Voyager mosaics and images of Jupiter, A fresh look at some ancient stuff Options

[Ian R]

This raw movie footage was taken by the Voyager 1 spacecraft as it raced towards Jupiter in February 1979. Clearly visible is the constantly changing attitude of Voyager's scan platform, which houses the narrow angle camera that took this particular sequence.

In total, 3531 frames were aligned to produce this film. The flickering is caused by the Blue filter frames, which are higher in contrast than the Green/Orange counterparts. It's best viewed at 720p (HD):

http://www.youtube.com/watch?v=9nf9nBtd2dM

[machi]

Amazing work! Especially moons are fantastic.
I like when Jupiter is moving up and down. It's in fact imperfection, but it looks like some action documentary.

[tedstryk]

It's in fact imperfection, but it looks like some action documentary.

Yes and no. The way it goes through filter cycles and spacecraft turns really give the viewer the feel of what the spacecraft was doing and of the dataset it was sending in a way that looking at individual frames or looking at a perfected movie focused more on how Jupiter appeared (i.e. color, corrected for spacecraft instability/turns/motion, etc.) never could. Thank you SO much for posting this.

[ElkGroveDan]

That is just smokin cool! It must have been an incredible amount of work.

[nprev]

Second that...whoa!!!

[Ian R]

Thanks for the overwhelmingly positive feedback guys! I have done a colour version, using sequential G/O/B color frames:

http://www.youtube.com/watch?v=CPOrZvIZf08

I'm not entirely happy with it, because I can increase the frame rate by three if I use this sequence instead:

G/O/B
O/B/G
B/G/O

Stay tuned for further developments!

[Ian R]

Here's a merger of all of the frames that feature a moon; Callisto, alas, is sadly absent:

[Ian R]

That is just smokin cool! It must have been an incredible amount of work.

It looks more daunting than it actually is, thanks solely to the ability to record macros in Paint Shop Pro, as seen here:

http://www.youtube.com/watch?v=0XjW0vZZZXw

Most of the frames for the B&W movie were rotated, resized and rendered in a similar fashion while I was busy doing the washing-up!

[Bjorn Jonsson]

Here is what's probably the final version of the Voyager 1 approach movie that has the GRS at the center of the disk:

 jup_8x_30fps.avi ( 1.4MB ) Number of downloads: 974


I posted a very early 16 frame version of this several months a ago but now I'm posting the entire approach movie. The movie is based on 58 orange-green-blue color composites obtained on every Jovian rotation from January 6 to January 29, 1979. Over this period Voyager 1's distance from Jupiter dropped from 58 to 36 million km so the resolution and sharpness of the frames increases from start to finish.

This version has been 'tweened' as described by Ian R earlier in the thread. The parameters are slightly different though; I increased the number of frames by a factor of 8 and then changed the frame rate to 30 fps.

As described earlier when I posted a very early 16 frame version of this movie, the frames are processed to keep Jupiter's size constant. This is accomplished by reprojecting the source images to simple cylindrical projection and then rendering everything using the same viewing geometry. I also sharpened the images a bit to better reveal various details.

Because I didn't have any information on the viewing geometry and camera pointing I had to 'reverse engineer' all of that information. I was unable to do it to perfect accuracy but I still think the resulting geometric parameters are fairly accuarate. Another source of errors is that the images had to be corrected for geometric distortion and that correction probably isn't perfect, especially where Jupiter is close to the edge of a frame (or even partially outside the frame due to bad pointing). These errors manifest themselves as lots of features that appear to 'wobble' a bit and I'm pretty sure that motion is not a real feature. In fact this looks a bit like an earthbased telescopic video of Jupiter except that Jupiter appears very sharp.

Some minor but spurious color variations from frame to frame may be apparent. This is because in many cases the source frames had data dropouts and I had to use partially synthetic color. In some cases Jupiter was partially outside the frame due to pointing errors and in one case the pointing was so bad that none of the three images (OGB) could be used. To keep the apparent frame rate constant everywhere I used Sqirlz Morph to generate a totally synthetic frame from the two adjacent frames where this happened. This worked remarkably well and I think it may be almost impossible to spot where I did this.

