How to merge data from several nights

[yesyes]

Hi,

quick image processing / stacking question.

When you have lots of frames of the same object taken over several nights, how do you stack these to combine the data?

Do you stack all frames individually? Or do you stack each night separately and then stack the resulting TIFFs?

I have already stacked each night separately and I'm now wondering how to proceed. I take it stacking all individual frames again would take a lot longer than just stacking each TIFF...

I'm using DSS for stacking.

[martin_h]

If they are all the same exposure time and the same temperature, just chuck the lot into DSS with darks,flats etc (you need the same temperature to match darks to lights)

[Glider]

If the exposures, temperatures and flats are not the same from night to night can you not put each night in a different Group tab in DSS and still stack them?

[yesyes]

I have darks and flats for each separate session. They are all 60 sec exposures and the same temperature (give or take a couple degrees).

So you would recommend staking them again individually, not just the TIFFs from each session?

I need a faster computer. Stacking last night's frames already took 50 minutes.

[jgs001]

In DSS... add any single frame, doesn't matter what, then look along the bottom. You'll see a two tabs, Master and Group 1. When you add something to Group 1, Group 2 will appear and so on.

Anything in the Master group will be applied to the data in all groups, so you can just leave this blank. Load all of night 1's data, Lights, Flats, Darks etc into Group 1. Night 2's data into group 2 and so on... Just make sure to save the file list. This way your callibration frames (flats for instance) apply only to the correct data set... And your not restricted to only capturing the same exposures each night etc.

[yesyes]

If the exposures, temperatures and flats are not the same from night to night can you not put each night in a different Group tab in DSS and still stack them?

Yes, I could. I was just wondering if it was OK to save some stacking time by only using the TIFFs from each night. But if using the individual frames is the way to go, then I'll do that and find something to do away from the computer for 2 hours..

[Glider]

So you can have nothing in the Master Group?

What stacking method do you use for lights, darks and flats?

[jgs001]

You can put whatever needs to be used for all the data into the master group... for instance, Bias.. or if you never change anything... flats... if you have one of those cooled cameras with set point, the darks etc...

[ollypenrice]

This came up on another thread. Dennis did a deliberate comparison between combining stacks and restacking from the start. He found no discernible difference. That agrees with my own subjective impression, too. If there was a big difference in the depth of the stacks then you would want to weight them accordingly or noise in the shallow stack will be over represented in the final one, I would have thought.

There are subtleties at a theoretical level but at a practical one I would just stack the stacks, myself. Well, I already do!

Olly

[yesyes]

Ah, ok, I see. Thanks Olly.

So both should work, with stacking the stack probably taking less time.

When stacking the stack, you don't add darks and flats and so on any more as they are already included in the TIFFs, I take it?

[jgs001]

I guess you'd put the tifs in the master group, and the image data with flats and darks etc, from that nights run in group 1

[martin_h]

Ah, ok, I see. Thanks Olly.

So both should work, with stacking the stack probably taking less time.

When stacking the stack, you don't add darks and flats and so on any more as they are already included in the TIFFs, I take it?

Correct, you have already done that the first time round.

[martin_h]

If the exposures, temperatures and flats are not the same from night to night can you not put each night in a different Group tab in DSS and still stack them?

Yes you can, just put the "common" calibration files in Main Group...ie bias and then specific calibration files for the lights in the other groups.

[yesyes]

I guess you'd put the tifs in the master group, and the image data with flats and darks etc, from that nights run in group 1

Ah, that's a combination I had not considered, thanks. Use older stacked TIFFs plus new individual frames...

My CPU is going to be rather busy tonight ... ;-)

[jgs001]

I'd select one of the tifs, right click it and set it as reference frame (probably only the one right ??).. That'll force DSS to align all the other lights onto the TIF.

[martin_h]

I'd select one of the tifs, right click it and set it as reference frame (probably only the one right ??).. That'll force DSS to align all the other lights onto the TIF.

Yep that'll do it

[martin_h]

If you are only stacking a group of ready stacked Tiffs go in to the Registration settings and set it to stack 100% of the frames or it might not stack them all, also find the stack with the lowest star count and use that as the refrence frame.

[yesyes]

If you are only stacking a group of ready stacked Tiffs go in to the Registration settings and set it to stack 100% of the frames or it might not stack them all.....

That must be it.

I tried that last night with 2 TIFFs from 2 separate sessions. It kept complaining that it will only stack one image. Come to think of it, it's set to stack the best 80%. Of 2 pictures that would be 1.6 pictures, rounded down to 1. Doh!

Will try with 100% tonight.

I did, however, run the whole stacking again with frames of each night (light, dark, flat) in a separate group and only bias frames in the main group. That has turned out quite well and has brought out quite a bit more detail than in the 2 separate stacks.

[martin_h]

Sorted then. I always run the whole stack again, but thats just the way I've always done it - and I do it on a dual core PC so it only takes a few mins to do 20/30 frames...considering how long it takes to get them a few mins is nothing!

[Tim]

Ack, just bung 'em all in the melting pot

[yesyes]

OK, I've now tried both. Stacked all individual frames in one run (as mentioned above). And last night I tried stacking just the 2 stacked TIFFs. Setting DSS to "use the best 100% of the frames" it did allow me to stack these 2 TIFFs. Bu the result was far worse than when I stacked the individual frames. A lot of the subtle differences in colour that I had with the individual frames had been lost and I had large areas with the same colour, as if I had used the fill tool.

So I will always stack individual frames, even if that takes a lot longer to stack (50 minutes vs. 2 minutes in my example).

Thanks for the advice, guys. Much appreciated.

[old_eyes]

In DSS... add any single frame, doesn't matter what, then look along the bottom. You'll see a two tabs, Master and Group 1. When you add something to Group 1, Group 2 will appear and so on.

Anything in the Master group will be applied to the data in all groups, so you can just leave this blank. Load all of night 1's data, Lights, Flats, Darks etc into Group 1. Night 2's data into group 2 and so on... Just make sure to save the file list. This way your callibration frames (flats for instance) apply only to the correct data set... And your not restricted to only capturing the same exposures each night etc.

Ooh! that's a great tip. I hadn't spotted that facility.

All I need to do now is to learn how to point the scope at the same piece of sky on two separate nights

old_eyes

[jgs001]

aha... now that is a challenge ... If you process the first nights data, and have the image handy... it helps a lot... any clearly defined pattern you can pick out and framed in the image, you can get close in the setup. Totally accurately doesn't matter, as long as the target doesn't need the entire frame ...

[martin_h]

I place a set of crosshairs over a previous image and use that to line up the new image

Als Reticle (for your PC) - Page 3 - IceInSpace

[jgs001]

Nice one Martin... I use Al's Virtual Lightbox for flats already.