M81 & M82 on 10/04/19

[RyanParle]

Ive recently started a little Imaging again after a few years away from the hobby. Back at the end of Feb I spent a few nights re acquainting myself with my kit and getting everything set up and working reliably, i took a few test images of the flame & horsehead nebulae as well as a few sample subs of M81 & M82, since then I have changed my light pollution filter to an IDAS D1 and on Wednesday night I had the opportunity to set up an get a solid night of imaging, I ended up with 24 subs of 15 minutes shot with my EOS 50d (unmodified) on my ED80 refractor, a few subs were discarded due to guiding hiccups, passing aircraft & satellites but most of the data was good and the highest quality i have managed to date. My Processing could probably do with some more practice as there was much more detail within the spiral arms of M81 which I couldn't bring out without messing up the background of the image, but here is my best attempt so far:

I have used the Photoshop Lens Correction tool in manual mode to try and improve the stars on the edges of the field, but couldn't get away with much correction as it was messing up the middle portion of the Image.

Last night I also got an additional 12 subs of 30 minutes I'm currently working on adding this data in to my image, I'm not really sure of the best way do do so, Would it be best to combine all the data in one hit with Maxim DL and process as normal in photoshop or combine the two sets of data separately in Maxim DL then bring them together in Photoshop?

I feel that the data is similar enough in both sets that combining it all as one dataset would work ok, as maxim has floating point image depth it should cope with the dynamic range ok.

Any comments or critique on the above image or advice on how to combine the two data sets would be greatly appreciated.

[alan potts]

Is it possible and I'm sure it is if you know the button to press, to layer mask the Lens correction tool so the inner stars are not affected, I have never tried using it myself?

[RyanParle]

1 minute ago, alan potts said:

Is it possible and I'm sure it is if you know the button to press, to layer mask the Lens correction tool so the inner stars are not affected, I have never tried using it myself?

Thanks Alan, I've got another thread about the lens correction stuff where I am looking in to creating a custom Lens Profile for my APO. Here:

 

 

[Xplode]

Your guiding must be great since you can make so long subs!
I do question the long subs thou, i'd suggest you do 5-10min subs instead of 15-30min. With dithering enable and taking shorter subs you will remove a lot of noise during stacking, no darks needed (and they are in my opinion pointless for not tempcontrolled cameras)

If you get trails from planes and satellites and the sub is otherwise good there is really no need to throw them away, by using some form of sigma clipping during stacking they will be removed automatically.

Do you image without a flattener/focal reducer since the stars are elongated in the corners?

[RyanParle]

28 minutes ago, Xplode said:

Your guiding must be great since you can make so long subs!
I do question the long subs thou, i'd suggest you do 5-10min subs instead of 15-30min. With dithering enable and taking shorter subs you will remove a lot of noise during stacking, no darks needed (and they are in my opinion pointless for not tempcontrolled cameras)

If you get trails from planes and satellites and the sub is otherwise good there is really no need to throw them away, by using some form of sigma clipping during stacking they will be removed automatically.

Do you image without a flattener/focal reducer since the stars are elongated in the corners?

Thanks Ole, My guiding has been pretty good since giving the mount a tune up & belt mod back in Feb. I am typically seeing an RMS of 0.2-0.3 arc seconds and a peak of around 1.5 arc seconds My image scale is 1.76 Arc seconds / pixel so i am happy that the stars stay on the same pixel for the whole exposure.

I may grab a load of shorter subs tonight if i decide to get more data on M81 & M82, I'm not sure that i'll stop using darks though,

I have a pretty extensive library of dark frames containing 60, 120, 240, 300, 600, 900, 1200, 1500 & 1800 seconds, each of those is split in to different temperatures from 5 degrees C up to 25 degrees C in 5 degree increments and each group has a minimum of 25 individual darks, add to this the Bias frames & Flat Frames for every 45 degree of rotator movement bringing the grand total of all calibration frames up to around 1400, of course these have now been replaced by master files to save time & disk space. So I pretty much have a set of calibration frames for any temperature, exposure or rotator angle I may wish to use.

