Thanks to modern computers and software the old, 'official' Voyager Jupiter images can be reprocessed into something much better. There is also a lot of Voyager data there that was never processed into color composites and/or mosaics (or at least it has not appeared on the WWW). With proper processing the apparent image quality approaches the quality of the Cassini images but needless to say the wavelength coverage is (vastly) inferior.

I have recently been taking a close look at the high resolution Voyager 1 images, i.e. the images obtained in early March 1979. This is going to result in some new and/or reprocessed mosaics. The first one is now complete and I'm working on another one.

The image below is a 12 image mosaic (12 orange + 12 violet + 12 synthetic green images). The images were obtained on March 2, 1979 at a range of 4.3 million km. The first image (C1629045.IMQ) was obtained at 05:09:23 and the last one (C1629131.IMQ) at 05:46:11. The resolution is roughly 43 km/pixel.

Click to view attachment

The raw images were calibrated, reprojected to simple cylindrical projection, mosaicked and then rendered using a typical viewing geometry (there is no such thing as a "correct viewing geometry" because the images were obtained over a 37 minute period with Jupiter rotating). I then fixed the color balance. I still haven't 'standardized' how I process the Voyager color. I wasn't completely satisfied with the color I got using an approach similar to what's described in another thread but I think the color could be improved a bit. The final step was to sharpen the resulting image a bit, mainly to compensate for all of the resampling that the previous processing steps required.

This image shows lots of features: The Great Red Spot and one of the three white ovals present during the Voyager flybys, smaller spots, scallopped belt/zone boundaries, gravity waves, a bright equatorial plume and the dusky south polar region.

I don't think I'm bragging by saying that this is probably the best Voyager 1 Jupiter mosaic that I know of, mainly because of its size (12 images).

I will be posting more Jupiter stuff in this thread in the coming days/weeks, both mosaics and interesting images (and needless to say, others are welcome to post images and mosaics as well).

Fantastic pic! Did you use ISIS? If so, what were your processing steps?

You have every right to brag as this is the best Voyager Jupiter product I've seen. The resolution and image quality reminds of New Horizons LORRI images and yet the colors remind of Cassini. With the filter set you had to work with, I think the color looks great.

Jealous mode = ON

Wow! That's stunning!

This is wonderful stuff!

Its so nice to see these images processed with today's technology.

The Images from the 70s and 80s where always over saturated.

Awesome. Voyager images are quite good once you remove the gunk.

Bjorn,

S T U N N I N G

thank you!

Fantastic result Bjorn! I think, that it's biggest published mosaic of Jupiter.

You're throwing meat in the tiger cage.

We want MORE!








If you're looking for ideas for more work; were there any mosaics shot on the departure side from either Voyager? A big enhanced crescent/dark side shot might be pretty dramatic.

Golly. I spent a good 15 minutes just wandering around the infinite complexities of that image.

Are we seeing cloud shadows new the top of the image?

P

Really nice job! This is beautiful!

Absolutely hypnotic. I've looked at it several times over the past two days, and Antipode's right: you see new things each time. Phenomenal, Bjorn.

Taking Bjorn's awesome image above as "reference", I created a blink animation that compares it to a Galileo image of the Great Red Spot almost 20 years later.

The Galileo image had used a synthetic green channel, so I adjusted the colors to match closest to Bjorn's image. I also took some liberties with warping and resizing the Galileo image to match Bjorn's image. I wanted to highlight structural, not color or size differences.


Click to view attachment
[Animated GIF: click to animate]

In the Galileo 1996 image, the clear zone to the S of the GRS is much thinner. The S belt also seems less affected by the GRS. Interestingly, the thin red band to the SE (standing wave?) is present in both images. There is also much more chaos in the region to the NE. In the Voyager 1 image, the NE section shows a smoother transition between zones.

I didn't use ISIS. Voyager calibration isn't available yet in ISIS 3 and ISIS 2 doesn't work properly on my computer. I think I now know why but I didn't when I made the mosaic.


Yes, that's my experience as well. The problem is that this wasn't done carefully enough ~30 years ago so those old, 'official' color composites and mosaics really need to be reprocessed.


