Astrodon vs Baader Ha Filters

[MartinB]

Boobed with Olly's images.  They were using 530mm f/l and different cameras 7nm SX H36 and 3nm Atik 1100.  This changes the FWHM calc to 4.08 for the 7nm and 5.86 for the 3nm.  

[ChrisLX200]

Boobed with Olly's images.  They were using 530mm f/l and different cameras 7nm SX H36 and 3nm Atik 1100.  This changes the FWHM calc to 4.08 for the 7nm and 5.86 for the 3nm.

And the number of stars used for that calculation in each image was what?

ChrisH

[Krzychoo226]

Show me where the 3nm Baader filter is and I might consider doing the test

No problem

Third from the top: http://www.baader-planetarium.de/sektion/s43d/s43d.htm

[davedownsouth]

No problem

Third from the top: http://www.baader-planetarium.de/sektion/s43d/s43d.htm

 3.5nm is close; just need to make it in the 1¼" size.

[Merlin66]

For what it is worth - the bandwidth of the Ha emission line is close to 3A (0.3nm) So any filter with a bandwidth greater than 1nm will always pass all the available Ha light.

[MartinB]

And the number of stars used for that calculation in each image was what?

ChrisH

Taken from the help file

CCD Inspector employs a proprietary algorithm for star filtering and extraction. For each image, it will extract up to a few thousand stars from the entire image, ignoring hot pixels and other non-stellar structures. As part of the analysis, CCD Inspector will throw out stars that are bloomed or saturated, and any stars with too low a signal-to-noise ratio that may yield an inaccurate measurement. It will then pick the median FWHM value, and the median Aspect Ratio value of all the stars remaining in the list. These will be the values displayed next to the image name.

By its nature, the FWHM and Aspect Ratio displayed represent an "average" value for the image. There will be some stars with higher and some with lower FWHM in the image. The same applies to aspect ratio value.  The values chosen are meant to quantify the image for a meaningful comparison between similar images, such as multiple sub-frames of the same field of view.

For images that contain some field curvature at the edges, the resulting measurement will not be skewed by such curvature, as long as the majority of the stars are not on the periphery of the image.

[MartinB]

For what it is worth - the bandwidth of the Ha emission line is close to 3A (0.3nm) So any filter with a bandwidth greater than 1nm will always pass all the available Ha light.

Doesn't the light cone affect that though?  I understood that very fast systems might need a broader band pass

[MartinB]

Crikey, that was interesting.  Decided to do an image stretch on both subs without trying to match precisely because of different background brightness but aiming to match the nebulosity brightness (close).  There was a very noticeable difference in sub quality.  The s/n was so much better with the astrodon sub.  Now it could be that the moon had just popped out to play for the baader or the neighbours flood lights  went off but I think Dave would have known about that!  The difference isn't marginal!

[ChrisLX200]

Taken from the help file

CCD Inspector employs a proprietary algorithm for star filtering and extraction. For each image, it will extract up to a few thousand stars from the entire image, ignoring hot pixels and other non-stellar structures. As part of the analysis, CCD Inspector will throw out stars that are bloomed or saturated, and any stars with too low a signal-to-noise ratio that may yield an inaccurate measurement. It will then pick the median FWHM value, and the median Aspect Ratio value of all the stars remaining in the list. These will be the values displayed next to the image name.

By its nature, the FWHM and Aspect Ratio displayed represent an "average" value for the image. There will be some stars with higher and some with lower FWHM in the image. The same applies to aspect ratio value.  The values chosen are meant to quantify the image for a meaningful comparison between similar images, such as multiple sub-frames of the same field of view.

For images that contain some field curvature at the edges, the resulting measurement will not be skewed by such curvature, as long as the majority of the stars are not on the periphery of the image.

Well the thing is, as you can see by comparing the images, there are (apparently) far less stars on the 3nM image than the 7nM image. If the algorithm only chose the brightest stars then it is not comparing like with like (i.e., the same stars in each image). All you are left with are the brightest (hence largest) in the 3nM image. There are hundreds of small stars in the 7nM image which siimply don't appear in the 3nM image. You would need to choose exactly the same stars making up the sample from each image otherwise the comparison is invalid.