You're right about imaging without a flattener, I never got around to buying one, and probably wont bother for this OTA, I'll be getting a Officina Stellare RH Veloce 200 MK2 AT some time this year, considering that it will be around the same focal length but MUCH faster at f/3 it will make the APO some what obsolete!

[RyanParle]

Ok so I've added in last nights data and got a bit more depth, I also sorted out the an issue I had with the flats not being applied when I calibrated the images, I think I've still got colour balance issues though, the dreaded pinky stars are back!

anyway here it is, 23x 15min subs & 10x 30 min subs:

 

Any comments welcome.

[tooth_dr]

1 hour ago, RyanParle said:

Ok so I've added in last nights data and got a bit more depth, I also sorted out the an issue I had with the flats not being applied when I calibrated the images, I think I've still got colour balance issues though, the dreaded pinky stars are back!

anyway here it is, 23x 15min subs & 10x 30 min subs:

 

Any comments welcome.

Hi Ryan. 

Ok, straight to the point - if that is what I got after 11 hours of data I have to say I think I would be disappointed.  But all is not lost since you clearly have an excellent guiding setup!

As advised, ditch the darks, do a large dither and only use bias and flats.

I have no idea about the scope you have mentioned buying, apart from just looking it up and noticing the enormous price tag.   The ED80 is a fantastic scope and certainly well worth persisting with.  A used 0.85x SW FF for £100 will both correct the field issues and speed up the scope.  If you search the forum, it is full of images taken with this scope, most of which using the field flattener (myself included).

Can you post the raw stacked data TIFF here?

I don’t use maxim but do use photoshop to PP. I would be more than happy to try the data, as I’m sure a lot of others on here would too. I’ve learnt so much from the forum  

 

 

 

 

 

[RyanParle]

12 hours ago, tooth_dr said:

Hi Ryan. 

Ok, straight to the point - if that is what I got after 11 hours of data I have to say I think I would be disappointed.  But all is not lost since you clearly have an excellent guiding setup!

As advised, ditch the darks, do a large dither and only use bias and flats.

I have no idea about the scope you have mentioned buying, apart from just looking it up and noticing the enormous price tag.   The ED80 is a fantastic scope and certainly well worth persisting with.  A used 0.85x SW FF for £100 will both correct the field issues and speed up the scope.  If you search the forum, it is full of images taken with this scope, most of which using the field flattener (myself included).

Can you post the raw stacked data TIFF here?

I don’t use maxim but do use photoshop to PP. I would be more than happy to try the data, as I’m sure a lot of others on here would too. I’ve learnt so much from the forum  

Thanks Tooth_dr 

I would be telling a lie if I said that I was ecstatic with the image considering the amount of data I have, I thought that my local skies bortle 4 rating was mainly the reason I couldn't get more out of it.  I will take the recommendation of larger dither & no darks, I'll try RE processing what I have without the darks to see if that gives better results, I think my dither is set to 2 px at the moment, what sort of dither amount would you recommend for future sessions? I will also try getting a lot more shorter subs on the next session. I'm also beginning to think that my choice of ISO setting could be an issue, is ISO 1600 too much for a 50d?

I'm happy to post my stacked data for you and anyone else to have a go at, I'm sure there is much I can do to improve my processing and would love to improve. I am not able to post the data straight away but I'll add it when I am at the PC.

The Veloce 200 Mkii has long been one of those "if money was no object" wish list items, but recently I was given an offer that was too good to turn down, the final deal is yet to be agreed but I'm expecting to pay around 1/4 of the trade price, it's all totally legit, no stolen or second quality gear. If all goes to plan then I should have it in my hands at the end of the year. The ED80 is a nice scope to use but the Veloce is so much faster & Sharper.

[RyanParle]

After careful analysis of my .fits images and my .tif images I've concluded that i lost a lot of information in the conversion so i'm going to post the .fits files instead, you should be able to get more out of it this way. these .fits files are the combined RGB files, they will need splitting in to individual RGB files before running through FITS Liberator (or similar)

Anyone got any good tips on getting the most out of FITS Liberator?

