IFN around M95 & M96?

[gorann]

About a week ago I gathered 3 hours of lum data with my Esprit 100 and ASI1600MMpro (89 x 2 min @ gain  139, offset 50, -20°C) to add to 3 hours of RGB data collected simultaneously with the Esprit 150 and ASI071MC. I posted the result previously:

Now when I stretch the lum data I see something that could be IFN and wonder if anyone else have looked for IFN around these galaxies and if it would be worth while getting more data to bring it out better - I stretched the hell out of the current data so it is in no way a pretty image.

I do not see any IFN in the RGB data collected with the OSC (ASI071) but maybe that is not so surprising since OSC are probably less sensitive and the RGB was 3 hours at f/7 while the lum was 3 hours at f/5.5.

[vlaiv]

This could be a real thing, but I can also be consequence of high alt clouds passing above (or something else).

Here is a few things you can do with data that you already have to determine if it's a real thing or transient phenomena.

I would first do two separate stacks. Take your registered subs and split them into two groups - first half and second half - and stack them to two stacks. Use heavy binning to enhance SNR - you are interested in background and not detail in the image, so something like 4x4 or higher will be good.

Stretch two images the same to bring out background - again, you are not after pretty image - just go crazy on stretch to show features, just make sure you stretch them both the same. If you are doing linear fitting / equalization of frame prior to stacking (and you should) - do it on same reference frame so that levels in both stack are the same.

Blink resulting images and you will see if pattern changes between the two. If it does not - it is probably genuine signal.

Another thing that you can do is take your registered / equalized frames and do silly high binning - like 8x8. And then create animated gif out of them - if there are passing clouds it will show.

[ollypenrice]

I don't know if there's IFN in the area or not. However, I'd be slightly suspicious of those dark regions around the galaxies. While they might be perfectly genuine they are also characterstic of DBE or ABE applications in which the background sky has been sampled too close to the galaxies. Are these stretches performed after ABE/DBE? Obviously if they are not then you can ignore my point.

Olly

 

[gorann]

1 hour ago, vlaiv said:

This could be a real thing, but I can also be consequence of high alt clouds passing above (or something else).

Here is a few things you can do with data that you already have to determine if it's a real thing or transient phenomena.

I would first do two separate stacks. Take your registered subs and split them into two groups - first half and second half - and stack them to two stacks. Use heavy binning to enhance SNR - you are interested in background and not detail in the image, so something like 4x4 or higher will be good.

Stretch two images the same to bring out background - again, you are not after pretty image - just go crazy on stretch to show features, just make sure you stretch them both the same. If you are doing linear fitting / equalization of frame prior to stacking (and you should) - do it on same reference frame so that levels in both stack are the same.

Blink resulting images and you will see if pattern changes between the two. If it does not - it is probably genuine signal.

Another thing that you can do is take your registered / equalized frames and do silly high binning - like 8x8. And then create animated gif out of them - if there are passing clouds it will show.

Thanks Vlaiv!

doing two stacks is a good suggesting and I will try that. Except for the stacking I do everything in PS. How do I do the binning?

[gorann]

1 hour ago, ollypenrice said:

I don't know if there's IFN in the area or not. However, I'd be slightly suspicious of those dark regions around the galaxies. While they might be perfectly genuine they are also characterstic of DBE or ABE applications in which the background sky has been sampled too close to the galaxies. Are these stretches performed after ABE/DBE? Obviously if they are not then you can ignore my point.

Olly

 

I have not done any PI routines like ABE / DBE. However, I had done a run of Gradient Exterminator in PS, but now when I check what the image looks like before the Grad Ext I still have the dark areas around the galaxies and the other dark patches. Actually there is not that much dark around the left galaxy so I expect it may be real. I will do the double stack that Vlaiv suggested to see what that shows.

[vlaiv]

22 minutes ago, gorann said:

Thanks Vlaiv!

doing two stacks is a good suggesting and I will try that. Except for the stacking I do everything in PS. How do I do the binning?

I usually do it in ImageJ - Image/Transform/Bin ... command. You can also do it in PI by using integer resample:

https://pixinsight.com/doc/tools/IntegerResample/IntegerResample.html

Use average method

Or you can post fits files here, and I'll bin them for you.

[ollypenrice]

19 minutes ago, gorann said:

I have not done any PI routines like ABE / DBE. However, I had done a run of Gradient Exterminator in PS, but now when I check what the image looks like before the Grad Ext I still have the dark areas around the galaxies and the other dark patches. Actually there is not that much dark around the left galaxy so I expect it may be real. I will do the double stack that Vlaiv suggested to see what that shows.

