Second Mosaic Attempt (same error!)

[AndyUK]

Having bought one of the £6.99 phillips 880's and flashed it, I thought I'd give it a whirl to see how it performed... and when stacking it, I thought I'd give Avistack a go as well for the first time.

I was well pleased with the performance of the 880/900 and I found Avistack was actually better with these avi's than registax, but that's probably just a personal thing or me failing to understand the more twechnical aspects of registax!. ) The only thing that failed (again!) on this 2nd mosaic attempt was me - The drift on the 2000 frames per avi (10fps) was atrocious - I then discovered that I'd set the wrong tracking (DOH!) which meant that when I got the stacked frames, yet again (as before), I missed a couple of sections .

Still, having spent so much time on processing the 29 frames, I thought I'd post this anyway - Aside from the 2 missing sections, has anyone any comments on how I could improve processing this?

[lukebl]

Really great shot. Real shame about the missing bits. The first thing I would do would be to fill the white blocks with black!

[AndyUK]

Cheers Luke - You're very kind... Here's a version where even the white bits are black (and tweaked exposure)

[resonator77]

nice one

how many panes in this mosaic?

[AndyUK]

There's a total of 29 panes... and I used Microsoft ICE to put them together (which I found a LOT easier than the last time when I used imerge!)

Oh yes, and another first for this one was using WxAstro capture - I used Craterlet previously, but (for me) I found that Wx had better shutter speed control. The seeing wasnt that great though...

[resonator77]

pity, you missed some bits

[AndyUK]

Yes... The major hole was actually between panes 1 and 2, but it didn't cross my mind how much effect the drift / wrong tracking mode was going to have on each final pane until I processed the first 2 panes and then found they wouldn't fit together... which was a bit late by then!

Never mind - I'm sure I'll get the chance to have another go fairly soon, and next time I won't forget to use the right tracking mode...

[neilm]

Great photo's nice and sharp.

Now to the questions! Are you just taking the web cam and using it in an afocal mode or are there some other tricks up your sleeve to get the original images?

And the reasoning behind the mosaic, is this because your using Avistack on a small section of each image at a time and then stitching it all back together? I've played with Registax on stills to improve the final image but always done the complete image in one go, never broken it down into smaller frames.

Cheers

Neil

[AndyUK]

Hi Neil - Thanks...! To begin with I removed the plastic focuser from the webcam and replaced it with a nosepiece (to which I've also attached a 1.25" IR block filter). The video's themselves are then taken at prime focus (for me through the diagonal).

My process for taking mosaics (after polar and star alignment, and please note I'm far from being an expert as those two holes denote!) is as follows:

1. Slew to the moon and centre (using a 10mm eyepiece),

2. Attach webcam to PC and open up capture software (check there's signal from the camera by pointing it at laptop screen)

3. Remove the eyepiece from the diagonal and replace it with the webcam via nosepiece - The webcam has a similar FOV to a c. 5mm eyepiece

4. focus and then aadjust exposure/brightness/gain etc on laptop screen using capture software settings until screen presentation looks right [i think I slightly over-exposed the highlights in this attempt]

5. Slew to start point of mosaic and then capture first video (For this one I used 10 frames/second at 1/250s and set max number of frames at 2000 in WxAstro Capture)

6. It's then a case of looking at the screen and moving the scope on each time so as to ensure there's an overlap between the videos [Again, the bit I got wrong!]

I ended up with 29 videos (AVI's) - Although I missed a couple of bits, with a 100mm OTA and an Phillips 900 webcam, if I'd have been more careful/selective (and hadn't been battling with the drift caused by wrong tracking mode!), I could probably have got away with maybe 20 or so, but for me too many is better than too few!

So each of the 29 video's covers a small section of the moon (each video was 1.2Gb). One by one these are then processed in Avistack (or Registax etc). Depending on the seeing, the quality of the video's can vary quite a bit, but I like to try and use half of them if I can (and for these didn't use any below c. 95%). For this image I then saved each processed stack as a PNG file - I didn't realise I could save them as TIFF until later (DOH!). For reference, I've attached one of the 29 tiles that was used for the mosaic - this is as it came out of Avistack so it's not been resized, sharpened... nothing.

When I had all 29 processed. I then opened them all up in Microsoft ICE and let it do all the stitching work, after which I then took it into Photoshop to tweak the exposure, colour (this is a B&W version) and sharpness - I don't know, but I suspect the wavelets feature in Avistack / Registax may be more powerful and next time when I get a FULL set of frames I may play around with some "standard" wavelet processing on each frame BEFORE stitching to see if it comes out any sharper...

