Excellent Pete, nothing like a new toy to keep the interest levels up
You'd need to add all the glass to that 16mm. I normally add up the thickness of all the glass including the glass over the sensor, and divide it by 3 and then add that to the light path. I doubt it would total 6mm though.
I have got the STL11K dialed in and humming away. I got a bunch of V curves done for FocusMax and now I'm doing some test shots on the area around M8. The FOV on the FSQ-106 is nice and big!
That camera sounds like a beast Rick. Did you find a lot of difference in the peak of the v-curve from one filter to the next? I'm just judging mine by eye for now as I don't have a stepper motor on the focuser however than may change.
Quote:
Originally Posted by Tandum
Excellent Pete, nothing like a new toy to keep the interest levels up
You'd need to add all the glass to that 16mm. I normally add up the thickness of all the glass including the glass over the sensor, and divide it by 3 and then add that to the light path. I doubt it would total 6mm though.
Thanks Robin. I looked at the glass thicknesses however these didn't amount to much. Their engineering drawings cover all of the models however there isn't one drawing that gives all the dimansions for my camera so I was having to look at differences between the 5 and 8 position wheel versions and between the wsg and the ws model. In their knowledge base there is a document that lists the back focus dimensions for each model and this turned out to be correct - should have gone there first.
That camera sounds like a beast Rick. Did you find a lot of difference in the peak of the v-curve from one filter to the next? I'm just judging mine by eye for now as I don't have a stepper motor on the focuser however than may change.
Yes, it is a beast, Peter. I have a bit of tilt but I think I know how to fix that.
I run V curves on the Luminance filter and then do a bunch of focus runs on each filter to calculate an average offset for each filter relative to Lum. The maximum offset I calculated was about 50um (600 steps out of 105,000 on the Atlas focuser.)
Managed to put together my first LRGB image today. I decided to give NGC5139 a try last night due to the near full moon and dewy conditions. I captured 12x5min Lum and 4x2min RGB binned 2x2 before it disappeared behind the trees - that was the easy part. Had to take more Darks and Bias frames today and then work out how to align the binned RGB images with the unbinned Lum. Luckily Pixinsight makes this pretty straightforward.
So here it is - not pretty but at least it has some colour and although I need to touch up the colimation the stars aligned!
A good start, Peter. Are you doing any colour calibration yet?
Rick,
Do you mean flats etc for the colour subs or calibrating the monitor?
I have a Spyder to calibrate the lappy monitor however haven't checked it in a few months.
In terms of the RGB subs I did 2x2 binned bias, darks and flats (only one set of binned flats using the L filter instead of each RGB as I was running out of time). I also did 1x1 bias/darks and flats for the Lum.
I think I need more/longer colour subs than what I did for this image in future.
Sorry, Peter. I meant getting the right ratios of R, G and B when you combine the colour subs. Typically this isn't 1:1:1 because of the characteristics of the filters and the QE curve of the camera. There are lots of ways to do this. I've mostly used calibration from a G2V star in the past.
Cheers,
Rick.
Quote:
Originally Posted by peter_4059
Rick,
Do you mean flats etc for the colour subs or calibrating the monitor?
I have a Spyder to calibrate the lappy monitor however haven't checked it in a few months.
In terms of the RGB subs I did 2x2 binned bias, darks and flats (only one set of binned flats using the L filter instead of each RGB as I was running out of time). I also did 1x1 bias/darks and flats for the Lum.
I think I need more/longer colour subs than what I did for this image in future.
I've just combined them 1:1:1 and eyeballed the colour. Still a very steep learning curve for me. Not even sure if I could find a G2V star at this stage
My (current) colour calibration method in PI for what it's worth:
- This should be done before stretching, ie with the linear data
- LRGB combine with 1:1:1 weights on the channels
- Open STF, turn off the "link RGB channels", and do an autostretch so you can see what you're doing
- Find a patch of sky with relatively few stars and no nebulosity, no matter how small. Make a preview of that patch.
- Make a preview of the whole image. Simple way of doing this is just drag the image identifier tab on the left hand side down. You'll see a "+" appear on the mouse pointer. Let it go and a full sized preview will be created.
- Run BackgroundNeutralization (BN), do the "region of interest" (ROI), "use preview", and select the first preview you created of the background sky. Apply this process.
- The image may look different so do another auto STF. Since the background has been neutralized, the colour balance should be pretty good, so you can link the RGB channels now in the STF
- Run ColorCalibration. In the white ref part, tick the ROI and use the full sized preview you created. Make sure "structure detection" is ticked.
In the background part, again ROI and use the preview you used for BN above. Apply it.
- Done.