I don't think so. You need higher resolution and/or lower solar elevation angles (mainly the former) for cloud shadows/vertical relief to become apparent. Vertical relief (if one can speak of that in the context of gaseous planets) is visible in higher resolution images but it is usually not obvious unless the images are sharpened. Incidentally, the Voyager images are better than the Galileo images if you are interested in cloud shadows and vertical relief. The highest reolution Voyager images are of higher resolution than the Galileo images. In addition, compression artifacts can be an issue if you want to examine small scale features in the Galileo images.

Actually I now think (after taking a careful look at lots of Voyager images) that the Voyager dataset still is in some ways better than the Galileo dataset. While the Galileo images are more carefully targeted and the wavelength coverage is far better (near infrared filters) the Voyager images have no compression artifacts, there is a significant amount of images of higher resolution than anything Galileo obtained and the Voyager images are a better data source for atmospheric movies. So if you are interested in very small scale details in the Jovian atmosphere the Voyager 1 images are what you want.

I'm currently working on another 12 image mosaic that should be finished soon, possibly this week.

Even when they did go through the proper steps, doing all the steps with the 8 bit data rather than converting to 16 bit, something beyond the capability of the vax computers they were using, really degraded the images.

Bjorn, I neglected to congratulate you on an absolutely staggering mosaic -- this is a sure-fire contender for APOD, if ever I saw one!

A new, big mosaic of Voyager 1 images, this time showing the Great Red Spot at high resolution. The first version has normal brightness and contrast and its overall color should be fairly accurate. Small scale features have been sharpened a bit. In the second one the contrast and sharpness have been greatly exaggerated:

Click to view attachment

Click to view attachment

This is a 4x3 mosaic of images. The processing is almost identical to the processing described in the first message of this thread so I'm not decribing it further here. One caveat though: There is probably some slight geometric distortion but it shouldn't be a problem in this case.

The images I used were obtained on March 4,1979 at a distance of about 1.85 million km. The first image (C1635314.IMQ) was obtained at 07:08:36 and the last one (C1635400.IMQ) at 07:45:24. The resolution is roughly 18 km/pixel.

Mosaics of some of these images have appeared before as 'official' image releases but interestingly, only 3x2 images were used in all cases. The offical mosaic can be seen at http://photojournal.jpl.nasa.gov/catalog/PIA00022 . There is a false color version (lots of blue color) that is better known: http://photography.nationalgeographic.com/...t-103317-sw.jpg

My new mosaic reveals an enormous amount of details, especially in the sharpened version. Some of these details I didn't know were visible in any images of Jupiter until relatively recently. The sunlight is coming form the east (right) and because the GRS is in the southern hemisphere it's really coming roughly from the ENE over the GRS and the regions south of it. With this in mind, vertical relief and cloud shadows are clearly visible at many locations around the GRS' periphery. A particularly good example can be seen near (2350,1600) and to the northeast of this location. There is another example near (2480,1080). The wind speeds are especially high there as the elongated appearance of these clouds may suggest. More examples can be seen near (600,980) and (370,920) and (540,1600) and possibly (925,550). These clouds have changed the appearance I imagine Jupiter would have from low altitude (1000 km or something like that). There are also some interesting clouds at (2270,605) and further east.

This is the highest resolution color mosaic completely covering the GRS that I have ever seen. Galileo didn't obtain GRS mosaics at this resolution and Cassini passed too far away from Jupiter. This image looks sufficiently different (and better!) from the old, official versions that in a way I feel like I'm processing stuff from a new planetary encounter when I see this. We will probably not be seeing anything comparable to this until EJSM (or some future spacecraft) starts orbiting Jupiter. Hopefully it will be carrying a camera capable of obtaining even higher resolution images than this from the distances it typically images Jupiter at high resolution.

Interestingly, the orange and violet images I used here were followed by a green filtered GRS mosaic ~40 minutes later. However, I couldn't used these instead of synthetic green because some of the clouds (especially in the GRS' northeast periphery) move so fast that the three color channels couldn't be properly aligned if I used the green images.

Wow.

Oh my.

Goodness O_o.

Whoa! The detail in that is A M A Z I N G ! ! !

Beautiful!