ChrisH

[MartinB]

Hmm, I a lot of stars are selected, not just the brightest.  However, it's easy enough to compare side by side.  I've performed an indentical curve stretch on the 2 images cropped round the same star and enlarged x3.  So here is the same star seen through the 2 different filters.  To my eye the 7mm  star is marginally smaller.  I don't think this has anything to do with the filter though.  I can't really see that a narrow band pass will tighten stars.  I'm pretty impressed with the contrast enhancement though!

[davedownsouth]

Crikey, that was interesting.  Decided to do an image stretch on both subs without trying to match precisely because of different background brightness but aiming to match the nebulosity brightness (close).  There was a very noticeable difference in sub quality.  The s/n was so much better with the astrodon sub.  Now it could be that the moon had just popped out to play for the baader or the neighbours flood lights  went off but I think Dave would have known about that!  The difference isn't marginal!

Hi Martin,

I choose these two images as the were taken straight after one another.  The Astrodon was taken first (15 minute sub), I then messed around for probably 10-15 minutes slewing the scope to Caph, refocusing with the Baader filter, swinging back to the Soul Nebula / doing a quick framing & starting PHD2 up again, then I took the 15 minute sub.  The Moon was pretty high and bright for both images and as far as I could tell, the weather condition were the same for both; maybe 1 degree colder for the Baader shot, but no signs of cloud, although they did turn up ~30 minutes after I took these images.  It is possible a neighbour turned on some lights, but the scope is reasonable shielded where it stands?

[MartinB]

No, I think the differences are down to the filters.  Well thanks Dave, you're just about to cost me a lot of money!  Do you mind if I explain to my wife that it's all your fault 

[ChrisLX200]

Hmm, I a lot of stars are selected, not just the brightest.  However, it's easy enough to compare side by side.  I've performed an indentical curve stretch on the 2 images cropped round the same star and enlarged x3.  So here is the same star seen through the 2 different filters.  To my eye the 7mm  star is marginally smaller.  I don't think this has anything to do with the filter though.  I can't really see that a narrow band pass will tighten stars.  I'm pretty impressed with the contrast enhancement though!

Achieving perfect focus with each filter is also essential for any comparison - and preferably using optics that are unquestionable. I suppose I could run my own test seeing as both those filters are in the wheel...

ChrisH

[MartinB]

That would be good Chris.  I can understand that stars will be dimmer through the 3mm filter and therefore appear smaller but they should still have the same point spread function.  Having said that, dimmer stars are very handy when it comes to NB so it's a plus regardless!

[davedownsouth]

That would be good Chris.  I can understand that stars will be dimmer through the 3mm filter and therefore appear smaller but they should still have the same point spread function.  Having said that, dimmer stars are very handy when it comes to NB so it's a plus regardless!

I would offer to repeat it as well, but the weather forecast is soooo bad, it might take a while .

[Merlin66]

The apparent brightness of the stars when using Ha narrowband filters will depend very much on their Black Body Curve - which is temperature dependent. Red giants should obviously appear brighter than OB stars.

[Stratis]

I imaged with Baader filters for about two weeks, then switched to Astrodon 5nm (not got deep enough pockets for the 3nm Ha just yet )

I noted an immediate increase in SNR, but a decrease in overall signal. With a noisier chip like the KAF8300 this is significant, it means I need to run the cooling at the maximum practical level in order to increase the sub length without adding too much noise and eliminating the gains made by the superior filter. Going from an 8min Baader sub to a 15min Astrodon sub was just night and day, and even on equal-length subs the difference is notable.

Of course 15min subs turned out to be more than my HEQ-5 could manage reliably, so I bought a CEM60 and have had weeks of overcast skies ever since

[ollypenrice]

This week we ran both filters side by side during a gibbous moon on the tandem FSQ106, Cave Nebula. The image is in the DS section now. Over four hours per scope. The 7Nm went in the bin. It simply wasn't worth using at all.

On moonless nights it remains the case that the AD really gives far smaller stars but is not always radically better. When it is better, though, it really is.

Olly

[opticalpath]

For me, the improved signal to noise ratio is at least as important a benefit as the smaller stars; SN ratio is king when imaging faint objects.  Most of us (with notable exceptions .... you know who you are (Olly) ) have to live with far from perfect dark skies and sky-background is by far the biggest source of noise in our images.  A narrow pass-band dramatically reduces the broad-spectrum background whilst transmitting all of the desired Ha signal.  I.e. a big increase in SN ratio with all the attendant benefits when it comes to processing.  And the narrower the pass band (up to a point), the greater the SN benefit.

Adrian