Lodestar Live - V0.12 (beta)

[HiloDon]

Ram,

Dom is ahead of me in time zones, so he beat me to the answers to your questions. He is on target with everything. I can only add that the 694 sensor looks interesting. Both SX and Atik have it. I think you will find that use of these will move out of EAA and into imaging, and it's an expensive experiment. The cooling of these cams does eliminate the need for dark frames from my understanding.

I also agree with Dom that we should probably move this to it's own thread if you want to discuss more on this. It is an interesting topic that warrants further exploration.

Don

[aparker]

I

Ram,

Dom is ahead of me in time zones, so he beat me to the answers to your questions. He is on target with everything. I can only add that the 694 sensor looks interesting. Both SX and Atik have it. I think you will find that use of these will move out of EAA and into imaging, and it's an expensive experiment. The cooling of these cams does eliminate the need for dark frames from my understanding.

I also agree with Dom that we should probably move this to it's own thread if you want to discuss more on this. It is an interesting topic that warrants further exploration.

Don

I'm also toying with the SX694 idea.  Agree that 2x2 binning would be the right way to go, and LL would need to support that.  A binned 694 would have pixels the same size as the Lodestar's, in an array twice as wide and deep.  I'm loving the idea of a bigger sensor with similar sensitivity especially if I ever get around to adding tracking to my 18" Dob.  Now there's an EAA fantasy - 2MP sensor, 450mm aperture, 1900mm FL, realtime imaging via LL.  I believe the download times for a binned image would not be limiting - perhaps stacking would be.  But that can ultimately be solved via code optimization and/or more processor power.

[Paul81]

[sorry for taking ages to reply..]

Hi Paul. Many thanks for Lodestar Live! I have both X2-M and X2-C and am curious if there is any way to run two instances of LL on the same PC, one for each camera?

Not currently - but this is on the radar for the future...

As for other SX cameras, on OSX communications with the camera are 100% my own code (I use libUSB for low level comms, and on Windows I only use the very low level component of SX driver, the rest is all my own code and is the same OSX uses - I wrote my own code so it would work on OSX).

SX are very good and supply the register maps for their cameras on the developer part of the website, so I can develop my own control code 

LL **should** work with all SX cameras, at the moment it doesn't bin or control shutters and coolers, but I think EAA is soon to start migrating to the higher res cameras....

Currently, LL only searches for the Lodestar camera, but via the command line you can override that to connect to other cameras.

As its all a bit experimental, for now if anyone has a higher res SX camera and would like to try it out, PM me. It may not work 100%, but can normally work through issues. If using a high res, the display processing might be a bit slow, as I have done all my performance testing based upon the resolution of the Lodestar (I still want to optimise things more so it will only get faster...)

Interesting times ahead 

[jam1e1]

Hi Paul, just bought a LX2C for Use with LL, great program only managed to use once so far on manual mount for testing , pointed towards Polaris area to reduce star trailing, was promising and quite addictive [emoji1] managed to capture accidental meteor or satellite too! Look forward to spending more time on having a proper go once gets darker. Will be interested if this does develop with other cameras too.... Though that will take a lot more saving up!

[Ain Soph Aur]

Not currently - but this is on the radar for the future...

....

Interesting times ahead 

I did further research after my original query and thought your reply would be that would not be possible. The ability (considering a beefy PC) to use both X2's with two LL's apps at the same time on the same PC with separate OTAs would be simply amazing.

But... after reading further into your plans of the possibility of....

on the fly multi-spectral stacking and processing - EAA does the hubble palette!

is if MUCH more interest.

Thank you Paul!

[ramdom]

I am having fun with my Lodestar X2m  (I tried it properly for the first time today) - it is incredibly sensitive. Thanks a lot Paul - it is your software that makes it all work out in this case.  

I am wondering what effect the colour controls have on the mono version of the camera. When I turn on the one shot colour option, I am able to colours in the image and seem to be able to change them with the two colour controls but I've not really played with them to see if they are really meaningful. So before I do that, I wanted to ask if there's any point to playing with the colour controls if you have the mono version of the camera?

I'm also getting a Mallincam Skyraider DSc soon so I'll be able to compare/contrast the two cameras/software with similar sensors.

--Ram

[Paul81]

Hi Ram,

You do not need the OSC checked if you have a mono camera - you can still play with the colour controls. With a mono camera there isn't a great deal of point - just keep the Modify All on the display processing tab checked. 

The controls are the same for colour and mono to keep things consistent - plus they will have a use for the multi-spectrum stuff if you choose to use filters with a mono camera (up and coming in V0.13).