My (current) colour calibration method in PI for what it's worth:
- This should be done before stretching, ie with the linear data
- LRGB combine with 1:1:1 weights on the channels
- Open STF, turn off the "link RGB channels", and do an autostretch so you can see what you're doing
- Find a patch of sky with relatively few stars and no nebulosity, no matter how small. Make a preview of that patch.
- Make a preview of the whole image. Simple way of doing this is just drag the image identifier tab on the left hand side down. You'll see a "+" appear on the mouse pointer. Let it go and a full sized preview will be created.
- Run BackgroundNeutralization (BN), do the "region of interest" (ROI), "use preview", and select the first preview you created of the background sky. Apply this process.
- The image may look different so do another auto STF. Since the background has been neutralized, the colour balance should be pretty good, so you can link the RGB channels now in the STF
- Run ColorCalibration. In the white ref part, tick the ROI and use the full sized preview you created. Make sure "structure detection" is ticked.
In the background part, again ROI and use the preview you used for BN above. Apply it.
- Done.
Thanks for the tip Troy. I have tried that once following the on-line tutorial for some QHY8 data however I'm worried if I try it again my brain will explode . Think I need some sleep.
It's easy after doing it once or twice. The trick is the white bal using the whole image. The PI guys are recommending that versus G2V stars. Could be a bit controversial method for some. The results seem to work for me so far.
I like to capture a bunch of R, G and B frames of an unsaturated G2V star and calculate the ratios (you need to adjust for extinction). Then I use appropriate factors from my little spreadsheet when I do the RGB combine. If seems to give a reasonable approximation of the right colours some of the time If that fails then I mess around with PI colour calibration like Troy
It's easy after doing it once or twice. The trick is the white bal using the whole image. The PI guys are recommending that versus G2V stars. Could be a bit controversial method for some. The results seem to work for me so far.
Where did you get that process from, Troy? I've tried something similar but I think it assumes that your average star colour is white and I'm not sure that's true in many FOVs I have seen.
I also tried Harry's (Astroshed) trick of assuming that overall a galaxy is white. It worked OK a few times but did a miserable job on NGC 6744.
I also tried Excalibrator once a long time back but didn't get much joy there. Must try it again now that I'm not quite so ignorant about processing.
Summary: I still haven't found a simple, fool-proof method for getting colours right.
There probably isn't an all-purpose method that will work for every target and every type. Always the way, isn't it.
The method came from reading the PI forum. On the PI YouTube channel there's a 2 part video on colour calibration. That' the bulk of it, but I think they used the galaxy as the WB ref in that one. Just been in random posts from Juan and Vicent regarding using masses of stars. I think their reasoning is that, on average, all stars in a galaxy average out to what we perceive, or choose to be, as white.
The conversation gets very detailed and way above my head.
I've tried the G2V star method, used Excalibrator etc. Trouble is finding a G2V star on the night nearby and losing time imaging that with all the filters and mucking around. Just my opinion, but I'd rather be on targets. Time consuming enough with filters etc.
I don't know about maxim troy, I use it with the sky. There is a txt file there so I guess I converted that to a sky data base file somehow in the dim dark past.
It looks like you use the sky database manager under data, to plug the SDB file into sky and you can turn it on/off with the display explorer under view.
Trouble is finding a G2V star on the night nearby and losing time imaging that with all the filters and mucking around. Just my opinion, but I'd rather be on targets. Time consuming enough with filters etc.
I just do it the once with a new camera and/or filters. In theory it won't change much although extinction factors may vary a bit with atmospheric conditions.
Quote:
Originally Posted by Tandum
I got a g2v data file for the sky here somewhere. You add that data base, turn it on and the g2v stars all light up
Found it but can't remember how to install it.
I have one of those but it was for TheSky 5 and didn't work with TheSkyX when I tried it. I just have a text file with a list in RA order. It's not too hard to find one.
I like to capture a bunch of R, G and B frames of an unsaturated G2V star and calculate the ratios (you need to adjust for extinction). Then I use appropriate factors from my little spreadsheet when I do the RGB combine. If seems to give a reasonable approximation of the right colours some of the time If that fails then I mess around with PI colour calibration like Troy
So Rick, can I clarify? Do you adjust your ratios for each combine based on extinction and the ratio you calculated for your camera/filter/scope combination?
If so, do you work out an "average" altitude for your run of RGB images to calculate the extinction value?
So Rick, can I clarify? Do you adjust your ratios for each combine based on extinction and the ratio you calculated for your camera/filter/scope combination?
If so, do you work out an "average" altitude for your run of RGB images to calculate the extinction value?
DT
Yep, exactly right, David. From memory, I think that AIP4Win is able to adjust for the altitude of individual colour frames but I've never tried to get that sophisticated.
. Think I need some sleep.
If that fails then I mess around with PI colour calibration like Troy 