An interesting fact I recently realized is that I have been using a large number of programs/software packages when doing this. I don't know if that's unique to my image processing. In addition to at least six programs and command line utilities written by myself I used these and maybe more (list in no particular order):

Photoshop
Paintshop
VideoMach
VirtualDub
AviSynth
Sqirlz Morph
MSU Frame Rate Conversion Filter
Excel

And also some simple programs that come with Windows, e.g. Calculator and Notepad .

Counting everything that's at least 16 programs and command line utilities.

[Floyd]

Bjorn,

The movie is fantastic. Watched it many times watching different parts of the atmosphere swirl and twist. Really outstanding work as always.

[Bjorn Jonsson]

Yes, that's one of the many interesting and fascinating things about Jupiter - something interesting is going on just about everywhere. Belts and zones moving at different speeds, cyclones and anticyclones of various sizes and colors, spots/clouds that enter the Great Red Spot area, circulating currents, and even the polar regions are interesting.

[machi]

Excellent animation!

An interesting fact I recently realized is that I have been using a large number of programs/software packages when doing this. I don't know if that's unique to my image processing.

Definitely, it isn't.
I use normally between 8 to 16 programs + 2 internet programs for downloading data.

[Tom Tamlyn]

>>Because I didn't have any information on the viewing geometry and camera pointing

I'm not familiar with using the PDS archives, but little googling reveals the following comment:


>Traditionally, the pointing information for the Voyager instruments was
>distributed in the form of Supplemental Experiment Data Record (SEDR)
>files. * * * Today, many of the SEDR tapes have been lost. The one major
>exception is Voyager's Imaging Science Subsystem, for which the SEDR files
>were later converted to Vax binary format and are still widely available. However,
>the ISS SEDR files only describe the camera direction and orientation at the time
>when images were taken, leaving large intervals during each flyby when the pointing
> is unknown.

http://pds-rings.seti.org/voyager/spice/ck.html

The discussion goes on to discuss a project for regenerating approximate values for the missing information.


Bjorn, did you need more detailed information than is included in the SEDR files, or were the crucial SEDR files missing?

I'm continually puzzled by the limitations and imperfections of the archiving of Voyager data. The images themselves were archived, but the ancillary information like pointing direction was simply distributed to the initial researchers without anyone taking responsibility for making sure it was preserved? I do understand that that documentation and archiving is a big, picky job, and you can't expect people to do it in their "spare time" (See Commons, the Tragedy of), but it's surprising that it took so long to realize that this function needed to be funded.

TTT

[Bjorn Jonsson]

The main problem is that to the best of my knowledge absolutely no kernels are available for Voyager 1 for the period prior to February 1979. So I not only had to determine the camera pointing from the images but also the spacecraft's position relative to Jupiter.

Another problem, had the spacecraft position been available, is that "C-smithed" kernels are not yet available for the Voyager 1 Jupiter encounter. In these kernels a subset of the images has been used to correct the pointing - without this correction the pointing is far too inaccurate for me to use. What I really need is Jupiter's position in an image with subpixel accuracy (or some equivalent geometric parameters) and this requires very accurate position and pointing information.

And yes, this lack of position/pointing information is frustrating and this is not the first time I express this frustration here. Had it been available the work required for the movie would probably have been reduced by a factor of three.

[Tom Tamlyn]

Are the data in fact missing, or are they packed away in some obscure legacy format that only a few overworked old timers know how to decode?

Or perhaps they exist only as derivatives of other values that need to be calibrated somehow ...? OTOH, the engineering team must have had the ability promptly to extract useable information from the raw transmissions for mission operations.

Unless there's a problem at the physical layer, such as was involved in the interesting LOIRP project to rebuild a 1960s tape drive for reading the Lunar Orbiter tapes, it's hard for me to grok why it's so hard to get at those data.

Of course, I'm not an engineer or a programmer, so my inability to imagine what the problems might be doesn't signify much. I don't doubt that good reasons exist, and I don't assume that the problem is a backlog at the PAO. I'm simply curious to know more about why this is so hard, because I'm interested in the general problem of the fragility of certain kinds of knowledge.

And of course, many thanks for your hours of hard work to give us this beautiful and mesmerizing movie.

TTT