[Ain Soph Aur]

The controls are the same for colour and mono to keep things consistent - plus they will have a use for the multi-spectrum stuff if you choose to use filters with a mono camera (up and coming in V0.13).

This! Thanks   Paul!

[Dom543]

I am having fun with my Lodestar X2m  (I tried it properly for the first time today) - it is incredibly sensitive. Thanks a lot Paul - it is your software that makes it all work out in this case.  

 

I'm also getting a Mallincam Skyraider DSc soon so I'll be able to compare/contrast the two cameras/software with similar sensors.

 

--Ram

Ram,

Nice to hear that you are enjoying your Lodestar X2. Please post images as soon as you have some that you are happy with.

There seems to be some problem with the Mallincam Skyraider DSc. The odd and even scans don't line up and Rock Mallin says the problem is beyond his abilities to fix. He offers to use payments already made towards his other higher priced cameras or a refund. Here is his post. https://groups.yahoo.com/neo/groups/mallincam/conversations/topics/66945.

You were smart to hedge your bets. This way you have at least one camera that works and are not losing the precious summer observing season.

Clear Skies!

--Dom

[ramdom]

Hi Dom, thanks. I am confident Rock will get the issues with the colour version of the SR-DSc worked out but for now,  I've realised that as with scopes, different objects require different types of cameras and I think mono cameras/images are particularly well suited particularly to galaxies and also to globular clusters. So I eventually would be happy to have both cameras and approaches to EAA and image acquisition in my arsenal.  I think having LL will serve me well in my quest for higher resolution since the Mallincam Universe doesn't have software that works with the Mac.  I also think the "live stacking" approach that LL takes is a really awesome - I don't know of any other piece of software that does this.  If you look at my stacks, you really can see the image improves on the fly as it goes.  So even though I'm not full polar aligned and haven't optimised my mount for tracking, I am able to use 15-30 second exposures and build upon it easily. 

In the mean time, I have taken a few images with LL v0.12 and they've come out okay but today I had the best session working on the Fireworks galaxy NGC6946.  I still am feeling my way through the settings and the effects of stacking multiple images and the effect of parametres, etc. and I still have a number of questions. I've put my entire series of captures at the URL below, starting with a 2 x 25s exposure (I found at 30 seconds my stars started to look slightly elongated---I am not fully polar aligned), and going all the way to 33 x 25s without any filters and the last two images (including a final 43 x 25s exposure) are with a couple of filters that seem to reduce the artifacts but cause the image to look a bit blurry.

The Untitled.jpg image is a stacked image of all these processed using Nebulosity.  So if one wants to look at only a single image, they should see (also attached):

http://ram.org/images/space/scope/1/1/ngc6946/ngc6946_2015.7.17_01.24.48.png

and if one wants to check out the progression (and help answer my questions below), then all the others are here:

http://ram.org/images/space/scope/1/1/ngc6946/ 

So my questions/issues - I apologise if these are answered somewhere but aside from the Live Stacking notes I've not found any other documentation. 

As you can see, the more I stack, the better the image starts to look. But I also start seeing these inverted V shaped artifacts - are they stars or noise or satellites or what? They seem to show up consistently. Yet I like my final images since the galaxy itself looks good.  And if I do any kind of noise reduction (either with the filters or in Nebulosity) they seem to go away or be minimised. 

Do the png files generated by LL have some kind of layering? Sometimes when I look at them at a particular size, I see a LOT of detail/info and sometimes it looks very dark. 

Speaking of layers, what is the effect on the stack if I change the parametres - I know some parametres like the filters are only applied to the output image but what about things like the contrast - if I change the contrast on the fly, does it get applied to the entire stack or just from that point forward and then does it get averaged (or summed or median shifted)? 

If you look at my processed image in Nebulosity, you see these cavity-like structures. I personally like them aesthetically but are they an artifact of processing or something real? 

--Ram

[ramdom]

In the above URLs, there is character at the end that is causing a 404 not found error - if you get that and point your cursor to the very end and delete it (it should look like a space) and then refresh, then the URL should show.

Anyways, this should give people a good idea of what a total noob without any experience with anything can do - this is really the second time I've used this camera/software and even the images from yesterday are pretty decent for EAA purposes but I think the picture of the Fireworks Galaxy I took is as good as the one I got from Slooh for my purposes anyway...

Thanks again Paul - this is a great contribution to EAA! 