I've also done an additional stack this time with SD Mask algorithm and no darks. I was pleasantly surprised at the quality of this stack so I'm beginning to believe that darks are less important than I first thought. this is posted here as well.

M81 & M82_Bias_Flat_Dark_Sigma_clip.fit is the original stack

M81 & M82_Bias_Flat_SD_Mask.fit is the latest stack

M81 & M82_Bias_Flat_Dark_Sigma_clip.fit M81 & M82_Bias_Flat_SD_Mask.fit

[Xplode]

It's nice that you've taken the leap of trying without darks
The problem with darks when using a DSLR is that the temperature just isn't stable enough for them to be good enough quality, dark noise doubles with a temperature difference of 5-8C depending on the sensor.
If you stack with +5C lights and +10C darks you're likely to remove too more than  just the noise darks so you're introducing noise. Might be a better idea to use darks from a colder temperature so you will remove some of the noise, but not all.
I can see by stretching your images extremely that there is a "walking noise circle" so i guess you don't dither your images? Dithering will greatly help removing the walking noise and also also the dark noise in general. For dithered images sigma clipping will have to be used when stacking.

[RyanParle]

17 minutes ago, Xplode said:

It's nice that you've taken the leap of trying without darks
The problem with darks when using a DSLR is that the temperature just isn't stable enough for them to be good enough quality, dark noise doubles with a temperature difference of 5-8C depending on the sensor.
If you stack with +5C lights and +10C darks you're likely to remove too more than  just the noise darks so you're introducing noise. Might be a better idea to use darks from a colder temperature so you will remove some of the noise, but not all.
I can see by stretching your images extremely that there is a "walking noise circle" so i guess you don't dither your images? Dithering will greatly help removing the walking noise and also also the dark noise in general. For dithered images sigma clipping will have to be used when stacking.

Yeah, I understand why darks with a DSLR can be tricky due to the unregulated temperature, which is exactly why I built up such a big database in the hope that every image should have a matching dark somewhere (combined with Maxims dark scaling feature)

I do dither when I image, i am currently using the default dither in CCD Commander which is a set to a mimimun of 1 px and a maximum of 4px how much dither would you recommend?

Do you think you could pull out much more details than in my previous posted attempts?

[Xplode]

A dither that small is not very effective on DSLR images, i would suggest 15-30 pixels. A larger dither will also help a lot in fixing small defects that flats doesn't fix 100% because of dust moving around a little and the small difference in size of dustmotes that comes with changing focus that is caused by temperature differences.

I tried PI's dark scaling feature before and found that many well dithered images beat using scaled darks

[RyanParle]

Ok i'll give it a go tonight if it clears up enough... theres a little bit of a haze out at the moment! Fingers Crossed! Should be able to get 90 x 5 min subs easy enough I've set a Minimum dither of 15 px And a max dither range of 100px hopefuly that should smooth out the background a little

[tooth_dr]

21 hours ago, RyanParle said:

Ok i'll give it a go tonight if it clears up enough... theres a little bit of a haze out at the moment! Fingers Crossed! Should be able to get 90 x 5 min subs easy enough I've set a Minimum dither of 15 px And a max dither range of 100px hopefuly that should smooth out the background a little

Hi Ryan

I downloaded both of those files, and opened them both and the integration time is 1219s and 1238s respectively.  Is this a mistake?  Normally when I open my stacked files it has the total integration time (for example 18000s ie 5 hours).  These files look not dissimilar to a single sub of my own data.

[RyanParle]

I've just checked everything through and it appears that it is correct, i'm not sure why MaximDL has saved those integration times to the fits header, but they are definitely the stacks, I've posted a single 15 min sub as well for reference (It has not been de-bayered yet).

The file name is messed up as I had an issue and restarted my CCDC script part way through, this for some reason stopped the "object info" being written to the filename & fits header, i manually updated the Fits header but didnt get around to the file name.

_900s_270.0deg_20190411_0048_000000720.RAW.FIT

[tooth_dr]

15 minutes ago, RyanParle said:

I've just checked everything through and it appears that it is correct, i'm not sure why MaximDL has saved those integration times to the fits header, but they are definitely the stacks, I've posted a single 15 min sub as well for reference (It has not been de-bayered yet).

