Dithering preperation

[ianpwilliams]

I'm going to try dithering for the first time tonight using APT and PHD2. I've tried to work out what I need to do, having looked through various SGL threads and the APT manual. It seems like this is what I would need to do, although I do have questions when it comes to the troubleshooting:

PHD2

 

Click Tools, then click Enable Server to enable server mode in PHD2 

Adjust the dither multiplier from the "Brain" dialogue window - set it to maximum

 

APT

 

Shift-click the "Guide" button to set up dithering parameters

Guiding program - PHD2

Dither Settle Time - 45 (could try 30 too - the dither notification in the PHD2 guiding graph will show you whether the dither returns to zero before APT takes the next sub)

Once APT has taken the first sub, check the image in APT and make sure there are no star trails

 

But I'm not sure what to do when it comes to troubleshooting:

 

Not enough movement (what does this mean?) - increase the "Auto-Dithering Distance" (how do you know if the movement is not enough?) (increase by how much?). If it is still too small then change "Dithering Scaling" in the PHD2 "Brain" box

 

Star Trails - decrease the "Dithering Stability Distance" value (decrease by how much?)

 

Timeouts (what are timeouts?) - increase the "Dithering Timeout" (increase by how much?)

 

Please can someone advise me as to what the above means? If it turns out that dithering doesn't work using the default parameters then I'd like to have some idea of what I can do to fix it, rather than just not dithering.

 

Also, do I set up dithering on PHD2 before or after starting guiding? My guess would be before, as I'm not sure if PHD2 would allow me to put it into server mode if it was already guiding. But I may be wrong.

 

[PhotoGav]

Hi Ian, you are very wise to sort out dithering as it will definitely help to improve your final images and don't worry too much, it's really not too tricky. Thankfully APT & PHD2 do all the hard work for you. I will try to answer your questions and at least let you know what I do with my set up...

PHD2 - the only thing I have ever done in PHD2 is to Enable Server so that APT can talk with PHD2. I didn't even know that Dither Multiplier existed in PHD2, so I'm sure the default settings will do you fine!

APT - this is where all the values need to be set. As you say, Shift-Click the Guide button to access the menu, then the settings I use are roughly as follows:

Guiding Program = PHD Guiding

Auto Dithering = On

Dithering Distance = 4 or 5

Mgen Mode = not relevant for PHD

Dithering Stability = 0.5 or thereabouts, depending on your exact guiding system and the normal stability you expect while happily guiding, so for my 80ED I reckon to be down at around 0.10 whereas with my EdgeHD 800 it is higher at around 0.4. So, this value lets APT know what you consider the value to be that indicates that PHD guiding has settled down to normal again and everything is good to go for the next exposure.

Dithering Settle Time = 30. This is the minimum time for APT to wait for PHD guiding to get back to normal before even considering the next exposure.

Dithering Timeout = 60, or thereabouts. This is the time after which, even if PHD hasn't settled down to below the Dithering Stability value, APT will just start the next exposure anyway.

Auto Cancel Exposure = Off. I don't use this, so can't offer further info.

PHD IP Address = Leave as default, I don't remember having to change this, so the default must be fine. I think it is there so that you can control a different PC from here if necessary.

As for troubleshooting... Give this a go first and it should work pretty successfully. You should notice a big bump in the guide graph as dithering happens and then the settle down period elapsing. The next exposure should then happen and there should be a small but discernible difference between the exposures when you flick between them. Good luck and report back if it doesn't go to plan!

[ianpwilliams]

Thanks PhotoGav, that's extremely useful info.

 

So out of those parameters it looks like Dithering Stability is the one that's most open to interpretation. Unfortunately as I'm also new to guiding I'm not entirely sure what value I would be aiming to begin with based on my setup. Is there a way I can come up with a rough idea for it?

 

Also, regarding the Dithering Settle Time and Dithering Timeout, that suggests that there will be a gap of 30-60 seconds between each sub, which presumably I will need to factor into my total imaging time? I do have intervals anyway to allow the sensor to cool, but I guess the Dithering Settle Time will override any interval that I specify (should my value be less than 30 seconds)? In which case I could just set the interval time to zero? 

 

Also, do you set up dithering before or after you start guiding?

 

And should I need to abandon dithering during the session because it's not working for whatever reason (but I still want to get some imaging done and take darks instead), is that easy to do? Presumably I would take PHD2 out of server mode, but what about APT? Would I need to reset the settings via "shift-click guide", or would taking PHD2 out of server mode be enough?

[Marky1973]

Thanks Ian and Gav....I've been dithering about dithering for a while and wasn't quite sure what to do, so this is useful info....especially as there is a clear night forecast tonight...