--Ram

[aparker]

Ram,

Welcome to the Lodestar club!  Your results look great.  The V-shaped artifacts are "hot" pixels - individual pixels on the Lodestar chip that are "stuck on".  You can minimize this effect by taking dark frames before capturing light frames of your object of interest.  LL allows you to do this.  It is very interesting that your hot pixel artifacts look like little sections of a sine wave.  This might possibly point at some very substantial periodic error in your mount - i.e. it went back and forth and started back again during your series of exposures.  Are you guiding in any way?  Does the Star Sense thingy do anything whilst the mount is tracking?

In general I think a C9.25 is a big scope for imaging applications on an AVX mount.  I run my C8 on a CGEM, by way of contrast... 

Also, you likely don't need subexposures as long as 25 sec.  Martin Meredith has posted extensively here documenting that equivalent results can be had by stacking 10 or even 5 second subs (just more of them).  I essentially always default to 10 sec subs these days.  I could go even shorter but occasionally LL has trouble stacking fast enough to keep up with e.g. 5 sec subs.

Good luck,

Alex

[ramdom]

Hi Alex, thanks - no, I am not guiding yet and my polar alignment error according to the mount is 32 minutes N and 1.14 degrees E (which seems large but I've been incrementally working on this and the movements I have to make to get to this point have been very very slight---it's a PITA to move my mount in AltAz). I've not done any of the PEC calibration, etc. My long term plans are to get  a carbon fibre refractor that is about half the weight as the maximum supported by the AVX for more serious imaging work - this is the only mount I can lift easily.  (I'm thinking an ES 127mm CF ED Apo.) I don't think StarSense does anything during the tracking - it is only for GoTo and I even keep the lens cap on most of the time. But all that shouldn't really matter for EAA right? It's WYSIWYG sort of thing? I can see the object and there is little vibration.  I will be getting an OAG soon I can guide with the ASI 120 MC-S and image with the Lodestar (and also the SR-DSc). 

I did take dark frames but they were from the night before.  Are they not used automagically? I guess one needs to plan a bit ahead---if the live image starts to look good, take a master dark with the same exposure as the last one.  Thank you letting me know these are the notorious hot pixels. I thought they were randomly distributed from frame to frame. 

Yeah, I think shorter subs may help with everything but I was just seeing how long I could go with all the errors I still had to fix.  

BTW, I'm using only the F6.3 reducer for this but when I get my C mount to 1.25" adapter that I've ordered I should be able to try out my 0.5x reducer additionally and while that'll give me a wider FOV and less exposure time needed, I think for this object I had the right setting. So not all objects should be imaged or seen the same way... 

--Ram

[HiloDon]

Hi Ram,

Congrats on your new LS.  To answer your question about changing parameters I can say the following.  LL image adjustments are done to the displayed image. It makes it real easy to try different things, even using the non linear scaling.  If things get out of sack a bit, just hit reset and the controls go back to default and you can start over.  Even if the image is stacked, the changes will be made immediately to the final complete stacked image.  You don't have to wait for another exposure cycle to see the adjustment.  My understanding is that the new SkyRaider DS will require another cycle to see an image change.  One thing to focus on in LL is the histogram.  The black and white adjustment are critical in getting the image and background adjusted properly.  The contrast setting will broaden the histogram, and brightness will move it right or left.  Usually to get a nice dark background The black level needs to be moved close to the histogram image curve on the left.  Going past the peak will usually be too dark. once you get a dark background, move the white level to the left to brighten the image to your liking.  It usually a good idea to broaden the histogram image curve with the contrast setting first, but there isn't any special order required.  If things start looking bad you can always reset.

I suspect that you may not have tried these adjustments, so I took your image and opened it in Nebulosity 3.  By just adjusting the white level I got the image below.  The Fireworks is a tough subject to start with, but I was really impressed with what you got.  I agree with Alex that the sine shaped artifacts are from tracking with periodic error.  Usually regular tracking errors result in a straight line.

Hope this helps.

Don

P.S.  Hope it was OK to change your image.

[HiloDon]

Ram,

Just one thing to add.  If you make all your image adjustments before you stack, that's ok, but even mean stacking sometimes shifts the histogram curve, so the black and white levels need to be readjusted.  Sum stacking will definitely require adjustment.

Don

[ramdom]

Hi Don, thanks a lot and the image you produced looks awesome. Yeah, I think the fact that the camera and LL do so well on this target is indicative of how good the system is and I felt it was just right for the FR I was using (the galaxy frames the image size entirely).  

My brightness and contrast settings are so that brightness is set to 0 and contrast is set to maximum. I will be sure to pay attention to the histogram and try your advice out next time I go out... thanks a lot for your help! 

