Unusual behaviour processing Oiii lights

@daraobriain Which camera and telescope are you using? Which filters? How long do you expose per sub for Ha? Did you load the Ha and OIII lights at the same time or did you first process the Ha and then the OIII separately?

@wvreeven Hi! I'm using an Altair 26MM, with an APM 100/800 refractor, and Altair narrowband filters. Each sub was for 360 seconds, and they were loaded (with their own flats) and done separately. No problem with the Ha and then huge problems with the Oiii.

@daraobriain Thanks for the info. Would you mind uploading, say, 5 subs of each Ha and OIII plus 5 each of the calibration files for each filter? See the upload instructions at the top right of any forum page. Please create a folder called daraobriain_oiii_slow and put the files there. Let me know when the upload is done and I'll have a look to see if I can find out what's going on. Thanks!

Oh and before I forget: what version of APP are you using?

Okay have opened a file under that name, and uploaded (in order) 5 Ha flats, 5 Ha lights, 5 Oiii Flats, 5 Oiii lights.

I'm using 1.083

Thanks for this!

D

@daraobriain Thanks very much. I'll probably have a look at the data tonight.

That's great. I'm just going to add two things to the mix. The first time I ran a test integration on Ha it gave a very pleasing result; but every time I run the same integration again I get a very blown out picture. I've added those results to the download; what am I doing wrong THERE? and how did I get it right the first time?

oh and when I tried to integrate my Sii lights, the star analysis went fine but I got the same blank result as the Oiii. There is a slim line on the Histogram but it was impossible to stretch into anything useful.

Thanks again. I've really enjoyed APP for the last two years, nit sure what I'm doing wrong here.

D

@daraobriain Dara, the files all are called very similarly (date and time in the filename only). Can you please either rename them or organize them in folders per type? That would considerably ease my investigation. Thanks very much in advance.

@wvreeven I've renamed them, so hopefully it is clearer now.

Nice to see you here @daraobriain would love to see some of your images......

@daraobriain Sorry for the delay, I am quite busy at the moment.

I have taken a look at your data and I am afraid that several things are wrong with the Ha and OIII lights as well as the OIII flats.

First of all, all images have been recorded as "Flat Frame", even the lights. This is no issue per say because APP ignores the IMAGETYP FITS keyword. On top of that, the filter used is not recorded in the FITS header. Again not a problem per say because it is possible to load the images per filter and manually assign a filter to them. It merely is a lot of work.

Regarding the Ha lights, apart from what I wrote just above this, the signal looks very strange to me. It is almost as if the data has been clipped. This is clear from the histogram when I open one of the lights in APP:

Instead of a smooth curve, there are very distinct spikes to the left (where the pixels with little signal are represented) and a more or less smooth (though still too spiky) distribution to the right (where the bright pixels are represented). This means that the faint nebulosity is lost and explains why the nebula seems to float on top of the sky background and why the sky background is so pitch black.

Regarding the OIII lights, apart from what I wrote just above this, there barely is any nebulosity in the frames and even barely any stars. The histogram looks slightly different from the Ha one, but still doesn't look correct:

Again the faint signal, which contains all nebulosity in this case, has been lost and the sky looks way too black.

Regarding the flats, both the Ha and OIII flats look much better when it comes to the faint end of the histogram. Here the one for an Ha flat:

and here for OIII:

However, unfortunately, both don't look well for typical flats either. Both are very clearly clipped to the left. This means that there are lots of black pixels in (probably the corners and edges of) the flats and that means that in those places no flat correction can be performed. It essentially means that the flats are underexposed OR that your optical train has a huge vignetting. Also, the peak of the OIII flats is too far to the left. I see that you shot the Ha flats with gain 6982 but the OIII flats with gain 444 so you may want to up that latter one.

I am afraid that something went very wrong when you were shooting al the lights and flats. I am not familiar with Altair cameras nor with the AltairCapture software (which is mentioned in the FITS headers) so I am afraid I cannot help you further with that. Perhaps you could ask the Altair people for help using their software or at least finding out what the issue was with these images and flats? Sorry not to be able to be of more help.

@daraobriain Mabula, Vincent and I had another look at the data and they are all 8 bit. Please make sure to capture your data in 16 bit. This most likely is the cause of all troubles so give that a try and see if that solves the issue.

@wvreeven Thank you so much for looking at these!

I have been using Altair capture with an OSC for a year or two and I just presumed it would default to 16bit raw with the new camera. I will add that to the ever-growing checklist.

it was my first attempt at mono shooting, but even though the sensor is slightly larger, it’s surprising that there would suddenly be huge vignetting. I will also have to check the software for how it includes flats/lights and filters in the fits header information.

It was also my first time using a Pegasus flat master panel for the flats, but again, I’m scratching my head as to how that would lead to vignetting.

I presume the original problem, that the “star analyse” step was taking so long, was because there was so little information in the Oiii lights?

also, I presume that ideally, an hour of Oiii imaging of the Rosette should have SOME nebulously in the image?

as I mentioned in another post, the Sii images were similar to the Oiii, so something went strange when I swapped out the filters.

I’ll just have to do more experimenting next clear night I get.

thank you so much for looking at these; hopefully the next set will be more successful.

d

Posted by: @daraobriain

I just presumed it would default to 16bit raw with the new camera.

I made exactly the same assumption and only when Mabula asked me to check that, I noticed that it was 8 bit data.

Posted by: @daraobriain

it was my first attempt at mono shooting, but even though the sensor is slightly larger, it’s surprising that there would suddenly be huge vignetting

I think my assessment about vignetting was wrong and this is because the data are 8 bit.

Posted by: @daraobriain

I presume the original problem, that the “star analyse” step was taking so long, was because there was so little information in the Oiii lights?

Correct. The number of stars simply is too small.

Posted by: @daraobriain

also, I presume that ideally, an hour of Oiii imaging of the Rosette should have SOME nebulously in the image?

Correct again. In general the OIII signal in emission nebulae is much weaker than the Ha emission but, given the strength of Ha in your subs, even a 360 sec light should show some nebulosity with OIII.

Posted by: @daraobriain

as I mentioned in another post, the Sii images were similar to the Oiii, so something went strange when I swapped out the filters.

The SII signal in emission nebulae is even weaker than OIII so you'll need a LOT of data to catch that. But again, given the signal strength of Ha in your subs, I would expect SII to show at least some nebulosity with 360 sec exposures. Mind that for both OIII and SII you may need to go up to 600 sec per sub.

Hi, just to confirm, I went to test the camera and software today and yes, it defaults to 8 bit. So, I've redone all the darks and bias in 16 bit, and if I ever get a break in the cloud I'll try some lights again.

just to check though; presuming I actually do the images correctly this time...

when processing I do each filter in turn with corresponding flats, bias for the camera and darks depending on exposure time...

save each final stacked image as Fits file

then I clear everything...

and then I reload the three stacked fits files, register them together (to align) and then go to "combine RGB" and assign colours of each in turn to get my final image.

Is that, broad strokes, the order?

Thanks,

D

@daraobriain That is one way of doing this, yes.

Another way is to load all lights into 1 session and to make sure to assign them to the correct filters. The same for the flats. Darks, dark flats and bias can be reused for all filters since those don't depend on light and therefore the filter has no influence on them. Then go straight to tab 6 and click the integrate button. That way you'll end up with the same 3 integration results and now they'll be aligned w.r.t. each other already.