I have a curiosity, here.

It seems that the HiRISE team is still learning how to handle their data products. Y'all have known exactly what formats, etc., your data products would be coming down in for years before you had to start handling them.

I'm just highly curious as to why all of these data handling techniques have waited to be developed until after MRO is in orbit and sending back data. Was there some kind of lack of test data for calibration purposes? Did the calibration data have flaws or lacks that didn't allow you to develop these techniques until you were working with live data?

I know it's not because y'all didn't know that you would be wanting to put together color images, for example, or wanting to release these massive images in a manner that the public could look at them easily. And yet, HiView seems a long way off (still), color images are trickling out at a very slow rate... and all because, it seems, the techniques for doing these things haven't been finished.

Is it a manpower issue? Or did the original budget and schedule account for all of this, but as with all budgets and schedules, things have just run behind...?

Just curious...

-the other Doug

Well, color imaging is quite difficult no matter what, and even more so with a camera like HiRISE. There are several things which can cause problems in the color images specifically, let me list a few of them.

1) The spacecraft is vibrating sometimes when we take a picture. This is known as jitter. Jitter can cause the pixels to be off by just a little, which for a black-and-white, doesn't make alot of a difference, but for a color, it can be a huge difference. Also closely related is the sub-pixel problem, what do you do if the pixels are off by a fraction of a pixel?
2) The cross-calibration between different CCDs to make sure the color is accurate is difficult, in fact, most of the color images released to date have been stretched in such a way that isn't accurate.
3) Few of these steps could be tested before MRO was in space, and in fact, most of them require a surface to image, which means waiting for MRO to arrive at Mars.
4) Once we arrived on Mars, the operations folks, who traditionally are the ones who have wrote the pipelines, are busy with other tasks.

I can very much assure you that the color CCDs are still fully-intact, we have the format of them perfectly well worked out, and in fact, the data is checked quite regularly for future images to identify problems by HiRISE Team members. The color data will be released soon, I think there's a June 8th deadline for release to the PDS, and at that time, we'll release our RDR products, which will include color images, as well as the larger black-and-white swaths. It's only a matter of time.

I should also add that I don't know that much about the color processing, but everything I have wrote here is the processing as I understand it. I could be wrong. The one thing I will promise, is that it is far more complex than your standard digital camera pointing and clicking, due to the multiple CCDs that we have to work with.

If you're referring to the lower left portion of the image, that looks like a genuine lighter-toned unit(s) to me.

The upcoming PDS release will not include standard color products. We aren't building standard color products yet, so we don't have any to release to the PDS. However, the complete set of EDR's will be released. So anyone that wants to try their hand at producing their own color images will be able to download the raw EDR's for each filter and attempt to stitch them together to create their own color products. It's possible that some new non-standard color products may be released as extras, but these will not be any different than what has been released previously.

Okay, I stand corrected. I think I can get away with saying that some day there will be a standard color RDR product, but that day apparently isn't June 8th. Guess I should catch up with the latest news before posting something like that...

If I were you I'd just leave.



Leave the executive fridge keys with Jason.

Doug

How could anyone complain about not having the color images,yet? I can barely handle single image files at any one time. I have to pick and choose what files I want to download. Having color added to the mix would likely mean files at the larger end of what has already been released. Perhaps I have an active imagination, but I keep seeing 10 and 15 Gig sized files in my head! My hard drive cannot store very many of those!

Do not get me wrong. I want to see color images just like everyone else. But I am confident that they will be released with time...perhaps when I gets a computer upgrade with storage measured in 10s or even 100s of terabytes. So I'm not complaining until then.

Besides, I cannot think of many other missions that have released data as timely as this one and MER. We should be thanking these teams.

Oh, please, don't get me wrong. I meant no disrespect for the MRO teams. I was just curious as to why some of the data handling tools weren't ready to go when the craft got to Mars, is all. I figured it might be due to a lot of the things tuvas pointed out, but I wasn't sure, and was simply curious.

After all, guys -- these machines are made and operated by human beings. They do incredible things, but their makers and keepers are only human. And if there are any jobs out there that demand that you can *never* make any mistakes, or misestimations, etc., etc., then I know for sure I'm not qualified... and I suspect that no other human being is, either.



-the other Doug

I bet it's got something to do with budget, or lack thereof.

Actually, it's not so much the budget, as it just takes time to learn and understand the instrument itself (that's not to say that throwing some extra bucks our way wouldn't help, but that's really a secondary issue). I don't know of any cutting edge scientific instrument that started producing fantastic data immediately. There's a lot of engineering that goes on just to understand the instrument itself. It may sound strange given all the design and planning that goes into these things, but it's true. These scientific instruments often have a personallity all of their own, and it takes a little time to get to know them, and understand how to make them perform at their best.

To put it a different way, I've used a lot of different telescopes around the world over the years, some of them multiple times, and I've never once acquired a dataset that didn't require a bit of tweaking to get rid of some electronic noise, or some thermal issue in the detector or other odd thing that cropped up in the instrument during an observing run, and it just takes time to figure it all out. As an example of what I mean: The data that is going to be released on June 8, will not be the end of the line for that data. As time goes on and we gather more calibration data from the instrument, there is littlle doubt that we will be reprocessing all the data and release improved versions of all the products we have released already.

In the case of MRO and all the instruments on board, we're still getting to know them while trying to keep up with the huge data volumes, which is a tricky feat at best. So the color products will come, eventually, but it's probably going to be a while longer before any of the planned color products come out.

As DataMiner stated, HiRISE has adequate budget for it's needs. The problem is more of time than money. There are some tasks that you simply can't share easily among several people, it's more efficient to be done by just one person at a time. As the saying goes, too many cooks spoil the broth. If too many people were working on the color processing, they could be taking tangents, and in the end, would have to splice everything back together again. It could lead things to become much more complex, and not really accomplish anything. Also, the calibration becomes much easier when you have hundreds or even thousands of images to work with, as we have right now. No amount of ground work can ever fully replace time actually working in it's final configuration, and using that data to help. There can be hundreds of tests done, but only by actually seeing the final product can all of the kinks be worked out. Models work good to find the source of first order problems, but second and third bring variety, and aren't always as easy to find until you see them. Also add the fact of the detector noise that we've seen in some of the CCDs, increased warm up time which creates some image artifacts, and you can see that things get more complex.

Well, I think that's about as much as I can offer on the subject now. The color images are simply waiting for absolute calibration, which takes a while. But I can promise that they are coming.

I guess work has an irritating way of intruding on one's blogging time