--Ram

[ramdom]

I have a question about the dark frame subtraction: I took a bunch of them at different exposure lengths at the start of session and they are saved as Masterdarkframe*. I'm wondering how they are used and if they are used automatically and how does the software know which ones to use. Are all the ones with the same exposure length used? In the live stacking documentation it says that one can collect these dark frames can be used across multiple observing sessions. So my question is: how does the software determine which frames to use if I've collected a bunch of them across different sessions with different temperatures, etc.  Thanks. 

--Ram

[HiloDon]

Hi Ram,

Once you have your dark frame masters saved, you can restore them by going into the Dark calibration tab in theDark Frame Aquisition mode and hitting the restore button. Then select your saved master for the exposure you are running and it will overwrite the existing dark frame if there is one and then automatically appliy to your next exposure.

Paul has suggested updating your darks as the evening goes on due to changing conditions. Usually four or five frames are sufficient to get a good master. I would not use masters collected at a different session. You can, but it's best to get new ones for each session.

Don

[Paul81]

Hi Ram,

You should expose your darks for the same exposure period that you use for the object you are observing - i.e. if you are using 20s exposures then in Dark Acquisition mode also use 20s. I recommend getting at least 5-8 darks.

Once you have these then switch to observing, I tend to take darks as I am setting up, and as Don pointed out, top it up every 30 mins or so with 2-3 new exposures.

The Dark subtraction can only handle one exposure length at a time - i.e. the dark you see on the display is the master dark and that is what is subtracted from each new exposure. Therefore it must match the exposure length you are currently using.

You can save and restore your master darks. This is useful if you have a couple of evenings viewing in a row and the outside temperature is similar. So save your mater dark from night one, restore it on night two and add a couple of top ups. Really though its best to take new ones each night - and I find it doesn't cause any issues if you just do it whilst setting up (I leave my scope & camera outside cooling for at least an hour, then as I am setting my mount up just before observing I start the darks off).

The other good use is say you started out with 20s, and then switched to 5s, and then wanted to go back to 20s, you can restore your 20s master dark from earlier.

Taking darks of different lengths at the start is not recommended, keep them fresh as temperatures change.

To be honest, I tend to stick to one exposure length in a session and just do routine 2-3 top-ups each 30-60 mins.

Paul

[Hibou]

Hi Paul81. As a simple physicist I don't even own a telescope :-) But I do use various SX (& Atik) CCD units to construct cameras for scientific imaging, and find LL quite useful (along with PHD and SX/Atik's own software). It would be good if I could link to a download site that hosted the latest version of LL, not just a specific version on Dropbox, or else tell people "join this forum and search for it". (BTW Is V0.12 the latest version, and V0.8 an old version ?) Why not set up a website to host LL ? Websites are easy and cost only €10 per year. I'll even host it on my own site if you wish, and probably sxccd.com would do the same.

But then the Readme.txt that comes with V0.12 says:
"Redistributions may not be sold, nor may they be used in a commercial product".

I thought the whole idea was that LL was meant to be used with commercial products (SX cameras)? Does that mean that no-one should give away a copy with their product, or put it on their own website? Normally freely available software only requires that any redistribution be complete and unchanged, and include the original copyright notice. 

Thanks again for creating a great application. It appears to handle the lodestar even better than PHD; I particularly like the simplicity of the LL interface (except that the exposure time should be automatically set when entered, without a button :-)

[Paul81]

Glad LL is proving to be useful and nice to hear it is being used outside EAA 

Version V0.12 is the current version, V0.8 if quite old now! I can say that sxccd.com will be hosting future versions, and I still intend posting here on SGL too (was wondering if the mods wanted to make the latest version a sticky post)?

As for the licence, I see the hardware as a separate entity. I don't mind LL being distributed, but it that instance I'd just like to know. As for being used in a commerical product, my aim there is I didn't want the software being integrated into something (for example a portable device) and that being sold as an out of the box device for EAA without my consent.

Keep any eye out for V0.13 - that pesky set button has gone...

[Hibou]

Thanks for promising to get rid of the "Set" exposure button Paul. I have been testing V0.12 with a lodestar X2 monochrome, background only, but that is already instructive. Changing the exposure should also reset the current image, and it should be possible to set a specific number of exposures.

The biggest problem for me is that although the individual frames are exported as 16-bit *.FIT files, the "Current Processed Image" stack is only exported as *.png RGB - a simple window dump. It would be much better to export it as FITS, otherwise people will always need to restack the individual *.FIT frames to obtain the best image. But perhaps that's what users are expected to do, and the stacked image is only to look at in real time ?