Oh, yeah, I probably shouldn't have said that......

[PhotoGav]

My pleasure. Now to try and answer the questions:

Dithering Stability. That is the figure that appears in PHD2 as RMS Error. It is the TOT value in brackets, the arcseconds one that you are interested in. You could start guiding and see what that is on average, that will give you your 'settled' value. You will also see the value in APT once you have connected to PHD2, it appears in the top right info area as 'PHD State', keep an eye on that during your exposures and see what it is running at. Plug a bit higher than that into APT as your Dithering Stability and APT will wait for the value to fall down to below that before it even thinks about the next exposure. To be honest, this is a bit of a safety net to stop APT from starting the next exposure if things have gone wildly wrong with the dithering! Your system should settle down very quickly and be more than ready to carry on with the next exposure well within 30 seconds.

The Gap. Yes, it does create a delay between exposures, though I think that the dither delay and your delay in the imaging plan will start at the same time and nothing will happen until the longest one has expired. Only a bit of experimentation will prove that one way or the other. How long is your plan delay between exposures? I would leave everything as is and see what happens!

When you 'Enable Server' in PHD2 in the Tools menu, it will stay selected until you deselect it. So every subsequent time you open PHD2 it will start the server - Fire and Forget! The same is true in APT, when you have set everything up in the 'shift-click Guide' routine, it stays selected until you alter it. So each Lights plan executed will use dithering unless you tell it not to or turn off dithering. I generally open APT first, do my focussing, get on target, then open PHD2, calibrate and then start guiding.

If you need to stop dithering, then you should just turn 'Auto Dithering' to 'OFF' in APT. You shouldn't need to worry about the server in PHD2. I am confident that you won't need to turn it off though, it works really quite easily. Give it a go and you might just be surprised!

One other setting I have just seen as I have now opened up APT on my astro-laptop, is 'Dither on # images'. I have that set to '1' so that it dithers after each exposure.

I intend to be imaging tonight, so long as Marky1973 hasn't completely cursed the night, so, feel free to PM me or ask more questions should things not go quite to plan! You'll be fine though and you will have beatifully dithered subs with much less noise as a consequence. Just to say though, I always use Darks as well as dithering - the more noise that is eliminated the better for the signal and all I want in my subs is lovely clear sharp bright signal!

[ianpwilliams]

Great stuff, thanks again, looks like I'm ready to go.

 

As for Darks, my time is always fairly limited due to having to set up and pack away each time, combined with early starts in work the next day, not to mention that the nights are now getting shorter. But I'll see if I can find time to take a few Darks as well as dithering.

 

[PhotoGav]

When I used a DSLR I created a library of darks by putting the camera in the fridge. It's not perfect, but it does work and I got pretty good results. The temperature is near enough to a cold winter's night to get away with it and it is certainly a lot easier to use MasterDarks from a library than to go through the hassle of taking darks every time.

My family did give me some strange looks while the camera was in the fridge for a couple of days with a lead coming out to the laptop!

 

[ianpwilliams]

Now that's an idea, I might even give that a try at some point!

 

I have attempted to keep a library of Darks from my previous sessions (which are actually from last winter). The temperatures tend to range from around zero to five degrees.

 

I always assumed that in addition to the sub length and the ISO having to be the same as per the Lights, that the temperature of the darks had to be exactly the same as the temperature of the Lights. But it sounds like that might not be the case? And if the temperature doesn't have to be exactly the same, how many degrees could you risk being out by I wonder?

[PhotoGav]

Ideally the darks would be at the same temperature as the lights, but practically you can get away with a few degrees difference. I don't know how great a tolerance there is, but I never had any problems with the fridge darks. Well, it was more a case of having plenty of other problems that were worse than the fridge darks not being perfect!!! Noise was always the killer with the DSLR at long exposure times for me. I love my cooled CCD! Definitely worth a go with the fridge darks.

[ianpwilliams]

Yes I'll definitely give that a go. I loathe taking Darks, sitting there in the dark and cold with beautiful clear skies overhead, while my camera spends an hour taking photos of the inside of my telescope cap! Very frustrating!

 

And would you suggest that I set up my dithering after I've started guiding? After all, it looks like I'll need to be guiding anyway in order to find out the value for my Dithering Stability setting?

[PhotoGav]

It doesn't matter when you set up dithering as it will only kick in after the first exposure. However you do need to Enable Server in PHD2 before you start the APT plan otherwise APT won't be able to talk to PHD2.

In reality, this evening, you could get on target, start guiding with PHD2 and give it a few minutes to see what your Total RMS value is. Then go into the 'shift-click Guide' settings in APT and enter all your settings. Then hit go on a plan with a few short exposures, say 15 seconds so that you can at least see the stars move from sub to sub. Then sit back with your fingers crossed and see if everything works perfectly!