That seems promising, then. I'll have another look at my own data on these two and see if I find a hint.

Olly

[gorann]

3 minutes ago, vlaiv said:

I usually do it in ImageJ - Image/Transform/Bin ... command. You can also do it in PI by using integer resample:

https://pixinsight.com/doc/tools/IntegerResample/IntegerResample.html

Use average method

Or you can post fits files here, and I'll bin them for you.

Thanks,

I assume I would have to do the Integer Resample on each sub before alignment and stacking (integration as PI calls it)? If that is the case it would mean doing it separately on 89 subs - a rather tedious job. I am now just integrating the star-aligned subs as they are (in two stacks) and see what I find

[gorann]

6 minutes ago, ollypenrice said:

That seems promising, then. I'll have another look at my own data on these two and see if I find a hint.

Olly

That wold be great. I have searched Astrobin for images of these galaxies showing IFN but found nothing - most seem to keep the sky quite dark.

[vlaiv]

27 minutes ago, gorann said:

Thanks,

I assume I would have to do the Integer Resample on each sub before alignment and stacking (integration as PI calls it)? If that is the case it would mean doing it separately on 89 subs - a rather tedious job. I am now just integrating the star-aligned subs as they are (in two stacks) and see what I find

No, you can just do it after stacking while still in linear stage - just twice - once for each stack

[ollypenrice]

Hmmm... this is a very aggressive stretch and cutback of our luminance data from the ODK14. It's been done to the linear stack with no other processing, so no DBE or gradient removal. You can have a look for the faint galaxies which might some kind of depth comparison with your data. To be honest I can't see anything IFN-like in ours.

If you PM me a regular email address I'll happily Dropbox you the linear TIFF. I don't want to put a link on the open forum because the data is jointly owned with Yves but I'm sure he'd be happy for you to search it for IFN.

Olly

Edit: running the Ps Equalize filter didn't find anything either, other than some defects in our flats and an LP gradient. This is usually a good way to search for faint signal.

 

 

[Knight of Clear Skies]

You may have something but it's very difficult to tell. There is some dust in that area but it shows very little structure in the SFD infra-red dust map. (The easiest way to view this is to download World Wide Telescope, search for M95 and zoom out to get a similar framing, then select the SFD dust map from the Imagery drop down.)

[gorann]

So I have done the double stack and the pattern is the same in both. However, I now start to worry that this could be a result of my habit of not using flats and instead try to keep everything as clean as possible. There are two ring-shaped structures there that looks like something that could be in the image train, but the rest ave much more irregular structures. Still, could it all be dustbunnies etc? Then gentlemen I apologize for taking your time.....

[gorann]

6 minutes ago, Knight of Clear Skies said:

You may have something but it's very difficult to tell. There is some dust in that area but it shows very little structure in the SFD infra-red dust map. (The easiest way to view this is to download World Wide Telescope, search for M95 and zoom out to get a similar framing, then select the SFD dust map from the Imagery drop down.)

Thanks for the link! Seems like Windows only and I am sitting by a Mac....

[ollypenrice]

3 minutes ago, gorann said:

So I have done the double stack and the pattern is the same in both. However, I now start to worry that this could be a result of my habit of not using flats and instead try to keep everything as clean as possible. There are two ring-shaped structures there that looks like something that could be in the image train, but the rest ave much more irregular structures. Still, could it all be dustbunnies etc? Then gentlemen I apologize for taking your time.....

Certainly no need for apologies. You may still have something but even if you haven't it's worth looking carefully at your data and getting other opinions. I believe someone had found pulsar signals before Jocelyn Bell but chose not to believe they were real...

However, I firmly believe in the virtue of flats.

Olly

[Knight of Clear Skies]

12 minutes ago, gorann said:

Thanks for the link! Seems like Windows only and I am sitting by a Mac....

There is a browser client as well that you could give a try. I thought it wasn't working any more, failing to load the imagery, but I've just tried it again and it seems to be OK. I've used it to match up IFN in my images before using the crossfader.

The correspondence between dust density and visible IFN seems to be pretty good. It could also be used as an aid to framing images.

[gorann]

21 minutes ago, ollypenrice said:

Hmmm... this is a very aggressive stretch and cutback of our luminance data from the ODK14. It's been done to the linear stack with no other processing, so no DBE or gradient removal. You can have a look for the faint galaxies which might some kind of depth comparison with your data. To be honest I can't see anything IFN-like in ours.

 

If you PM me a regular email address I'll happily Dropbox you the linear TIFF. I don't want to put a link on the open forum because the data is jointly owned with Yves but I'm sure he'd be happy for you to search it for IFN.

