A Different Approach

[Polar Bear]

Always one to experiment I am trying a different approach to Lunar imaging.

The image below was processed in the following way.

Original 720P HD movie clip was in AVCHD taken with a Nex-5

This was run in Magix Movie Editpro and the best 7-10 seconds selected.

Next it was converted and saved as an AVI

Fully automated run in Registax with just one reference point on a 150 image stack and auto wavelets.

Quick tweak and sharpen in Gimp.

Whole process took about 3 minutes.

Considering the seeing wasn't great and it is a first attempt at this method I think it is a good starting point.

Does anyone else use a similar method?

[JamesF]

I'm not convinced it makes sense to take the best few seconds of the video stream. My experience is that the image can vary massively over the course of a few seconds. I reckon you'd be better off taking the best frames you can find in the entire video.

James

[long_arms]

Great image!

I have to agree with James though.

Use the whole avi and you can narrow down the number of good frames to stack in registax.

Not sure whether theres a theory behind the 1 align point either?

Dan

[Steve Ward]

You 'might' get away with this approach in perfect seeing but these are very ,very rare , as James has said the difference in quality frame by frame can be enormous , even at 30+ fps.

Much better results will come if you convert the clip into AVI first and then a combination of PIPP , AS!2 and Registax is employed , stacking only the very best 'frames' rather than seconds of film.

[Polar Bear]

I'm not convinced it makes sense to take the best few seconds of the video stream. My experience is that the image can vary massively over the course of a few seconds. I reckon you'd be better off taking the best frames you can find in the entire video.

James

Thanks James, I'm not convinced either hence the experiment to prove whether it is a good method or not.

The Nex-5 is good in that you can take video in full HD to give a big image.

Multiple high MP single shots are out of the question. The shutter is more violent than my 5D !

I will persevere for a while.

[Polar Bear]

Great image!

I have to agree with James though.

Use the whole avi and you can narrow down the number of good frames to stack in registax.

Not sure whether theres a theory behind the 1 align point either?

Dan

You 'might' get away with this approach in perfect seeing but these are very ,very rare , as James has said the difference in quality frame by frame can be enormous , even at 30+ fps.

Much better results will come if you convert the clip into AVI first and then a combination of PIPP , AS!2 and Registax is employed , stacking only the very best 'frames' rather than seconds of film.

Thanks for that, the results I am getting at the moment are no better than my 640x480 webcam shots from a few years ago.

Seeing was really bad last night which doesn't help. I will try the methods suggested.

[JamesF]

Are you just using the camera to capture the raw data, or is it on the back of a telescope? I'm assuming the latter.

James

[JamesF]

Thanks James, I'm not convinced either hence the experiment to prove whether it is a good method or not.

The Nex-5 is good in that you can take video in full HD to give a big image.

Multiple high MP single shots are out of the question. The shutter is more violent than my 5D !

I will persevere for a while.

What effect does the shutter movement have on the image? And what exposure/shutter speed were you using?

James

[michael.h.f.wilkinson]

I agree with the idea that best frames is much better than best few seconds. I also find auto-wavelets are often not as good as manually setting wavelets. I have no fixed recipe, although I tend to work from small scales (wavelet 1) up to higher ones. I generally fiddle and experiment until I like what I see. I think there is more to be got out of your data that way.

[Polar Bear]

What effect does the shutter movement have on the image? And what exposure/shutter speed were you using?

James

Hi James

video with the Nex-5 is AVCHD 1080p @ 60FPS . I bought the camera primarily to take single still images for my Plane / Moon project.

I am currently re-processing the movie files using PIPP and Registax

I was being a little harsh on the shutter on the Nex-5 for still images and the movement was probably down to not using a remote (one just arrived through the letterbox)

Catching planes at 600mph crossing the Moon will prove interesting as I will need a minimum 1/1000th sec to stop (limit) the planes motion (900ft/sec at 613mph) so 1ft movement at 1/1000 sec.

The best time I have found is during daylight hours an hour or so before sunset or in the morning just after sunrise. This ensures the plane is also illuminated.

Typically I have taken a few days off work to get started on this project and its 100% cloud cover and raining in Lancashire...

[Polar Bear]

OK folks, been re-acquainting myself with Registax and my new friend PIPP

I started with 4300 AVI frames that were converted from AVCHD

PIPP chose the best 1200

Registax ended up with 63 frames.

Applied the Wavelets manually.

This is better than the original and shows the advantage of using many frames to start from.

Comments welcome.

[JamesF]

I'd say that looks pretty good

Two or three more panels and you should be able to create a full mosaic image

James

[Polar Bear]

I'd say that looks pretty good

Two or three more panels and you should be able to create a full mosaic image

James

I'm off to Opticstar tomorrow for a 0.5x focal reducer