In the interest of keeping things easy to find I'd like to suggest that if you are doing some reworking with Registax 4, that direct side by side comparisons be posted here.
Stand alone reworked images can be pasted here or in their own thread as you like, but if in a new thread can you post a link to the old image so those that are interested can check out the changes.
This image is a reworked Endymion from the 11/05/06. I have reprocessed both images using the same settings for wavelets, gamma, B and C etc. The only difference is the one on the left was processed using Reg 3 with single point alignment and the one on the right used Reg 4 with 7 alignment points.
I don't know if I stuffed up somewhere along the line but I was definately not expecting the difference in tonal range in the two processings.
I've also been playing with an old Jupiter Avi, but I'm finding it a bit more of a challenge to get improved results.
Yes ving it will be something to take into consideration when processing. You are correct the adjustments were based on the original saved wavelets and gammas from the first time I processed the image back in June.
I've also been playing with an old Jupiter Avi, but I'm finding it a bit more of a challenge to get improved results
I think it's a mistake to assume that ALL avi's will automatically have a better result just by processing it in R4.
In fact, over the last 2 months of beta testing R4, in some cases (of lunar avi's) i've gone back to single alignment points because R4 stuffed up on joins etc.
A lot of my lunar images have been the result of some areas with MAP processing, combined with other areas of single alignment point, or another run through with only 2 or 3 alignment point on particular areas of the frame where it goes out of the field, etc.
MAP processing can/will be beneficial for some planetary images, especially at long focal lengths where the image scale is large, but on smaller image scales it's not going to add any real benefit when you can encompass most of the whole object in a smallish (eg: 128px) alignment box.
I believe the real benefit in MAP processing will come with lunar and solar images, where the object literally fills the 640x480 frame, or in large Saturn or Jupiter images.
Here is an example of R3 SAP and R4 MAPs and as you can see, there is not a lot of difference...if any. The seeing conditions were rather good at the time, and the operator also had his mojo firing on all cylinders, his chi was up and his biorhythms were in sync...
i am thinking that it will work better on saturn than on jupiter... given saturns lovehandles.... i mean rings. more places to align on, especially in bad seeing. i could be and probably am wrong tho
hmmm I didn't think I could see much difference when I processed the images separately, so here's my comparison.
I always save my stacked image before applying wavelets, so I took an R3 SAP image from 27/10/06 and a reprocess of the same avi done with a 5 point MAP in R4 tonight.
I had trouble picking any difference between the stacked only (unsharpened) images... so... I loaded both images into PS CS2 and assembled them into the same image file. Applied an aggressive unsharp mask (500%, 2 pixel radius and 0 threshold).
There is a visible difference. The R4 version is noticably sharper I think ( the USM is a bit over the top!) but the R3 version is softer.
I agree with Mike that I think there is going to be varying degrees of success, based on quality of the image due to seeing and transparency conditions, collimation and thermal equilibrium of the mirror, etc (you know the usual hair pulling things ). But I also think that there is going to be a moderately steep learning curve, particularly in the quantity and arrangement/distribution of alignment points, the use of different sized alignment boxes in a multipoint alignment, and not to mention trying all the little whizz bangin' options buttons for each step of the way.
Looks like a few fun months coming up in preparation for the next planetary apparition. Come on Saturn
Yes, I'm sure you're right Paul. I've had a bit of a play with different size alignment boxes so far and I can tell you the results are disasterous if the alignment boxes are too small when the image is jumping around due to bad seeing!
The biggest difference I have noticed so far between R3 and R4 lunar images is that for R3, to produce the best image, I have had to stack fewer frames from the avi.
With the MAP functionality of R4, I seem to get equally sharp results using 2 to 3 times the number of frames from the avi.
Here is an R3 / R4 comparison using Plato as the subject matter. As expected, when the seeing is very good and you have good data, there doesn’t seem to be too many differences between using a SAP and MAP’s.
I'd agree with that, Dennis.. MAP seems to be best suited to the nights where you have a boundary layer and/or average seeing where you gets "waves" of fluctuation across the image, where some of the image is sharp while other parts are blurry. MAP processing will combine the sharp parts of each frame so that the final result is sharp.
When the seeing is very good and the image is stable across the field already, then MAP processing won't help much if it all.
i had a quick squiz at it last night... definitely easy to use and quick. just a word of warning tho for those wanting to use the MAP function. you need 24 bit colour on your screen. i had 16bit and it kept crashing... took me ages to figure out it was my own settings doing it.
i ran one of my old avis (some crater... pythagorus i tink). it was a really bad avi and v.4 had problems finding good alignment points with the estimate function. it found 15 points but couldnt hang on to them. i manually picked 4-5 points that were fine tho.
did anyone try the function where registax finds points for you? what did you think?
![[1ponders]'s Avatar](../vbiis/customavatars/avatar45_9.gif){kind=avatar}
![[1ponders] is offline](/vbiis/images/statusicon/user_offline.gif)
Registax 4: post your reworked comparisons here.
). How about posting the comparisons here 
). But I also think that there is going to be a moderately steep learning curve, particularly in the quantity and arrangement/distribution of alignment points, the use of different sized alignment boxes in a multipoint alignment, and not to mention trying all the little whizz bangin' options buttons for each step of the way. 
