IDAS P2 LPS filter & mono camera - Can you tell a difference?

[swag72]

Here's a test I did a month or so ago as I was deciding whether it was worth putting my Hutech IDAS P2 LPS filter into my imaging train. I am using a mono camera and these are taken with the IDAS in the nosepiece and a Baader luminance filter in the filter wheel.

This is only 2 subs, one with the IDAS and one without, both are 300s exposures.

I would be interested if anyone fancies looking (They are fits files) and seeing if they can see a difference. I am particularly interested in whether there is any degradation in the image with the extra filter in the train.

Perhaps looking at this will answer some people's questions who use mono camera's whether they are worth using.

As FLO are now stocking these, this may be of interest to some. Sorry there's only 1 sub each, but I was battling with the moon at the time!!!

No IDAS used - 300s luminance http://dl.dropbox.com/u/41385123/Lum_300s_noIDAS.fit

IDAS used - 300s luminance http://dl.dropbox.com/u/41385123/Lum_300s_IDAS.fit

[Martin-Devon]

Sara, on single subs it's a bit tricky to really assess whether there is any meaningful/reproducible differences. I've opened both FITS files up in Maxin & flipped between them, but can not discern anything that stands out as a significant difference. To do a proper comparison I think you need to stack 10-20 subs +/- the filter and do this ideally under more demanding conditions such as full moonlight.

Martin

[James]

I think Martin is right - there can be enough variability between two subs that any difference caused by the filter can be lost in the natural difference between the two.

I've found generally that I don't really need to use mine for subs up to about 4-5 minutes where I live (fairly dark village 8-10 miles from Exeter) but I do if the subs are longer. Back in Taunton I noticed the difference after just 2-3 minutes. This was whilst imaging in mono...

James

[swag72]

I have found 2 different comparisons. I knew I had them somewhere!

Here is 5x300s subs - Again one with the IDAS and one without. These have been processed in exactly the same way - Black point taken from exactly the same point in Maxim and DDP'd with the same settings. Perhaps this may be a little more telling, although it's not the 10/20 subs Martin suggested.

5x300s with IDAS - http://db.tt/V8b60GbZ

5x300s luminance filter -http://db.tt/YuUQmzzH

[Gina]

I don't seem to have an app that will read FIT files

[steppenwolf]

Looking at the second set of images as these are a better test as they will be more of an 'average', the comparison is inclusive - some tiny features are better in the IDAS version and some are better in the L version. If I had to hazard a guess at which should give the better image, I go for the IDAS because you are collecting a broadband image with both with IR and UV cut also with both BUT the LP is passing through the straight L filter where it has the potential to 'muddy' the data whereas the IDAS version at least has attenuation at undesirable wavelengths.

A slightly better test would be to choose a good clear night and take a succession of images alternating between the L and IDAS filters throughout the session so that you get an even more 'averaged' set of data when you stack like types of data into the two sets.

[swag72]

Thanks for those thoughts Steve - I think I'll put the IDAS back on after the next image.

[gingergeek]

Gina, FITS liberator is a free download from ESA http://www.spacetelescope.org/projects/fits_liberator/

[opticalpath]

This is an interesting experiment.

Subjectively, I find it quite hard to decide which image is 'better'. To do a more objective comparison, perhaps we should go back to first principles. Our primary aim, technically, is to get good signal to noise ratio, so an objective test would be to measure the S/N of individual stars; especially faint stars which represent the marginal parts of the image that are in danger of being lost in the noise - the most sensitive area to test. If the measured S/N of such stars is consistently higher in one image compared to another, we could call that an objective improvement.

Using something like Maxim DL in aperture mode, we could mouse over a selection of faint stars in each image and note the S/N for each one reported in the Information window, and build up a comparison picture that way.

There are some issues with stacked and processed images, though. DDP processing with default background value setting clips the black point and that will distort things (differently) in each image. That clipping is evident in your stacked images, Sara. In fact any non-linear stretching could distort the findings. I think a better bet would be to measure stars in an UNprocessed stacked image. The raw images should really be fully calibrated if we want to be scientific about this - as you would for photometry- and stacking should use something like median or sigma-reject, in order to eliminate outliers caused by artifacts.

It would be interesting to try this if you have the unprocessed stacked images, Sara.

Adrian

[Gina]

Hmmm - doesn't look like there's an overwhelming case for forking out nearly £150 for yet another filter.

[swag72]

Unfortunately Adrian, I no longer have the original stack. However. I can at some stage put my luminance filter into the carousel and do another comparison with say 10 subs and not stretch them at all and post the fits. I will do that, but I have lots of cloud and have a project on the go. I will do this so that we can all get a better handle on the difference than an IDAS gives over the straight luminance.

[opticalpath]

Intuitively, you'd expect a benefit from the filter in a light-polluted sky; the filter will reduce brightness of stars slightly, but will reduce flux from the polluted sky background even more - which sounds like a recipe for higher S/N if sky background is the predominant noise source.

Adrian