Saving Registered F...
 
Share:
Notifications
Clear all

Saving Registered Frames  

  RSS

 Tim
(@tim)
Main Sequence Star Customer
Joined: 3 years ago
Posts: 39
October 2, 2017 16:26  

Hi Mabula - My normal stacking routine consists of first calibrating my lights and saving them. Later I load the calibrated lights then register and integrate each channel. This works well. However, I notice you also have a facility to save registered frames. I tried saving registered frames but when loading them later - they just show ca in the front as if they are only calibrated. If you try to integrate them they will repeat the 3. Analyze Stars and 4. Register again as if they have only been calibrated. 

So, my question is why save registered frames?  Are they any different than the calibrated frames already saved?  


ReplyQuote
(@gnomus)
Main Sequence Star Customer
Joined: 3 years ago
Posts: 46
October 2, 2017 17:03  

Hi Tim

I save registered frames.  My thinking is that it might be best to 'register' frames only once.  So I register all frames of all channels in one operation and then I integrate the already-registered subs.  To avoid a second round of integration, you go onto the Register tab and under 'registration mode' you select 'no registration'.  Once all the other options are set up it is just a case of loading all the registered frames of each channel and pressing integrate.  Do this for each channel and when you do the LRGB (or NB) combination, the stacks are already registered to one another.   

[The other way involves two registrations: register and integrate each channel, then register the channels.  Whether this makes any significant difference or not, I don't know.]  

EDIT:  Incidentally, I think that for 'archiving' I will only keep the calibrated frames - that will allow me to add more (calibrated) frames at a later date (for example, the following year) and just run a new round of registration.  


ReplyQuote
 Tim
(@tim)
Main Sequence Star Customer
Joined: 3 years ago
Posts: 39
October 2, 2017 17:12  

Thanks Steve - That makes sense. Its the "no Registration" that I was missing. Appreciate you responding.


ReplyQuote
(@gnomus)
Main Sequence Star Customer
Joined: 3 years ago
Posts: 46
October 2, 2017 17:17  

No worries, Tim.  


ReplyQuote
 Tim
(@tim)
Main Sequence Star Customer
Joined: 3 years ago
Posts: 39
October 2, 2017 17:53  

Steve, it didn't register twice following your directions. Tnx again. However, it does "analyze stars" again. That may not seem like much but with hundreds of subs - ASI1600mm-c   -  it can take awhile. Going through "analyze stars" twice for each channel, especially at one sitting seems to be not efficient. The other option is after registration untick the subs except for the channel you want to integrate. This would be better - assuming there was a quick way of unticking and ticking subs in groups. Unticking/ticking hundreds of subs one at a time isn't efficient either. I know its easy to do in DSS but if there is a way in APP, I don't know it.  

So, it seems to me the quickest way is to pick a reference frame yourself. Then use it to register each channel and continue with each channel separately through to Integration. This only uses each function once per channel. It may not seem like much but for hundreds of subs it seems more efficient to me.


ReplyQuote
(@gnomus)
Main Sequence Star Customer
Joined: 3 years ago
Posts: 46
October 2, 2017 18:08  

No worries Tim.  We all have our different requirements and our different ways of doing things.  Yes, APP does re-analyse.  I don't know why - I wonder if  it is something to do with picking the best frame for LN and stacking.  (And perhaps I prefer to have APP 'pick' the reference frame, based upon its algorithms rather than just having me eyeball it.)

Like you, I don't see a way of 'multi-un-ticking' (if that is a word).  I did try the single frame unticking approach, but, as you say, with not far short of a hundred subs it was a nuisance, and I knew I would make a mistake.  With the 100s frames that you generate with these CMOS cameras it is probably a non-starter.

I don't find that the 'extra' Analyze Stars takes all that much time.  (In any event, I should probabaly take more breaks than I do.)  But I have come to the view that an extra couple of minutes of pre-processing is insignificant when I have already spent 10-20 hours capturing the data (and will probably spend another 10 or so hours processing it).  However, I am not 'analyzing' hundreds of CMOS frames. 

In any event, I think it's nice to have the 'save registered frames' option.  

 


ReplyQuote
 Tim
(@tim)
Main Sequence Star Customer
Joined: 3 years ago
Posts: 39
October 2, 2017 19:01  

Thanks again Steve for taking the time to respond. In case another CMOS user is reading this, you don't need to manually choose a reference frame. So, for example, when using LRGB do your Lum channel first and let APP chose the reference frame. Then use that reference frame for the other channels.


ReplyQuote
(@mabula-admin)
Quasar Admin
Joined: 3 years ago
Posts: 2409
October 2, 2017 19:06  

Hi Tim and Steve,

Indeed, like Steve indicated, the "no registration" mode is to be used. That prevents an extra data resampling and time needed to (re-)register the frames.

The analyze stars function is needed to get quality scores calculated, besides needing to know star lokations for registration. Normally I integrate with the quality weight setting and for this to work, APP needs to know the sizes and shapes of the stars in all the frames. Maybe I can built-in the option to disable star analysis when no-registration is used, but then you will be limited in ithe ntegration weight settings.

I'll work on the possibility of selecting/deselecting frames in the frame list panel using ctrl and shift buttons probably, that should help in quick selection/deselection of many frames. Maybe there are even smarter ways to do it, I'll look into it.

Kind regards,

Mabula


ReplyQuote
(@mabula-admin)
Quasar Admin
Joined: 3 years ago
Posts: 2409
October 2, 2017 19:10  
Posted by: Tim

Thanks again Steve for taking the time to respond. In case another CMOS user is reading this, you don't need to manually choose a reference frame. So, for example, when using LRGB do your Lum channel first and let APP chose the reference frame. Then use that reference frame for the other channels.

Hi Tim, in this case, you need to be carefull with the integration composition setting. If it's set at "full" (default value) then the resulting integrations will most likely have different image dimensions, so they wouldn't be ready yet for a direct composite using the RGB compositie tool. If you use the composition setting of "reference" then it will work perfectly, but you will lose all data that is outside the field of view of your reference frame.

Kind regards,

Mabula


ReplyQuote
 Tim
(@tim)
Main Sequence Star Customer
Joined: 3 years ago
Posts: 39
October 2, 2017 20:35  

Thanks Mabula, I caught the setting of "reference" in time. I'm sure it must be often missed. Have you thought of making "reference" the default? At a minimum all mono users working on different channels will need it.

I now understand why you need to do "Analyze Stars" again. I think it would solve the issue easier to work on a better method of selecting/deselecting frames. Thanks again for being so responsive.


ReplyQuote
(@mabula-admin)
Quasar Admin
Joined: 3 years ago
Posts: 2409
October 3, 2017 22:15  

Hi Tim,

Excellent, the "reference" setting was default in older APP versions. I adjusted to "full" for all mosaic combinations.. but I can probably add some extra information on this in the tooltip of the composition.

Yes, selecting/deselecting frames should be made more easy, that will help the most.

Mabula


ReplyQuote
Share: