M16 - 4 hours 45 mins with the Eagle

[g__day]

I'm happy with this series of stacked shots - 20 shots between 10 minutes to 30 minutes. Very minimal processing besides levels and curves - two iterations, increased the saturation, Gaussian Blur on the mid zone and an Unsharp Mask (probably a bit too servere) on the core of the Eagle. Much happier with the original shots - but even this shrunk shot is revealing decent data.

The other thing I like is the stars are nice and round - meaning guiding is decent.

Matt

[Octane]

Matt,

That is amazing guiding. Very impressive.

I can see loads of intricate detail throughout.

What telescope and camera combination are you using?

Regards,
Humayun

[DavidU]

What Humayun said !
Well done

[AlexN]

Looks great for detail... I would have thought with 4hrs 45mins data it would be a fair bit brighter than that?

How did you go about stacking the different duration exposures?

If you just threw them all into DSS together, DSS will have replaced some of the outer extensions of the neb with the values in the shorter subs.. Your best bet for adding multiple exposure sets together is to stack all the sets separately (ie. all 10min subs, all 20min subs etc.) then load them into photoshop and merge them to HDR. that way you will get all the dynamic range from each data set. where as stacking them together will usually result in less data in the final image...

Like I said, Guiding and detail look great, and focus too.. Perhaps next time, just shoot the longest duration subs you can accurately guide for all night long... I'd love to see 4.5hrs worth of 30min subs... That would be insane!

[g__day]

Humayun

Thank you very much! The gear is a C9.25 -> Meade Motor focuser -> Lumicon OAG (focal reducer removed) -> Hutech LPS filter -> Canon DSLR 400D, on a Vixen Atlux on Losmany gear, guided with a Meade DSI II Pro mono -> PHD 1.10.9

Alex - yes I through it all into DSS - will try what you suggested, grouping similar duration light frames then playing with a merge in CS4.

It's the image processing skills I have to acquire now!

I am experimenting with long and short subs. Now guiding is good I can try all different length subs to get the best signal / noise happening.

Matt

[AlexN]

I agree mate.. Your data seems exceptional.. You have everything under control.. Now you just need to spend some time reading up on processing techniques and experimenting in photoshop.....

Your guiding is at a point that is unbelieveable.. those 1 hour exposures you posted of M20 were unreal! I look forward to watching your processing skills match your capture skills... When that happens, Maaaatteee!! you'll be posting some astonishing shots then.. No doubt about it...

Keep at it Matt..

Alex.

[Hagar]

Very nice Mathew, Guiding is magnificent, You have captured some fine detail in the core. A bit of layer masking to keep your stars from blowing out and you can lift the core to a new height.

Nice.

[g__day]

Doug et al,

Huge thanks - I can't wait to find out what layer masking is :)

Matt

[gregbradley]

I echo the comments. I can't believe you got such round stars for 10- 30 minutes subs.

Something is wrong here though - 4 hours 45 minutes gets a whole lot more data than this image is showing. I know you are imaging at F10 (no reducer is that right?) but data is missing.

What ISO did you use?

Perhaps its what Alex has said and the stacking has averaged out some data. I am only familiar with Images Plus which gives quite a few options about how the data is stacked. What stacking process was used? (there are different statistical methods for adding the images together for example - mean (average), median (midpoint) etc etc.

They have different effects on the data. For example median will get rid of artifacts and cosmic ray hits. Sum will give you a bigger number for the processing to play with but leave those artifacts in. I personally use median most of the time. Sigma reject is another good one (it rejects values too far away from the average ie. hopefully artifacts).

Greg.

[Jen]

Wow nice pic Matt :thumbsup:

[Octane]

Matt,

Thanks for the information. I'm really very impressed!

Did you also take darks, flats and flat darks?

Any way I could get my hands on that data to have a poke around with? :)

Regards,
Humayun

[dugnsuz]

Amazing Matt...the rock-steadiness of your guiding has really brought out some great detail. Your persistence in nailing the tracking will surely be the foundation for some great work in the future.

I'll have to trawl through your posts to try and pick up some tips!!!
Inspiring...
Doug

[g__day]

Humayan

PM me your address and I'll happily cut you a disk with my raw data if you want to experiment.

I take flats, darks of different duration and bias shots. I haven't worked out what a flat dark is - that one confused me!

I am still learning the basics of stacking - re-reading the documentation. My process with DSS has been:

1. take one light frame image and group all my darks of similar duration together so I get master darks for 600 secs, 900 secs, 1200 secs, 1800, 2400 secs, 3600 secs - stored as TIF files.

2. group all my images over 2-3 nights on a target and do a median combine with master darks (for the same durations), master flat and master bias.

3. on the processed image (camera is set to Canon 400D) set the saturation to +21%, balance the colours so they all align as closely as possible - in the middle of the histogram, then go to illumination - in the dim zone move the darkness right up to the far right, shift the mid zone so it just starts to rise when it hits the 3 colours left edge, and change the bright zone to flatten the curve some.

