detected. The stars have to be somewhat focused for the star detection algorithm to pick them up.

The ISO property is dynamically generated upon connecting to the camera, I will try saving it in the configuraiton and testing how it goes.

Sure, I'll add this to my TODO list.

This was pointed out before, another TODO.

Sure, Ekos can only get better from feedbacks such as yours. Will try to sort out these issues soon. Btw, which distro are you running? For you to be able to test the latest, you need to be using Kubuntu 14.10 to use the latest KF5 based KStars & Ekos (kstars-bleeding PPA). Of course, if you are compiling from source, you can use any distro that has KF5 packages in its repos.
9 years 3 months ago #2228

Please Log in or Create an account to join the conversation.

  • Posts: 9
  • Thank you received: 1

I did not but I thought about doing so. I will definitely test this at some point, it should work fine as you said. We will make sure. :)


Excellent, that is more than enough and yes that was the only limiting value.


RAW, always. :)


Excellent as this is quite annoying, don't know how it gets stuck either, does dcraw remain open after using it or does it keep a cache of its own somewhere?


Ah! Ok I did not notice that graph was absolute position vs. HFR. Ok so in the manual focus case I think it should be # of iterations vs HFR. It would also be nice if we had another graph of the FWHM and its associated number (another feature request to the list).


This one is not that important it would be nice to have it save but not required for now.


I am on Linux Mint 17.1 Cinnamon and I am using the kstars-bleeding PPA. What I am trying to find out though is where to get the KF5 packages from. I can build Indi just fine but Kstars of course needs KF5. Akarsh (my friend who also works on Kstars) pointed me here: community.kde.org/Frameworks/Building#Dependencies and it looks promising. Let me know what you think. I will gladly help if desired.
Best Always,

Daniel
Last edit: 9 years 3 months ago by Daniel.
9 years 3 months ago #2233

Please Log in or Create an account to join the conversation.

So this is what I got done:

1. You can select ISO from CCD & Focus Modules. ISO can be seen also in the Sequence queue and sequence lists with ISO can be saved and loaded. Finally, ISO can be saved to configuration file and loaded next time on connection.
2. Guide colors are configurable now via KStars options.
3. Added subframing support go INDI GPhoto driver (only in FITS mode), so after you initial image, you can subframe and it will make focus process easier.
4. Add # Iteration vs HFR plot in addition to Absolute Position vs. HFR

I wasn't not able to reproduce the problem you had while framing, but I _suspect_ it has to do with HFR calculations and on some images, it can get stuck and take forever to solve. Therefore, I'd recommend to subframe to a specific region of interest. Also, I think I took over 100 images and non had flipped values .. not sure what caused your issue. At any rate, all these changes are available in in SVN and GIT, but there is no kstars-bleeding for Ubuntu 14.04 (which is 17.1 is based upon). I'll see a way users of 14.04 can use KDE5 easily.
9 years 3 months ago #2253

Please Log in or Create an account to join the conversation.

  • Posts: 9
  • Thank you received: 1

Wow Jasem, thanks so much! You have just improved the usability immensely for DSLRs. Excellent that you did not see the dimension flip, that really was the most annoying of the issues. Maybe it is an older bug that does not exist anymore. I had thought I was on kstars-bleeding, well I am using the PPA but I am not getting updates from the "bleeding" branch which is for Ubuntu 14.10 I am guessing. You have to bear with me a bit as I am coming from the Windows world of code management (TFS and Sourcesafe) and am trying to wrap my head around how code management is done in the open source\Linux world (SVN, Git, etc.). I am learning though.

What I will do is use an Ubuntu 14.10 based OS, I have no preference really as this is for Astronomy and C++\Java development so it makes no difference what OS I am running. I am not picky as long as it works. Don't spend time on trying to find a way for 14.04 users to run KDE5, I took enough of Akarsh's time asking him already, lol! :)

I will let you know if I think of or find anything else we can fix or improve. Thank you again and I hope to be helping you fix some bugs myself soon.
Best Always,

Daniel
Last edit: 9 years 3 months ago by Daniel.
9 years 3 months ago #2254

Please Log in or Create an account to join the conversation.

  • Posts: 9
  • Thank you received: 1
Jasem,

I think I will go with OpenSuse 13.2 which seems to have QT5 and KF5 packages like ArchLinux, do you see any problems with that path to building Kstars? I should just be able to install the QT5 and KF5 packages needed to build Kstars. Wish I could do it on LM 17.1 but oh well no Trusty PPA to do so.

Thank you sir!
Best Always,

Daniel
9 years 3 months ago #2263

Please Log in or Create an account to join the conversation.

Well, the easiest way is to install Kubuntu 14.10. Then you'd have the PPA and you can build it yourself if you want to.
9 years 3 months ago #2264

Please Log in or Create an account to join the conversation.

  • Posts: 9
  • Thank you received: 1
That will work, one thing I want to clarify though, I do not have to install Plasma 5 right? Can I run Plasma 4 and still build Kstars? That was my original goal (not to have to install Plasma 5 yet) and Akarsh made it sound like that is what he was doing. Since he told me that, it had put me on path of just finding a stable OS that had all the QT5 and KF5 packages in its base PPA. Do you use Kubuntu and is it stable? I have not checked any reviews on it yet.

That is my end goal, to build Kstars myself. Originally I tried following the instructions in the forum post "How to build Indi\Ekos" but I guess that was a bit outdated (you updated since I first read it). :)

Thanks for all your help Jasem!
Best Always,

Daniel
9 years 3 months ago #2265

Please Log in or Create an account to join the conversation.

No you don't need Plasma 5, it's actually not very stable in 14.10, just install Kubuntu 14.10 with KDE 4 and then use the PPA (it will install all necessary dependencies) or build it yourself (I'll add instructions later)
9 years 3 months ago #2266

Please Log in or Create an account to join the conversation.

Time to create page: 0.599 seconds