Olly

 

 

Thanks Olly. To me it seems unlikely that I could find any IFN if you could not. I stretched your jpg file and could see nothing (except some dust bunnies - so is this without flats?). But stretching a small jpg file may say nothing. I will PM you.

[gorann]

13 minutes ago, ollypenrice said:

Certainly no need for apologies. You may still have something but even if you don't it's worth looking carefully at your data and getting other opinions. I believe someone had found pulsar signals before Jocelyn Bell but chose not to believe it was real...

However, I firmly believe in the virtue of flats.

Olly

Thanks a lot for the comforting words Olly!

[vlaiv]

17 minutes ago, gorann said:

So I have done the double stack and the pattern is the same in both. However, I now start to worry that this could be a result of my habit of not using flats and instead try to keep everything as clean as possible. There are two ring-shaped structures there that looks like something that could be in the image train, but the rest ave much more irregular structures. Still, could it all be dustbunnies etc? Then gentlemen I apologize for taking your time.....

Could you post linear fit stacks of these two images?

Blinking them shows at least 3 slow moving objects in these images - probably minor planets / asteroids

Stretch is different in the two, btw, so its a bit hard to tell, but I would say effect is there and not transient in nature (so not clouds or such). We just have to figure out why is it there.

[gorann]

6 minutes ago, Knight of Clear Skies said:

There is a browser client as well that you could give a try. I thought it wasn't working any more, failing to load the imagery, but I've just tried it again and it seems to be OK. I've used it to match up IFN in my images before using the crossfader.

The correspondence between dust density visible IFN seems to be pretty good.

Thanks again. I now used it and found M96 but both the galaxies and sky is quite dark. Is there some way of increasing the brightness?

[Knight of Clear Skies]

5 minutes ago, gorann said:

Thanks again. I now used it and found M96 but both the galaxies and sky is quite dark. Is there some way of increasing the brightness?

I guess you are looking at the Digitized Sky Survey imagery, it was shot on film so is not particularly deep. However, it is good enough to match up the framing with your image by looking at the asterisms in the corners. After that you want to switch to the SFD Dust Map (Infrared) in the Imagery dropdown near the bottom left. 

[gorann]

6 minutes ago, vlaiv said:

Could you post linear fit stacks of these two images?

Blinking them shows at least 3 slow moving objects in these images - probably minor planets / asteroids

Stretch is different in the two, btw, so its a bit hard to tell, but I would say effect is there and not transient in nature (so not clouds or such). We just have to figure out why is it there.

Here are the stacks without anything done to them (except calibration with a master dark). Their brightness is a bit differnt - maybe due to sky conditions changing. You will also notice that there is vignetting, which I removed with Gradient Exterminator in the two images I posted. The number in the file names show the time periods.

M95&96 half stack 0203-0331.xisf

M95-96 half stack 0333-0459.xisf

[vlaiv]

2 minutes ago, gorann said:

Here are the stacks without anything done to them (except calibration with a master dark). Their brightness is a bit differnt - maybe due to sky conditions changing. You will also notice that there is vignetting, which I removed with Gradient Exterminator in the two images I posted. The number in the file names show the time periods.

M95&96 half stack 0203-0331.xisf

M95-96 half stack 0333-0459.xisf

Unfortunately, I can't open those as I don't have instance of PI.

But if you calibrated with darks only and no flats, and using tools like Gradient Exterminator - that can explain things in background.

[gorann]

Just now, vlaiv said:

Unfortunately, I can't open those as I don't have instance of PI.

But if you calibrated with darks only and no flats, and using tools like Gradient Exterminator - that can explain things in background.

Well it is not the Gradient Exterminator since I saw the dust patterns before I used it (in the final step after stretching just to get rid of the vignetting). Here are the files as tiff and the master dark which as far as I can see only shows a bit of amp glow if I do an extreme stretch (a 3 min master dark - should ideally been 2 min but I had none in my library so it could have created some of the vignetting). When I get home in an hour or so I could do stacks of uncalibrated images.

M95&96 half stack 0203-0331.tif

M95-96 half stack 0333-0459.tif

Master dark ASI1600 46 frames -20°C Gain139,Off50, 3min.tif

[wimvb]

1 hour ago, gorann said:

Thanks,

I assume I would have to do the Integer Resample on each sub before alignment and stacking (integration as PI calls it)? If that is the case it would mean doing it separately on 89 subs - a rather tedious job. I am now just integrating the star-aligned subs as they are (in two stacks) and see what I find

Quite easy in pixinsight. Create an image container with the 89 subs. Make sure that the target directory is set, and that you DON'T overwrite the images. Then apply the resample process to this container.