Does this answer your query? It's not a case of breaking down an image to process it and then recomposite, simply that the target is too big to fit on the chip so in order to get the whole image you have to take it in chunks and then stitch them all together. And the reason for taking video's as opposed to stills is to get improved image quality (to iron out seeing difficulties).

For info, for my first moon mosaic I used a barlow and ended up having to take a total of 79 tiles to cover the entire lunar surface. I wasn't taking AVI's then though, they were single snap shots using Craterlet, but if it were possible to print it out, it would have been over 4ft squa re(!). When I get a better barlow I want to give that another go...

[Earl]

Great image!

[neilm]

Andy, yes thanks I think I've got the idea, dead easy really! I'm joking, I'm sure it take hours of practice and patience.

I'll summarise my understanding if you've got the time to read it.

Remove the plastic focuser from the web cam.

Focus the scope.

Remove EP and place in web cam that's attach to the nose piece.

Check out the image of the PC,the software allows me to get the focus exposure right not the scope.

Shoot.

Move to next section of moon and reshoot

And so on.

Am I right in assuming that the capture software comes with the web cam, either that or a standard installation of windows has something suitable?

Thanks very much for the info

Cheers

Neil

[AndyUK]

Neil

Hi Neal - It's really not actually that complicated - It's a LOT easier with pictures and/or actually doing it...

Firstly, removing the plastic foscuer from the webcam is a one-off job. With the SPC880 I unscrewed it as far as it went and then pried it off with a screwdriver. This then exposes the chip surrounding which is a thread. A number of people supply the nosepiece adapter but here's the link for FLO's - Adaptors - 1.25" nosepiece to webcam lens thread (ToUcam Pro SPC900NC) - You simply screw it in (tight, but not too tight) and I then fitted an IR block filter. This not only does what it says, but also protects the chip from dust - I therefore leave it all set up... effectively as a "sealed" unit.

With the scope pointing at the target (the moon in this example), remove the eyepiece from the focus tube / diagonal and replace it with the webcam - The nosepiece is a 1.25" fit.

Regarding capture software, there's 3 that I know of that are downloadable free of charge: Craterlet, WxAstro Capture and Sharpcam (I'm sure others will advise us of any others!) - A google search should enable you to find them. For some reason you need to attach the camera to the laptop before running the software (I guess so that it finds it?), but the support info mostly explains what you need to do (again, once you do it the first time, it's not so daunting!), but they all have a "viewing window" which allows you to see what the camera's pointing at.

You use that window to focus the telescope, but you may need to play with the exposure settings within the software to do this - This includes shutter speed, gain, brightness, gamma etc (not all applkications are the same, but trial and error here should quickly point you in the right direction). However, be aware that when you're at this phase if you start of overexposing you'll get a white blob... unless it's way out of focus in which case you may see nothing much at all... and if it's underexposed (and out of focus), again you may see nothing at all. Again, with a bit of trial and error and changing the setting eventually you'll start to see light and with a little more adjustment / fine-tuning the exposure (using the software) and the focus (using the scope) together you'll get the image that you want on the screen with the right exposure and in focus.

At this time, you then set the number of frames, frames/second, duration etc - Not all applications have all of these settings, but again, you'll get the idea pretty quickly. I stick with 10fps as although 15fps might give better physical frames, the images would be compressed. However, if seeing was REALLY bad, then I might go up to 15fps, and if seeing was good (and I thought I could get away with it) I might have a go at 5fps. Regardinmg duration / number of frames, again, suck it and see - somewhere between 1000-2000(?), but the longer the duration, the better the tracking needs to be (again, if using a goto, make sure you use the right tracking mode!)

When that's all done, press capture... and when it's finished, move the scope to video the next pane and press capture again etc etc until you've covered the whole Moon. (I try to make a mental note of specific features/craters in the corners so that I can inlude that feature in the next frame...). And when you've got them all, the processing / stacking can then begin...

Hope that helps(?) but if I've not made myself clear, please shout...!

[neilm]

Andy,

That's exceedingly helpful, going to have to give this a go.

Thanks very much indeed

Cheers

Neil

[AndyUK]

My pleasure... but please don't forget I'm very much a rookie at all this... but there's a LOT of very knowledgeable / nice / kind people on here!

