NASA has huge databases of photos, some numbering in the hundreds of thousands to millions. Most of them have been archived as-is for four good reasons. Archiving the images as-is preserves the historical quality of the images. The job of cataloguing and digitizing in various sizes and file formats is in itself a monumental task. Adjusting images can sometimes be rather subjective. Space is dark.

Many photo images from Nasa not only can be improved by but, also require adjustment. Simple steps can yield moderate to great improvements.
For those who may be unfamiliar with these basic operations we have included some screen shots.

We often discuss the differences
between various files types such as .tif or .jpg with people from the Bristol Biomedical Archive, the Amarican Society of Microbiology, National Geographic Society, PhotoDisc, TalkGraphics.com, NASA, the CDC, NIH, Wadsworth.org, NARA.gov and many others. We will discuss here tiff and jpg viz NASA images.

Due to the manner in which images are initially created there are all manner of artifacts which can be inherent in NASA space images. Lines both vertical and horizontal, checker patterns, moire, circles, speckles, blips, blobs, rays, glares, fuzziness, graininess, etc. may be present. Lossless image formats simply preserve defects faithfully. Larger pixel size for many NASA images often exaggerates these imperfections and makes them appear worse.

TIFF or JPG?
In practical terms the question becomes, Download image as .tiff or view as .jpg?
That is the question. Our answer is quite simply, for use on the internet, neither. PNG would be our first choice. For offline uses tiff would have an advantage. TIFF or Tagged Image File Format has its devotees to be sure. And we would undoubtedly ruffle some feathers if we were to say that we are not one of those devotees. The same would be true for jpg. If the graphics world would stop using both of these formats and use png instead, a lot of unnecssary ifs, ands and buts would be eliminated. Working with images would become much simplified. Png could well replace gifs as well. That said, we are reminded that the question is tiff or jpg. Our assessment of tiffs began originally about a decade or so ago. We exhaustively compared hundreds of images from a large variety of sources. PICT and BMP were perceptibly better in quality than TIFFs consistently. We also found that consistently the difference was enhanced when images were printed instead of just being viewed on a monitor. We could only barely dsitinguish a difference between the so-called lossless formats and jpgs most of the time. Our vision is less than 20/20. So we always wear glasses when making a comparison showdown of that sort. Over the years we have revisited tiffs many times on numerous platforms, systems, monitors, and software applications. We remain unconvinced that tiffs represent something other than just another file format which needs to be converted to something else. When a form of internet started we were not surprised that people would not waste time with a 2,400 baud modem transfering tiff images. Talk about a long wait. None of our viewers have likely seen such a slow modem. We had not cast our opinion in stone but, we had rather believed that tiffs was a dead issue. And then with faster transmission speeds, we started seeing tiffs again here and there on the internet. Is bigger byte size necessarily better? If so, why not .tga, the targa format? It's true color. It can be transparent. It's got large byte size like tiff. It also is not as good as png in our comparison testing. Some people could have tiff on the brain.

Maybe there is something wrong with us. Maybe it's a psychological defect. Maybe it's a genetic defect. What else could it be? Any compression changes an image some say. We can't argue with that. That's absolutely true. But if we can't see the change, the point is moot.

We are also tired of fussing with jpg images. The fuzzy static you can sometimes get upon applying JPG compression is a bother. They aren't all that bad though. Sometimes you can clean it up by switching to index color. That has the effect of reducing and forcing the palette and the static disappears. We rarely use anything but photo jpg for digital movie compression. In that case it gives great results. Jpgs have become an internet standard. Why? From the outset the two main requirements were compression and image quality. Many possible formats were considered. Many experts could agree that images compressed with this format were the closest facsimile to originals. Upon being compressed the reduction in byte size is very substantial. Quality? It is so close that you wonder which is which.

Png could be a standard too. The compression is similar, the quality is better, and it can include transparency. Browsers accept pngs easily these days, too. Some people could have jpg on the brain.

Actual examples:

Below are thumbs of three tiffs and three jpgs along with the file information.

PIA00024_16712
615x542pixels, left side cropped by 108 pixels, not resized
17.1 Kb

PIA00349_17160
482x690pixels
24.2 Kb

PIA01957_16041
541x615pixels
62.9 Kb

We randomly chose three pairs of identical color images from the planets Photo Journal. One tiff image, was just too large at 1 Mb and was resized down. The jpgs were quite small so both the original and adjusted images can be viewed. When the images were initially viewed, we were unable to distinguish between tiff and jpg. We would view an image an guess that it was a jpg and then check the name and find it was a tiff and vice versa. But we weren't going to give up. We were determined to try to discover what the difference between TIFF and JPG is. Not only that, we were determined to try to find a way to demonstrate to our viewers what the difference is so that they would be able to view it for themselves.

The first breakthrough came when we noticed that when we viewed thumbnails of the images in Ulead's Photo Explorer, which had come in a bundle with Ulead's Photo Impact, we could detect a definite and consistently demonstrable grainy appearance in thumbs of jpg images which was absent in thumbs of tiff images. The effect disappeared when the actual images were viewed at actual size, and compared side by side, but it was a start. See for yourself:

After another day of looking back and forth at the tiffs and jpgs, our eyes became accustomed to the nuances and subtleties of the images. We noticed a difference between images which can be demonstrated to viewers. We took screen shots of the two side by side comparisons and saved them at 24 bit uncompressed. This faithfully reproduced the differences as we viewed them on our screens. The differences were also faithfully reproduced when we compressed the images for you to view here. It has been pointed out that the image of the tiff has been compressed in these two comparison images. That's true but, the point is to show a difference between the jpgs and the tiffs. It is difficult to see a difference between the two formats. It's more difficult to demonstrate that difference. We have succeeded in demonstrating that difference.

The originals are unchanged. You may view and download them either from the thumbs above or directly from the NASA Photo Journal by entering the image id numbers and conduct your own side by side comparison.

Whether you download a tiff or a jpg, you still may want to apply PhotoShop filters. That is how you can really improve an image. It has been said that the image which can not be improved by applying the PhotoShop Variations filter does not exist.

Gaussian Blur: Using a gaussian blur filter will enhance the image by just slightly deemphasizing the rectilinear pattern of lines caused by the way the images are usually created. They are composited from many smaller pieces of data. The most subtle of the blur filters is the Gaussian Blur, so let's use that. Open file in PhotoShop. Select the Filter menu, then Blur, then Gaussian Blur. Enter a value of 0.3-0.4 and click OK. The result is that the .jpg image now looks better than the .tif image. After applying a slight gaussian blur, we applied the Variations filter in a further effort to minimize the inherent artifacts. Resizing down to a size which would better fit a web page again smoothed the image but the artifact remained.

Variations Filter, PIA00024

How about a direct eyeball test? Let's say our hypothesis is that we may have a genetic defect which will prevent us from detecting the difference between two closely similar colors at a bit depth of 8 bits per pixel. That is 256 possible colors, not including browser limitations. We have created a table with four colors. If we can see the difference we may not be genetically inferior. If we can't see the difference, well, we may be in trouble. We've used closely similar 8 bit colors that are as close as we can get with your browser. If we can't se a difference with closely similar eight bit colors, how can we hope to distinguish between two contiguous colors at 24 bits, which is over 16 million color variations?

OK. We can see the difference between these colors. No proof of genetic defect. Great.

What about a 24 bit color depth? Well in that case there may be as many as 15 thousand gradations of color between each of the colors in the table above. If you could make a color table with 16 million colors for a bit depth of 24 bits, two contiguous colors in the table would look very very very very very very similar.

32 bit? Forget it. No way, Jose. For us that is absolutely impossible.