[ianpwilliams]

Thanks, I'll try what you said. I've checked my so-called Darks library and it doesn't look good (30 seconds at 800 ISO, 90 seconds at 800 ISO etc). So maybe I'll give Darks a miss tonight, and try the fridge method on the weekend instead.

 

Exciting stuff! It wasn't long ago that I was imaging unguided, with no laptop, no light pollution filter, no dithering etc. How times have changed! Here's hoping the light pollution filter and the dithering work out fine. Good luck tonight, fingers crossed we all get clear skies!

 

[ianpwilliams]

Well that seemed to go pretty well. I had another go at Andromeda because it's always been my favourite object, despite being a pain to image!

 

Guiding worked again, and Dithering seemed to work too. I did seem to get star trails at first though with shorter test subs (which I'm pretty sure was caused by the dithering because with non-dithered guiding the stars are always perfect), but then for some reason the longer subs had no trails.

 

I was getting 60% histogram peak at 800 ISO 5-min subs, so I stuck with that. The Lights did look quite saturated (strangely looking worse in APT than in the camera screen, despite the histogram matching), but there is obviously processing to be done.

 

And I saw no light pollution in my Lights either, as opposed to my previous bright orange subs, so the IDAS filter is clearly helping.

 

I also took 20 flats. And the temperature is supposed to be the same on Thur and Fri night as it was tonight, along with clear skies tomorrow night and cloudy skies on Friday night. So my hope is to do another session tomorrow night, and then I think I might lock the camera in the shed on Friday night and let it take loads of Darks. It seems a bit more sensible than the fridge option!

 

So yes, I'm hopeful that I will have some good data. I'll find out when I get home from work tomorrow.

 

As for tomorrow's session (I'd it happens), presumably if I wanted to I could just do the same thing again, same target, and stack both sessions together. Is that worth considering do you think? Or is it a real pain to do?

 

[PhotoGav]

Excellent news, well done! I look forward to seeing your image of Andromeda.

As for data for the same target across multiple nights imaging - definitely, it is the only way to get substantial amounts of data. Plate solving will get you perfectly back on target.

[ianpwilliams]

I can't remember if I used plate solving to centre Andromeda last night or not. I have a feeling that I may have just used GOTO and chose M31. Because I remember being surprised that it was dead centre straight away. I wonder if there's a way of finding out? It's not the end of the world if I have to go for something else next. But certainly I'll make sure to use plate solving next time so that I can do the same thing both sessions.

 

Across multiple sessions, even if the target was dead centre, wouldn't it still be rotated slightly from night to night? Does DSS allow for that when stacking? Or would you stack the nights separately and then layer the two processed images (something that I suspect I might struggle with)?

[PhotoGav]

It doesn't matter if you used plate solving for a previous session or not. All it does is centre you accurately on a set of RA & Dec coordinates. So, if you know or find out the coordinates of your image from last night (by a blind plate solve perhaps), then you can centre up on that at the next session.

The rotation will be small, especially if you don't move the camera on the scope, and alignment / stacking programs can deal with that easily. I don't use DSS (I use Nebulosity on a Mac) so can't give you instructions, but I'm sure that is part of its standard routines.

[ianpwilliams]

Would it not also be an issue that I set up each time, and possibly in a different part of the garden (I have to move around to allow for a couple of tall trees)? And when you say a blind plate solve, would that be doing a plate solve in APT of one of my subs and then taking a note of the coordinates? Having said that, surely the coordinates would be different if I was set up in a different place?

 

Also, will my PHD2 TOT value be the same from now on, or will it change from session to session? Incidentally the values in PHD2 and APT didn't seem to match, so I went with the PHD2 value as that seemed to be the most relevant figure of the two.

 

[PhotoGav]

Thankfully the coordinate system of the sky doesn't change as we move around, so once you have RA & Dec for your particular target it is fixed in stone (well relatively, a few thousand years and it will have moved a bit, I digress...) and will work all over the globe, though it might be below the horizon!

Blind Solve - make sure that you have downloaded Plate Solve and set it in the Point Craft bit of APT. Open an image in APT (using the IMG tab), then click 'Blind' in the Point Craft window. It should eventually come back with the RA & Dec values. A really nice feature in APT is the 'Custom' tab of the Objects window, where you can store all your own targets with coordinates and other bits of info. Makes going back to a target really easy.

The PHD2 TOT value will change from session to session, as it does within a session. However, it shouldn't change massively, so your settings should remain good. Yes, the PHD2 value always seems to be a bit more than the APT value, not by masses though.