[johnh]

Impressive attempt for a big Moon, missing bits is par for the course when you start. I have never done a big mosaic only small easy ones but your wetting my appetite for a biggy.

John.

[AndyUK]

Cheers John - I must admit, seeing some of the MAJOR close-ups possible with a big scope (like your SPX350), I've always wondered what a full moon mosaic would look like... and also how HUGE it would be!

If you do ever get around to it, PLEASE upload the full image somewhere as I imagine it would like being in orbit (and it would make one helluva print!)

[Steve 1962]

Hi Andy

That's a really nice image. A tip to minimise the risk of missing sections of the mosaic is to use iMerge as you go.

I normally have iMerge running while taking the .avi files and drag & drop the .avi files straight from Windows Explorer into iMerge to keep track of where I've been.

Hope this helps

Steve

[AndyUK]

Hi Steve - I must admit I wasn't aware you could drop AVI's into imerge... I thought of using craterlet and taking a BMP before the AVI and dropping that into imerge, but WXAstroCapture doesn't seem to have a stills capture... Thanks for the tip - I'll try that next time... It'll save that pang of disappointment when I realise I've done it again!

[jgs001]

Nice one Andy, shame about the missing bits... I was gonna offer the iMerge tip... I was surprised to find it took AVI's, I found that one out with by mistake ... but it doesn't half make ensuring no missed bits easier. One other thing, if you orientate the camera, so that it roughly follows the lines you're going to track, you need less panes and it's easier to ensure full coverage.

[AndyUK]

Thanks John - I've learn't quite a few things from this exercise, the biggest of which is imerge taking AVI's. I think you posted this trick in my first attempt, but I didn't realise you meant AVI's - I thought it would only work with the BMP's taken through Craterlet (which although is fine I'm sure, I found WXAC gave me better exposure control). I suppose I could have considered using both Craterlet (for the BMP to imerge check) and then used WXAC for the AVI's but now I know I don't have too!

Another thing I've learnt (as you've mentioned) is getting the panes "level" - I did attempt this (as you can sort of see from the first image posted), but I started in the wrong place (about 7 o'clock on the lunar clock face). If I'd done a dummy run through the centre (9 o'clock to 3 o'cloc, not only could I have got the panes more level, but I would have probably chosen a different standard exposure setting as well... and then of course there's using the correct tracking mode [sIGH]. I'd like to think that the tiny hole at 8pm wouldn't actually have been there if it hadn't been for the drift (but I can't use the drift as an excuse for the enormous hole at 5 o'clock...)

A few queries though - Is 2000 frames per AVI excessive? And if not, should I have be setting my quality limit higher (ie top 500 maybe?). Also, is there any value when stacking in applying a standard wavelet setting to each pane and then merging panes with wavelets applied, or is it best just to stack without wavelets and then sharpen in PS?

[jgs001]

Andy I can't find it at the mo, on an iPod, but I did post my workflow as a sticky somewhere . I use a set of wavelets and apply identical settings to each pane prior to mosaicing. I normally use 1000 frame avi's and use the beat 25%. It's worth taking a look at avistack as it'll run through all the avi's with identical settings.

[Steve 1962]

Andy

Another thing to think about is imaging the terminator first, then imaging the rest of the moon afterwards. This is because the lighting changes quickest at the terminator affecting the light levels and shadows.

I've been tending to set the camera up so that it is "square" to the terminator at the moon's equator and then working from north to south down the terminator, move towards the sunlit limb and go from south to north, towards the sun and north to south again..you get the idea.

If you have an EQ type GEM mount - EQMOD and EQMosaic is pretty good.

I also have to say - read everything John (jgs001) writes - here's a link to his sticky - I've found it enormously useful.

http://stargazerslounge.com/imaging-tips-tricks-techniques/104876-making-lunar-mosaic.html

John - hopefully that's saved you looking it up - I had it as a "favourite"

Regards

Steve

[jgs001]

Cheers Steve, it does, I'd have looked it up later when I can get on a proper outer

[AndyUK]

Cheers both - That's REALLY helpful. I'll try and have another go at this in a couple of weeks time incorporating your tips... Thanks again!

[neil phillips]

Hi nice mosaic, check your frames on reg because of the drift if your ref frame was over the missing section, you can find a frame where the drift shows the missing bit, align on that, then hey presto you actually might have it. ive donE this before. Not sure of course if drift will have the missing section though ?