M20 in Ha, SII, Ha, OIII (90, 30, 90, 30 minutes), SII and OIII binned 2x2.
There's a rainbow effect off to the RHS of the nebula which was caused by my main mirror fogging from dew. It was exceptionally cold here last night, I didn't notice that I had condensation on the inside of the RCX until this morning when I went to put all the covers back on. Both the back of the corrector and the mirror had condensation, looks like I'll have to run the fans all night.
Again some nice detail in there Stuart, the star birth spike is visible - a good sign of above average resolution.
I must say though, I don't think the Trifid lends itself well to narrow band especially this dirty green look, I haven't really seen a visually appealing NB shot of the Trifid come to think of it..?
I must say though, I don't think the Trifid lends itself well to narrow band especially this dirty green look, I haven't really seen a visually appealing NB shot of the Trifid come to think of it..?
Mike
I think it needs the outer reflection envelope showing or it kind of looks wrong but of course this is invisible in narrow band.
Very nice resolution & details in there. What was the FL for this shot? Is it a 1:1 crop?
Hi Marc,
The scope is 2000mm FL (10" f/8). This was taken at native focal length, so it's way oversampled, which allows me to use deconvolution effectively. The final picture is shrunk to 75% so I can get it under 200k, it's either that or Jpeg the hell out of it (yuk!).
The scope is 2000mm FL (10" f/8). This was taken at native focal length, so it's way oversampled, which allows me to use deconvolution effectively. The final picture is shrunk to 75% so I can get it under 200k, it's either that or Jpeg the hell out of it (yuk!).
Cheers
Stuart
That'd be great if you could upload a full frame online with a link to it. I think IIS provides some space with an upload facility for this. Then you could put your M16 shot in there as well. THere's so much details in both of them would be cool to have a peek into it.
The scope is 2000mm FL (10" f/8). This was taken at native focal length, so it's way oversampled, which allows me to use deconvolution effectively.
Cheers
Stuart
whew, more up close work, the detail in the dust lanes is impressive, the work you have presented has a very natural unprocessed look about it, which tends to make me think maybee you could push the decon a bit more to bring out even more detail.... i think its there.
Great detailed image. Can't say I am keen on the colour though. It does look a bit odd without the reflection nebula attached but I suppose thats NB for you. Love the detail.
Great detail, but the colour is not for me Stuart. The Trifid is I think an object that just cannot be narrow band imaged. Maybe in the HST pallete perhaps, but not this one. Certainly spectacular detail though.
Nice one Stuart.
M20 can look great in narrowband if correctly weighted. Ha (green) is weighted too high in the presented image or the weights are correct, but the Ha signal is simply too strong for the other channels. Its pretty easy to judge this given its nearly washed out the OIII around the center and there's very little SII signal.
Extract more detail? Perhaps, but doubt it would be through deconvolution unless you're talking about a multi-layered approach. Looking at the stars, if it was pushed any harder through deconv, the image would take on a harsher appearance - not ideal, but a matter of taste. Probably some basic contrast enhancements is all it would need to provide more depth and sharpness where required. IMO, Great data that just needs a little more work.
Great detail, but the colour is not for me Stuart. The Trifid is I think an object that just cannot be narrow band imaged. Maybe in the HST pallete perhaps, but not this one. Certainly spectacular detail though.
Thanks Paul, this is in HST palette...
Quote:
Originally Posted by jase
Nice one Stuart.
M20 can look great in narrowband if correctly weighted. Ha (green) is weighted too high in the presented image or the weights are correct, but the Ha signal is simply too strong for the other channels. Its pretty easy to judge this given its nearly washed out the OIII around the center and there's very little SII signal.
Extract more detail? Perhaps, but doubt it would be through deconvolution unless you're talking about a multi-layered approach. Looking at the stars, if it was pushed any harder through deconv, the image would take on a harsher appearance - not ideal, but a matter of taste. Probably some basic contrast enhancements is all it would need to provide more depth and sharpness where required. IMO, Great data that just needs a little more work.
Cheers
Thanks Jase (and everyone else that's commented), I really need to work on getting the HST colour balance right. I've had this problem with NB images before. This target is dominated by the Ha signal, it's red in RGB images, so I'd expect it to be green in Hubble palette. But Jase is right, there is very little contribution from the OIII and SII data I have in here. Back to work on that one.
In the meantime I've done a synthetic green version of the same thing using only the Ha and OIII data. More visually pleasing.
Stuart I thought I had seen a HSt pallette with and orangy yellow and red on the outside. Oh well, I guess green is not the colour I like. Like I said though, the detail is very good.
It depends on how you mix the channels. The orange ones have a lot more influence from the SII data, I'm not sure how to weight the channels, anyone???
NB weightings are usually specified in an arbitrary manner. However, it helps to have a starting point. For the HST palette (SII:Ha:OIII), I would suggest a 4:1:3 weighting. Of course these weighting will mean nothing if the data set is not normalised between the masters. I'd recommend using pixel math to achieve this.
For those not following the term, normalisation is a process of equalising the background pixel values between subs. Subs that are of varying background brightness make it difficult for data rejection algorithms to detect outlier pixels. Normalisation resolves this by providing uniformity across the data set. While the normalisation of subs is easily performed using image processing software by analysing the differences of a selected reference sub, normalising masters is slightly different as the intention is not for data rejection, but accurate colour weightings (a master is the output of combined subs). Tools like MaximDL have a balance background check box on an RGB combine that perform the background normalisation for you. Its reasonably accurate, but for best results (or for troubled images with gradients for example), pixel math is the answer.
You may find that you'll need to play around with the weightings post normalisation, in particular push SII higher than 4. Trial and error. I've highlighted the process in a narrowband processing thread.
but of course this is invisible in narrow band.
... 
Can't wait to see the RGBilitated version! 