That's outrageously brilliant, Bjorn. That image deserves to be in astronomy textbooks for years to come.

Holy.... holy!

And I thought the Jupiter mosaic from Cassini post-flyby was good! (I cut it out of a calendar and hung it on my wall)

Brilliant work!

Phil

One gets the impression of a "wall of clouds" to the west of the GRS - it appears almost like one is looking down several hundred miles through an upper cloud deck and that the GRS is far below. This may entirely be an optical illusion, but with my screen completely filled with the image, it's hard for me not to see it that way!

Congratulations!

Jonathan

To echo comments above, this is astonishingly good. bravo Bjorn!!

Beautiful quality of detail. Congratulations!
My wife teaches 5th Grade and echoes Stu's textbook idea!

Good gravy! I now officially have Jupiter-envy...

beauty! would you call this close to natural color now?

the voyager cameras were quite good for the time and Ive gotten lots of mileage out of the images. the big difference was the lack of infrared capability, which allows you to look deeper into the atmosphere and see thermal updrafts and hot lavas (on io that is).

The upper one should be fairly close to natural color but this is subjective. It might be a bit too reddish but I'm not sure. After some channel mixing I processed it to make the oval SSE of the GRS roughly white.

A few notes on the mosaics and on images I have noticed in the Voyager dataset:

(1) As should be obvious I have been looking at lots of high-res Voyager images of Jupiter recently. When doing this I have come across several images that show interesting features that I had not noticed before. An example (processed to increase contrast and sharpness):

Click to view attachment
And a wide angle context image shuttered at the same time:
Click to view attachment

The features near the top of the image are interesting. These images where obtained on March 4, 1979 at a range of 1.3 million km from Jupiter's center.

Another example (processed to increase contrast and sharpness):

Click to view attachment
And a context image:
Click to view attachment

The features (clouds?) at upper right in the narrow angle image look somewhat weird. They are located near (435,375) in the context image. The images were obtained on March 5, 1979 at a range of 630,000 km from Jupiter's center.

(2) I noticed an interesting comment in several images in a brown barge imaging sequence. For example in image C1635958.IMQ:

"ATMOS DYNAMICS 3X2 NA GR MOSAIC OF BARGE, STEREOSCOPIC WITH 16359.50 - 60.07"

This seems to have been an attempt to image Jupiter in stereo - I don't think I'm misunderstanding anything here. However, I don't think imaging Jupiter in stereo using a single spacecraft would work due to cloud motions. Two spacecraft are a different story so here's today's crazy idea: Imaging parts of Jupiter in stereo if both NASA and ESA send spacecraft there simultaneously. I'm not sure it would work though - it requires high resolution and I'm not sure ESA's JGO spacecraft is going to be capable of that due to its distance from Jupiter and maybe due to its camera as well.

(3) I haven't been able to confirm that it is indeed vertical relief and/or shadows that is visble in my mosaic. I've been looking for scientific papers and I haven't found any but I haven't searched very carefully. I also did some simple 3D test renders to see which cloud thickness and altitude might reproduce the appearance of the clouds in the mosaic but the results are inconclusive. Still this looks a lot like vertical relief and it should be noted that cloud shadows are visible in high-phase Voyager images of Jupiter. It would be interesting if someone here knew something about this topic. Also see the cloud shadow thread which has a hi-res image showing what might be shadows or vertical relief.

(4) I'm in the early stages of starting work on yet another mosaic and it's going to be a bit different from the ones I have posted here because OGV is available, not just OV. Actually there is a huge amount of Voyager stuff I'd like to process but this takes a lot of time if high quality results are desired. In particular, the lack of accurate camera pointing angles is a constant source of frustration and requires me to reverse engineer this information for every image. Fortunately, the spacecraft location is available and that data seems accurate - without it this would be even more time consuming.

So true. When I was at the New Horizons team meeting in December, I was asked if I used the camera pointing information for my work. I had to be honest - only in situations where I could not ascertain the pointing by studying the image. When it could be done, determining the pointing from the image was significantly more accurate (though much more time consuming).

i spend far more time accurately controlling all the images than i do anything else!