I then store this picture - which looks raw (a bit pastel colours but quite good detail, some dim regions look like they need smoothing), but I store it as a 16 bit tif (not 32 bit rational)

4. Import it into CS4 (and here's where I probably botch it all). Stretch the data (levels and curves 2-3 times). Try to use select colours to grab an area I want to sharpen or smooth (no processing of layers) nor stacking of shots of different duration in DSS and then combining in CS4 using its HDR function.

Its generally here I get somewhat lost and destroy fine detail!

Finally get ride of any small smudges 2-3 smears from the LPS filter, shrink to 1200 and store as a JPG to publish.

So as I said - ALOT to loearn!

Matt

[Octane]

Hi Matt,

As I mentioned to Dave in the tips and processing sub-forum, think of a flat dark as a dark frame for your flat (light) frame. If you haven't already taken them, and, I know this is heresy, because this all should be done at the same time as the image (lights and/or flats were taken) but, would you be able to take a bunch, say 19 or so dark frames for your flat frames? Essentially, if your flat frames were 1/250th (or whatever your exposure duration was) of a second at ISO-100, put your lens cap on, make sure the viewfinder is covered so no stray light is entering the camera, and take 19 exposures at 1/250th (or whatever) of a second. Essentially, the images will look black. That's what we want.

If you can do the flat darks for me as described above, and, if all your stuff was shot in RAW, I would love to have a go with your data as I can tell you've captured some amazing stuff there.

Unfortunately, I am unfamiliar with DSS so will be of no help there at all.

I'll message you my address.

Regards,
Humayun

[atalas]

Nice shot Matt...nice pillar detail.

[g__day]

Humayan,

Thanks for that explanation. All my shots (flat, light, dark, BIAS) are ISO 800. My flat shots were between 1/20 sec - 1/30 sec thru a tee shirt at dusk. So if I understand it correctly a flat dark is the same duration dark frame (say 1/25th sec). So take 20 or so off these (I don't touch my rig - all the gear is firmly attached). So tonight I'll take these frames and send the lot to you and be very interested to see what you can do with it!

Cheers,

Matthew

[Octane]

Sounds perfect!

Thanks.

Regards,
Humayun

[gregbradley]

IfI you are shooting bias frames which are a really short exposure to show the noise from the readout process then there is no need for a dark for the flat. The dark is the bias frame. Take 3 and median combine them. Take 3 flats so you get about 1/3rd of the saturated exposure of your chip. Median combine them and subtract the master bias. That is your flat. No need to use the bias anymore after that you are done with them.

The dark then is 6 or more exposures of the same length as your light and ideally at the same temperature although adaptive darks in Images Plus reduces that need somewhat.

These are combined using sigma reject. That is your master and it is smart practive to make them a normally used exposure length that you can use over and over say 5 or 10 minutes for a DSLR. Have a master dark for each used common exposure time. Make a little library.

The darks will remain valid for a while but chips degrade over time due to radiation damage from cosmic rays.

So redo them from time to time and in your case as it is not cooled from season to season to best match the temperatures (typical for a CCD is thermal noise doubles every 6C or so).

To learn photoshop try Wodaski's Zone System book. I think it has everything you need just about.

Greg.

[leinad]

Wow!!! Matt, that is a great run of imaging and processn!

I maybe missed a post, but are you piggybck guiding/ side-by-side ?
And, what do you use as your guidescope?
Altered PHD settings or running at default ?
:)

Cool that cam down, noise killa!

[g__day]

Greg,

Thanks for the advice - I have the Zone System - just have to read it past page 43! I'm envious of how all Wodaski's shots start real dark, whereas mine have too much ambient light thanks to living near Sydney.

Also a very helpful explanation of the various types of shots. I reprocessed some of these shots adding dark flats and they look better to me.

Initially I was just thrilled that my one hour shots have round stars - I did it again last night (one hiccup where the guide star was lost (clouds) from four shots - so three good hours and one wasted).

Now that I know I can grab longer exposures I can start to concentrate on doing the optimal duration subs, combining shots over several nights - framing the shots well and the all important signal processing.

I stitched eleven shots together last night - totalling 8 hours of M20 data - and it looked absolutely excellent to me after one round of levels and curves - no filters applied whatsoever. So once I know how to remove gradients, subdue light pollution and mask, layer and filter data correctly I think I will be a happy camper.

It's a fun journey when you finally start to see things going well!

Matthew.

PS

Now that I know how to burn data to DVD - let Humayun confirm this - if anyone else wants the RAW data (about 300 files or 2.6 GB) on a DVD to experiment with - I'm happy to burn a disk and flick it to you. I'm keen to see how an experienced hand can process my data - as I am the first to admit I'm an absolute novice on the image processing side of the fence!