|
ONAG and M83
I was with Dr. Motta when we imaged M83, very low near the horizon for testing purpose.
In each cases (OAG and ONAG) the adaptive optic unit AO-L from SBIG was used.
Exposures were the same, likely stretching was different, but beside there is not post-processing, such as deconvolution as suggested. To be precise those images are not taken exactly at the same time. However they were both very low on the horizon to maximum seeing problems. The light travel has traveled a large chunk of the atmosphere leading to bad seeing anyway.
It is fair to say that there was no seeing monitor to record the exact values, therefore seeing may have been somewhere different for both images, but it was below average in any case.
I like to point out that the images posted in our website are in the JPEG format with medium compression for download speed. Standard JPEG compression creates some inevitable artifacts which could be seen sometime on the background (noisy patterns), also some stars may look like having a darker center, or a strange shape. This has nothing to do with any post-processing.
Attached a zoomed section (bit map) of the M83 ONAG XT image (as downloaded), you can clearly see that stars does not exhibit dark centers.
Guiding with NIR decreases seeing effects on the guider camera but not on the imager camera. However the auto-guiding algorithm (depending of the settings used) may start "chasing" the seeing. Since the isoplanatic patch is quite small, it is very likely that guide star seeing is not correlated with the target image (much wider FOV).
Therefore correcting for the seeing will make matter worse, excepted in a vicinity of the guide star (10" at most). This will be great if you try to resolve close binary stars, but for usual targets it does not work, it makes things worse.
AO does not work outside the isoplastic patch FOV, they are too slow to correct for seeing as well. You need at least a 100Hz correction rate to do so, way faster that you can afford in real situations (units not fast enough, guide stars not bright enough). AO unit could improve the quality of tracking when mounts are not able to deliver the expected performances.
Auto-guiding is always a compromise. In a perfect world we should use long guide star exposures to average out seeing and correct slow mount (an other) drifts. However a lot of mounts exhibit short term ("high" frequency) errors, beside the PE, and require correction every few seconds, or less. At this rate seeing effect on guide star becomes more an issue and NIR guiding will help (even a 23% reduction).
Dr Motta's telescope uses a large fork mount. Although it is truly a beautiful piece of equipment it is hard to expect his mount to behave like high end ones.
In his set-up the OA-L is used as an image stabilizer device (correction rate around 3Hz). AO units have a limited range of correction. When near their limits, it is typical to defer bigger corrections to the mount itself ("bump" the mount), until the AO is re-centered.
In the OAG M83 image we may have experienced some artifacts (trails) from "bumping" the mount, since the AO-L faced a more challenging seeing.
I do understand that our comparison is not a perfect science, there are a lot of parameters to account for when auto-guiding especially with AO unit in the loop. But I do believe it is a fair enough demonstration of the benefice of NIR guiding. It is quite possible touchy differences in the system and software settings may have improved the OAG image as well, but in this case they were not needed, and I think it is a valid point for itself.
|