Then if I flip between "Stacked" and "Raw Exposure" on the "Stacking" tab, I see the white noise as individual pixels for the latter, but slightly smeared out for the stacked image. I guess this is because the "Max Pixel Displacement" is at least 1, so LL always tries to register frames, even with just noise. I would like to have 0 displacement as an option to turn off automatic registration.

Again, the "Filter" is meant to smooth the image, but it also smooths the noise. I would like to see a true noise filter that identified hot pixels and replaced them with a 2x2 or 3x3 average of surrounding pixels (cf PHD). OK, I know that astronomers like to take Dark Frames to subtract background, but with LL you are attempting to make observation as "live" as possible. Collecting darks whenever the temperature changes seems a little contrary to that principle.

The objective IMHO should be to make real-time observation as simple as possible i.e. as simple as aligning and looking through the telescope, with automatic corrections where necessary. But to save the raw FITS data for off-line treatment to extract the best image. 

Incidentally, ImageJ is an excellent free program for scientific image treatment, comparable to Photoshop for illustrative imaging. ImageJ already has lots of routines for filtering noise, stacking frames etc. It is funded by the NIH but has been adapted for astronomy, since people write their own plug-ins to do whatever they want. I would recommend taking a look at ImageJ's "Process/Noise", "Image/Adjust" etc as well as plugins from Atik Cameras, for de-interlacing etc...

[Dom543]

Changing the exposure should also reset the current image, and it should be possible to set a specific number of exposures.

I disagree with this. There is a reset button to kill the current live stack and reset the image. Merely changing the exposure time shouldn't do this.

When stacking frames made with different filters, it is often useful to also use different exposure times. As an example, when I stack H-alpha filtered exposures together with O-III or S-II ones, I almost always use shorter exposure times for the H-alpha. This is to prevent the much stronger signal overwhelm the weaker ones.

Clear Skies!

--Dom

[Paul81]

The biggest problem for me is that although the individual frames are exported as 16-bit *.FIT files, the "Current Processed Image" stack is only exported as *.png RGB - a simple window dump. It would be much better to export it as FITS, otherwise people will always need to restack the individual *.FIT frames to obtain the best image. But perhaps that's what users are expected to do, and the stacked image is only to look at in real time ?

The objective IMHO should be to make real-time observation as simple as possible i.e. as simple as aligning and looking through the telescope, with automatic corrections where necessary. But to save the raw FITS data for off-line treatment to extract the best image. 

A button (on the stacking tab) to export the current live stack to a FITS file is a neat idea - have added this to the wish list for V0.14.

 

Then if I flip between "Stacked" and "Raw Exposure" on the "Stacking" tab, I see the white noise as individual pixels for the latter, but slightly smeared out for the stacked image. I guess this is because the "Max Pixel Displacement" is at least 1, so LL always tries to register frames, even with just noise. I would like to have 0 displacement as an option to turn off automatic registration.

 

You can turn off live stacking by un-checking the 'Enable Live Stacking' checkbox. The displacement sets the maximum image displacement from the last key frame before a new keyframe is generated.

 

Again, the "Filter" is meant to smooth the image, but it also smooths the noise. I would like to see a true noise filter that identified hot pixels and replaced them with a 2x2 or 3x3 average of surrounding pixels (cf PHD). OK, I know that astronomers like to take Dark Frames to subtract background, but with LL you are attempting to make observation as "live" as possible. Collecting darks whenever the temperature changes seems a little contrary to that principle.

 

Again a neat idea, and I have taken a note as something to look at in the future. Not having to take dark frames would be nice. I find with the Lodestar I don't need to update them that often, take the initial 5-8 darks and then maybe update once or twice in the session.

 

Changing the exposure should also reset the current image, and it should be possible to set a specific number of exposures.

 

​Also agree with Dom on this one!

 

 

 Thank you for the feedback - its is all valuable and all suggestions are welcome. Now the app has become quite popular it can be hard to cater for all uses! 

[arkosg]

Hi Paul,

I know you have a lot on your plate, particularly now that SX is using LL (congrats!), but would like to 

put in my couple of requests for future versions of LL:

1) a "gray point" histogram adjustment in addition to being able to set black and white point

2) look into feasibility of enabling "monochrome" mode during imaging for X2C (like what is done in Focus & Framing mode)

3) would it be difficult to implement flat frame corrections similar to how you have implemented dark frames?

Thanks again for a great bit of software - can't wait to get back out there again...

Cheers,

- Greg A