Hope that all helps.

[ianpwilliams]

That's great, thanks.

 

I've taken some Darks this evening, and I'm trying to stack everything in DSS, which I know you don't use, but maybe you'll know the answer anyway. I need to know what settings to use in DSS to let it know that I'm stacking dithered Lights. Now I thought someone once mentioned something about selecting a particular stacking mode (there is a choice of Average, Median, Kappa-Sigma Clipping, Media Kappa-Signma Clipping, and Auto Adative Weighted Average). The beginner's guide to APT that I have says to use "Average" generally-speaking, but for dithered Lights would I use a different mode do you think? Or maybe change a different setting? Or should DSS theoretically recognize the dithering and allow for it?

 

[PhotoGav]

Mmm, I don't know the answer for sure, but my guess would be that DSS will be absolutely fine with the fact that the subs are dithered and will align and stack them with ease. I think the best method of stacking is Kappa-Sigma Clipping as that will eliminate outliers such as satellite trails. A DSS user will be able to give you an exact answer though.

[wxsatuser]

If you have more than a dozen or so subs use kappa sigma or median kappa, this will remove planes, sats and colour mottle.

You don't have to tell DSS about dithering as it will register the stars and then stack them on top of each other.

The only other thing I do is use Mosaic Mode under the result tab.

 

[ianpwilliams]

Thanks, I'll give that a try.

[ianpwilliams]

11 hours ago, PhotoGav said:

Thankfully the coordinate system of the sky doesn't change as we move around, so once you have RA & Dec for your particular target it is fixed in stone (well relatively, a few thousand years and it will have moved a bit, I digress...) and will work all over the globe, though it might be below the horizon!

Blind Solve - make sure that you have downloaded Plate Solve and set it in the Point Craft bit of APT. Open an image in APT (using the IMG tab), then click 'Blind' in the Point Craft window. It should eventually come back with the RA & Dec values. A really nice feature in APT is the 'Custom' tab of the Objects window, where you can store all your own targets with coordinates and other bits of info. Makes going back to a target really easy.

The PHD2 TOT value will change from session to session, as it does within a session. However, it shouldn't change massively, so your settings should remain good. Yes, the PHD2 value always seems to be a bit more than the APT value, not by masses though.

Hope that all helps.

It's looking like clear skies tomorrow night so I want to get back out there. If I were to give Andromeda another go (assuming I can blind plate solve before tomorrow night and have the coordinates ready), what would the process involve? I guess I would do the same again, same settings for the Lights, guided, dithering etc, and as many as possible, and as many flats as I can manage too (I think I took 20 last time)? And then I guess I would use the same Darks in both stacks that I took last night (I have 20 of those)?

And more importantly, how would the processing work? Would the two sessions have to be stacked and processed separately and then layered? Presumably that would have to be the case, because although the Lights could hopefully be stacked together, I would guess that the Flats couldn't because the two Flat batches would be different as regards the camera rotation being changed, and no doubt the focus of the scope too. In which case, how complicated would it be to layer to processed stacks? It's something I dread contemplating to be honest.

[PhotoGav]

Yes, you just do the same again, the difference being that you go to your saved coordinates from the first session's image. Hopefully all the settings in APT and PHD2 won't need to be fiddled with as all will work from the first session.

Processing wise, it depends whether you have removed the camera from the telescope between sessions. If so, you will need to shoot new flats for the second session. Again I don't know how DSS works, but my workflow for multiple session data is to apply flats and darks to the separate session subs, then put those pre-processed subs into one folder and align and combine them all together. If you haven't removed the camera from the scope between sessions, you can just carry on adding subs to the same single folder. Don't dread it, give it a go as the systems are designed to do exactly this as it is the only way to get a decent enough quantity of subs to do a project justice. Even more so in the summer.

[ianpwilliams]

I did remove the camera from the scope as I took lots of Darks last night by leaving the camera in the shed. So I would have to take some new flats. But I could live with that.

It's cloudy tonight anyway, although this weekend looks possible.

I wonder if it would be worth taking another batch of Darks in the shed during a cloudy evening? I took 20 last night, but presumably the more the better?

Although I'm not clear on temperature differences. I agree that it must be OK for the Darks to be a few degrees out because the temperature can change mid-session anyway, resulting in some early-evening Lights not matching the end-of-session Darks.

Then I wondered if it would be OK to combine Darks of different temperatures in one stack, and at first I thought surely not. But then if someonr alternated between Lights and Darks during a session for example then those Darks would be different temperatures too, so maybe it would be alright?

It seems like it would be worth having more than 20 Darks in my library, albeit different temperatures? Or maybe 20 Darks of the same temperature  is enough?