Leo Triplet with tidal stream

[AndyG]

Working through some older data collected earlier this year (February-March). I've imaged this target a few times before from home, but I wanted to see if imaging from dark skies in New Mexico I could pick up the tidal stream flowing from NGC 3628, and I'm happy to see that I was able to pick it up!

[rjbokleman]

Really nice detail! Yep, darkness helps significantly.

I noticed you binned 2x2 the R:G:B. Is there a reason behind that?

[Chuck]

Very well done Andy and great shot as always. I always look forward to your image posts.

[Paul D]

Absolutely AMAZING!!!!! This has become my new desktop, thanks Andy.. LOL

[AndyG]

Thanks, guys!

I noticed you binned 2x2 the R:G:B. Is there a reason behind that?

Mostly it's out of habit since that's the way I learned how to do it when I started getting into AP. The theory is that we're looking to set the exposure time long enough to be sky-limited. The bandwidth of the RGB subs is about 1/3 the bandwidth of Lum, so the 4x flux per binned pixel cancels that out and the exposures can be the same duration as the Lum subs and still be sky-limited. Even binned 2x the RGB subs are only slightly undersampled (3.45"/px with the refractor); if there is some loss of resolution in the RGB's, our perception of detail in an image is mainly based on luminance and very little on chrominance.

I have seen some discussions recently about omitting the Lum subs and just doing unbinned RGB subs and I'd like to experiment with that.

[Apollo XX]

Hi Andy,

Really beautiful image! I'm just curious, that mag7.1 star HIP55262 - is that nebulosity we're seeing there or do you think it's color from the optics?

[AndyG]

Thanks, Mike!

is that nebulosity we're seeing there or do you think it's color from the optics?

Definitely color from the optics -- a round reflection, plus perhaps some dispersion?

[rjbokleman]

Andy,

Fascinating stuff. You forced me to pull out a calculator to tinker around a bit.

The bandwidth of the RGB subs is about 1/3 the bandwidth of Lum, so the 4x flux per binned pixel cancels that out and the exposures can be the same duration as the Lum subs and still be sky-limited.

I did notice that hence the reason most images I reviewed (including my first one here) generally seem to follow that pattern. So in a typical LRGB set, the total of L frames is equal or slightly greater than the total of R+G+B. In my case here, I set out to shoot a total 90 L frames, but only at 90 seconds. (Got 66 good ones.) I then decided to shoot 30 RGB frames which would total 90 together, but at 120 seconds. In reality, I could have gotten away with 60 second L frames and probably gotten 180 second RGBs. I did bin only 1x1 though.

Even binned 2x the RGB subs are only slightly undersampled (3.45"/px with the refractor); if there is some loss of resolution in the RGB's, our perception of detail in an image is mainly based on luminance and very little on chrominance.

The thing that's a bit confusing here is if you BIN 2x2, you've driven the resolution down in my case from 4656 x 3520 to 2328 x 1760. Yes, you get a 4x4 pixel array now instead of 1x1, but how do you account for stacking and aligning these smaller scale images in PI? Do you just let PI do the scaling? I do get the value in the shorter exposure times at 2x2 regardless, so I could have shot 60 second L frames and 60 second RGB @ 2x2 and probably come out in the same place, provided that PI scales them.

I have seen some discussions recently about omitting the Lum subs and just doing unbinned RGB subs and I'd like to experiment with that.

I thought that the L frames is where finer details in the image were coming from, so eliminating them seems like you'd have to compensate by shooting longer RGB?

[AndyG]

how do you account for stacking and aligning these smaller scale images in PI? Do you just let PI do the scaling?

Right, PI handles it automatically by up-sampling the bin 2 RGB subs to match the bin 1 Lum reference frame (in the StarAlignment process via the BatchPreprocessing script)

I thought that the L frames is where finer details in the image were coming from, so eliminating them seems like you'd have to compensate by shooting longer RGB?

Yes, that's the idea. You would shoot the same total amount of time, but just shoot bin-1 RGB subs, with long enough sub-exposures to be sky-limited. Then, a synthetic Lum is created by combining the RGBs. Like I said, I have not really tried this, but I'm interested in experimenting with it. Here's a link to one of several threads on this topic in the PI forum: RGB vs LRGB

One problem I have with LRGB is that it is always difficult to balance the SNR of the Lum vs the RGB. I tend to get much more Lum data, hence nice low SNR master Lum versus noisier RGB masters. The net result is washed out color in the dimmer areas of the image (since the Lum is much more stretched than the RGB.) The appeal of the RGB-only technique is that the luminance SNR is automatically balanced with the RGB (since it is synthetically derived from the RGB.)

[bluemax]

Andy & Ron, please be sure to discuss those sky-limited equations at the next basic imaging forum You guys are way over my head for sure! Last time I heard of Lums it was a restaurant!
Of course, achieving the fine results you guys get, one has to pay attention those sorts of details.
Guys like me who just slog through using default software settings can get fair to OK results. It's just a matter of how deep you want to get into the hobby. But I really enjoy seeing the varying levels of imagers within the club. i.e. You really can do imaging with simple equipment as some of the members do with DSLR's and wide view Milky Way images. But you can also move to much more complex equipment and go deep sky (or planetary) Once again I am feeling inspired!
Thanks for the great images and discussions here