I found the pointing information for the Neptune images to be limited.

yeah, all the published data are predicts based on the original pointing commands. Ive had to update all the pointing
vectors iteratively for all the satellites to get the images to properly mosaic. but then, thats my job!

OK, so you do have to correct the data somewhat. That makes sense.

Paul, I hope you realize how many people reading this forum wish "iteratively updating Voyager 2 spacecraft pointing vectors" was a part of their job description

Substitute Cassini for Voyager 2, and that's half my job...

Perhaps the most important question here is, will those corrected numbers get into the public record somewhere?

Phil

all my work, especially the stereo stuff requires precise pointing, so every cassini release is spent 2 weeks or more calibrating and geometrically registering and revising the pointing vectors for all the new satellite images (altho the past year has been rather a dry spell!). what fun. but then i get to make mosaics like these.
release of the pointing info is a bit more complicated. my Galileo pointing is now public in the form of the Atlas, at least in labeled image map format. beyond that Im not sure yet . . .

This is great, I have always found the lack of accurate pointing info a bit strange. It means that everyone working on images with inaccuarate (or missing) pointing info must either obtain it from someone else or spend lots of time determining the correct pointing. So I suspect the same thing sometimes gets done by more than one person/team when it seems a lot of work could be saved by doing it only once.

Even the spacecraft position can be a source of problems. Today I wanted Voyager 1's position in late January 1979 but my software responded with a SPICE error so I tried the Jupiter Viewer at the PDS Rings Node but got similar results. I suspect that SPICE kernels containing Voyager 1's position aren't available for dates before 6 February but I need to check this more thorougly. I'm working on a Voyager 1 atmospheric movie by the way.

I'm also fiddling with some of the Voyager 1 movie data, Bjorn -- it goes without saying that I'm sure your work will be superior to mine, but here's what I've done so far, in any case:

Jupiter Voyager 1 Approach Movie - Green Filter Only:
http://www.youtube.com/watch?v=4jBir5hix6g

Jupiter Voyager 1 Approach Movie - Color / Rotated:
http://www.youtube.com/watch?v=jZrJ0D8IfB8

Jupiter Voyager 1 Atmospheric Features Movie - Color:
http://www.youtube.com/watch?v=YVMc5PTChwE

I really am a glutton for punishment - at various points during this process, I have resisted the urge to put my fist through the monitor, and throw the tower out of the window!

Only time will tell - at the moment my 'movie' consists of exactly one image...

Paul, I'm really glad to hear you are doing this. To clarify my earlier comment, I was talking about desmearing images and aligning images for stacking purposes. As it stands, the current pointing data that is publicly available isn't adequate.

Ted,
typically, one or two active team members (Cassini is an example) do their own pointing updates as well, sometimes on an as need basis, but sometimes also for cartography. That is how the satellite maps they published are done. I do not know if they do all images or only those that are required for the global maps. I do just about everything because i do a lot of stereo and mapping work (I have a late DPS abstract in on this topic) and require a lot of images. I cant wait for any team releases (i dont know if a mechanism even exists for that) so I do my own! There are a few others out there who do it on as need basis but I may be the only non-team-member doing the whole bunch of it.
cheers
paul

How do you reconstruct it then? The only way I have found is, like I said, clues within the image.

These versions are only 'early drafts', mind you - the final movie should (hopefully) consist of stills that look similar to these:

Click to view attachmentClick to view attachment

Nice, both the images and the videos. I can appreciate how labor intensive a video with that many frames can be, you had to have a *lot* of will power to do that I imagine.

Do you use calibrated or raw data?

IanR:
Wonderfull job! It looks very promisingly.

My movie now consists of exactly two frames that look like this:

Click to view attachmentClick to view attachment

This is an approach movie that has the GRS at a 'constant' location and starts in early January 1979. This looks fairly promising but I don't think I'll finish the movie this year - it's going to be (probably) almost 100 frames. The two frames above really require more careful processing but time is not unlimited so I don't know how much time I'm going to spend processing each frame. As usual the lack of accurate pointing greatly increases the amount of work per frame - in this case probably by a factor of 5 at least.

Jupiter's size will stay constant throughout the movie. This is accomplished by reprojecting the images to simple cylindrical projection and then rendering the resulting maps.