[X] My Assistant

[elakdawalla]

This is going to be awesome.

Attached thumbnail(s)

 

[Ant103]

Work in progress, but I could not resist

Anaglyph Quicktime VR

[EdTruthan]

Work in progress, but I could not resist

Nice VR there! Yes, still waiting on those few AWOL frames to complete the full anaglyph mosaic... but like you, I can't resist. Here's a quick preview of the deck in gloriously detailed full resolution...

[mcaplinger]

still waiting on those few AWOL frames to complete the full anaglyph mosaic...

I think the last missing frames came down last this past Thursday. We're still slowly working on the individual mosaics but I'll be interested to see if anyone builds the complete mosaic and how well it can be fused.

[EdTruthan]

Gotta say Mike, this is a stunning portrait of MSL and just beautifully conceived and captured. Many thanks to everyone involved for the brilliant sequencing. But oh boy, what bear to render. Figuring out the right projection so that no part of the image is too stretched is pretty tricky. And anaglyphing from 2 mosaics that always seem to render slightly differently and aligning them for minimal eye strain even more of a headache. Whew! There's a slew of things one could fidget with for weeks but... ready for prime-time or not, here's my initial version of the full resolution mosaic - anaglyphed (and links to the 2D, Sol 85 version below it). Admittedly, probably the toughest anaglyph I've ever attempted (...and still far from perfect...)

Click the image for a half-sized 9000x7500px version:


...and the Full Resolution anaglyph version at 18000x15000px is here: Sol-84-85-MSL-Self-Portrait-Anaglyph-18000x15000px.zip (32MB Zip File)

Sol 85 in 2D (left channel for 3D)
Half Sized 2D/Sol 85 version at 9000x7500px is HERE (11Mb JPG)
Full Resolution 2D/Sol 85 version at 18000x15000px is HERE (27MB Zip File)

Sol 84 in 2D (right channel for 3D)
Half Sized 2D/Sol 84 version at 9000x7500px is HERE (11Mb JPG)
Full Resolution 2D/Sol 84 version at 18000x15000px is HERE (27MB Zip File)

11/15/12 EDIT: All files have been now been updated and refined...
~ Shutter smear substantially reduced.
~ Horizon flattening refined.
~ Additional anaglyph dirt 'floaters' removed from both channels.
~ Additional anaglyph channel alignments.

11/16/12 EDIT: Added links for Sol 84/2D image to pair with Sol 85 2D for 3D imaging applications.

[mcaplinger]

But oh boy, what bear to render. Figuring out the right projection so that no part of the image is too stretched is pretty tricky. And anaglyphing from 2 mosaics that always seem to render slightly differently and aligning them for minimal eye strain even more of a headache.

Yes, we've been having the same problems. Your version is really nice and definitely shows that you can get reasonable stereo out of the whole thing.

If anybody would find the metadata of use, let me know and I can probably provide it. We're just using hugin for this ourselves at this point and making no use of the metadata.

Also, removing the shutter smear would be nice. I'll see if I can post that as well if nobody beats me to it.

[iMPREPREX]

If anybody would find the metadata of use, let me know and I can probably provide it. We're just using hugin for this ourselves at this point and making no use of the metadata.

Yes - please provide us with the metadata for MAHLI and both MastCams if possible. Thanks.

[mcaplinger]

Yes - please provide us with the metadata for MAHLI and both MastCams if possible.

I'm only talking about the MAHLI metadata for the self-portrait, and it'd be nice if you gave me some indication of how it would help processing. I've already discussed my inability to release metadata in general; it will be included with the PDS archive.

[iMPREPREX]

Well, I apologize - I just wanted the data so I could plug it in to the stitching programs (I am a beginner), and there is a lot on this site to sift through. I apologize for wasting your time. Just wanted to help.

[mcaplinger]

I just wanted the data so I could plug it in to the stitching programs (I am a beginner)...

The metadata for pointing is in the form of the CAVHOR model vectors (see, e.g., http://www.agu.org/pubs/crossref/2004/2003JE002199.shtml ). Hugin at least requires pointing information in the form of roll/pitch/yaw. If anybody already has code to go from one to the other, or knows of mosaicking software that uses different types of pointing information, please let me know.

I'm going to have to jump through some hoops to get permission to post this, and I don't want to bother if nobody can really do anything with the data.