PPM Centre "jumpy"

[Muddy Diver]

Hi all

I have recently recorded some very "well Tracked" AVI's where Jupiter remains floating almost stationary in the centre of my fov. Having applied my now routine method of Binking then PPM centering,(thanks DP) i was dismayed to see that the resulting frames have misaligned quite badly and Jupiter jumps around badly to the degree that Registax loses the alignment.

PPM centre does a great job on my normal avi's which have suffered the poor tracking of my scope and I cant understand that on an avi which is hardly moving, it can produce such a result. Any others expereinced something similar? is there a fix?

Barry

[davidpretorius]

ok,

bink em and then go to that directory and have a quick look at the images and check to see if the images are mimicking the video you took!

make sure you are using the -outdir=out option so you have the new files in another directory.

are you using -width=400 -height=400?

[Muddy Diver]

Thanks David

OK well, i tried to find my jumpy jupiter then realised that in fact the AVI I had problems with was a scatty saturn! I'm sure thats a sign of something setting in. Actually it was an avi from over a week ago just didnt have time to process everything i shot.

I tried following your lead and altered the width and height to 400. I input 400 into the first width and height fields from the left on the conversion screen. Is taht correct? I couldnt see where to set the -outdir=out. Is this on the conversion screen? I guess that browsing and setting a new directory does the same thing right? Then I converted my 4 AVI's as a sequence to produce 1600 bmp's

The BMP's appeared to mimick the avi in that the position of saturn was very still. even the transition frames (400-401, 800-801 etc) were very similar due to the excellent tracking I had during capture.

Following ppm centre however, everything changed. when scrolling through the centred BMP's the image of saturn jumps erratically and each frame has an image of saturn displaced by 100%.i.e the image is not even overlapping and has moved to a different part of the screen. I retaced my steps for the benefit of this reply and got the same result. Could it be the high gain and therefore noise in the images pershaps?

Anything else I could try? perhaps I should use B&S to produce a complete avi of the four wirth no BMP conversion and ppm centre and just use the misalign and tracking functions in registax.

[davidpretorius]

sorry,

ppmcentre -width=400 -height=400 -outdir=out etc when using ppmcentre

this will crop the bmps and store them in a newly created out directory.

try comparing imagies from the original and the out directories.

[Muddy Diver]

OK David, with you now. The reason I thought you must be referring to B&S is that I didnt know we could change any of the settings in PPM Centre. How is that done? I tried to change directories then typed ppmcentre but got the message below. Is there another way?

By the way, did you get Jupiter in the end?

Barry

[davidpretorius]

try this, hopefully the screen shot shows the process. take into consideration the windows explorer underneath for my directory structure

[bird (Anthony Wesley)]

Hi Muddy, there are lots of settings in ppmcentre that you can change - you should have a read of the ppmcentre home page:

http://www.acquerra.com.au/astro/software/ppmcentre/

that lists all the command line options you can use to change the way it behaves.

The most likely reason for the erratic behaviour is that ppmcentre is having trouble differentiating between the "planet" and the "background", the default setting for ppmcentre sets a threshhold brightness value of 40 for the planet, any pixels dimmer than this are considered to be "background" pixels. If your video has a bright background then ppmcentre can get confused and start counting background pixels as part of the planet, so it gets the wrong idea of where the centre should be.

You should open one of your frames in an image editor and see how bright it thinks the background pixels are, and then run ppmcentre with the "-threshhold" option to set the background brightness a bit higher than this.

As an experiment you could just try running something like:

ppmcentre -threshhold=80 -outdir=test -width=400 -height=400

This will create a directory called "test" and fill it with the cropped images, then see if these are less jumpy. By setting a higher threshhold of 80 it should help ppmcentre differentiate between the planet and the background.

cheers, Bird