The file name is messed up as I had an issue and restarted my CCDC script part way through, this for some reason stopped the "object info" being written to the filename & fits header, i manually updated the Fits header but didnt get around to the file name.

_900s_270.0deg_20190411_0048_000000720.RAW.FIT 28.91 MB · 1 download

Thanks Ryan for that.  Although it's B+W with the bayer matrix present, it does look pretty good.  I think lack of dithering and calibration issues are introducing more noise than removing.

[RyanParle]

OK,  Have just run a stack with calibration turned off, and whilst it was a bit messy due to the lack of flats it looked like there was a lot more information above the level of the background noise, so I'm thinking that I have a bad set of calibration masters. I'm currently going through "set calibration" and sorting some settings, some how my default combine when creating master calibration frames had become set to Sigma Clip, and Add pedestal was set to 0. I've now set the combine to average for flats & darks (should i need them in future) and Median for Flats. I've also set a more realistic pedestal value of 3x BG noise level. Maxim is currently creating my Calibration master frames, once it has done I'll try another stack to see if its any better.

[RyanParle]

So after rebuilding my calibration library I haven't noticed much difference to before, but I've taken the time to make a comparison between SD Mask & Sigma Clip as well as Dark, Flat & Bias vs Flat & Bias

Each of the images have exactly the same screen stretch, I think that the Dark_Bias_Flat_SD_Mask file is the cleanest of them all but maybe slightly lacking in the fainter areas. the background noise in the Bias_Flat_SD_Mask file appears a lot less even, I guess this is where the dithering will help out. Sigma Clip seems very noisy by comparison to SD Mask, again I assume that dithering would help here.

[tooth_dr]

18 minutes ago, RyanParle said:

So after rebuilding my calibration library I haven't noticed much difference to before, but I've taken the time to make a comparison between SD Mask & Sigma Clip as well as Dark, Flat & Bias vs Flat & Bias

Each of the images have exactly the same screen stretch, I think that the Dark_Bias_Flat_SD_Mask file is the cleanest of them all but maybe slightly lacking in the fainter areas. the background noise in the Bias_Flat_SD_Mask file appears a lot less even, I guess this is where the dithering will help out. Sigma Clip seems very noisy by comparison to SD Mask, again I assume that dithering would help here.

Some difference in the stacking methods. Unfortunately I don’t use Maxim or know what SD means.

Id be very keen to see what difference dithering makes, that would be a nice test. 

[RyanParle]

3 minutes ago, tooth_dr said:

Some difference in the stacking methods. Unfortunately I don’t use Maxim or know what SD means.

Id be very keen to see what difference dithering makes, that would be a nice test. 

From the Maxim Dl Manual:

• Sigma Clip – a compromise between Average and Median. A standard deviation is calculated for each pixel location, and a standard deviation calculated. The pixel with the largest deviation is rejected if it falls outside Sigma Factor times the standard deviation are rejected. The average is then taken of the remaining pixels.

• SD Mask – a custom variation on Sigma Clip contributed by Ray Gralak. This is an iterative algorithm that more effectively removes outlier pixels while preserving "good" pixels. It is most useful when the number of images available is too small for effective use of Sigma Clip.

There are a few settings that go with Sigma Clip & SD Mask, I've not experimented much with them but when I have I've never noticed much difference.

I will of course post some dithering tests when I get the opertunity to get some more data, the weather looks good for next weekend so fingers crossed!

[RyanParle]

After a few months of being busy on any possible clear night I finally had a couple of nights where I could get some 5min dithered subs.

Things went pretty well and I got 4 hours worth of data each night which when stacked looked very clean but lacked in depth compared with the long exposures, but when combined with the previous data actually turned out looking better than expected although this time I have struggled to get the colour balance right, its all a little too yellow, & I cannot seem to get the blues to show, also the background is still a little noisy which I put down to the noise from the older data but I think its a big improvement on before.

I'm still not sure my results are great considering that there is around 18 hours of subs.

Anyway here is the latest image, comments and critique welcome.