All STScI Archive servers will be unavailable Thursday, May 23 from 6:00 PM through midnight ET. This includes archive.stsci.edu, hla.stsci.edu, galex.stsci.edu, mast.stsci.edu
When converting preview images from FITS to GIF, we have a problem: the FITS image uses 16 bits per pixel, while a GIF image uses only 8. In technical terms, the FITS images has a much wider dynamic range than grayscale GIF. That means that areas of fine detail, like the arms of a spiral galaxy, may get clumped into one color value in the GIF. The Pixel Correction selector on the Preview Form lets you decide how to brighten the pixels in the FITS preview before they're converted to GIF.
What a pixel correction does is to exaggerate the lower-valued pixels so they can be differentiated from each other, without changing the higher-valued pixels much at all. This works great on HST previews, because the higher-valued pixels are usually just a cosmic ray, or a very bright point right in the center of a galaxy, while the arms of the spiral, say, might have much lower pixel values than that.
Here's an illustration of what I'm talking about. It's one of the best images HST has ever done, and one of the first images done with the WFPC2, right after it was installed in the 1993 Servicing Mission. The target is M100. Note how as you go up in pixel corrections, the spiral arms of the galaxy become more prominent, until you get to "Brightest", where they're probably a bit overdone.
Here's another example: The gravitational lens in the galaxy cluster Abell 2218:
Abell 2218 (U2HS0303T)
I think you might agree that the "Brighter" correction looks best here. "Brightest" doesn't do much but wash out the background without contributing detail; "Bright" and "None" don't brighten the image enough to see it. Since most of HST images are like this- low signal-to-noise data of galaxies and such like- I made "Brighter" the default pixel correction value. However, it's not always the best choice...
Using "Brightest"The "Brightest" pixel correction can be useful particularly in images where the signal (the image of the target) is not very bright against the noise (the background clutter). Here's an example, using the Ring Galaxy AM0035-335.
the Ring Galaxy, AM0035-335 (U2JB0104T)
The ring is certainly visible in the "Brighter" corrected image, but it's still dim and doesn't show much detail. The "Brightest" brings out the ring clearly, though at the expense of bringing out the noise, too. (By the way, most of the little bright specks all over the image are cosmic rays hitting the camera chips.)
So how do you know which one to use? Well, you don't. You just have to take a look at the default image and then experiment from there.
PlanetsUsually, if you take the defaults for a preview image of a planet- say, Mars- you get something like this:
The problem with this image is that the default "Brighter" pixel correction has washed out the image of the planet, so no features are visible on the surface. But also, the image of Mars is so small that all its features will be compressed out anyway. The solution to this is to 1) use a lower pixel correction, and 2) to use the PC-only preview.
See how Mars is centered in the small square in the upper right-hand corner? That's the PC chip. The other three squares are from the WF chips. That's where the WFPC2 gets its name: Wide Field Planetary Camera, version 2. Anyway, most of the WFPC2 previews have a PC-only version available, which you can get by hitting the button labelled "PC chip only" on the preview form. (If you don't see this button on the form, then a PC-only preview is not available.) This will give you a preview not of the entire mosaic of all four WFPC2 chips, but rather of only the PC chip.
Here is the PC-only preview of the Mars observation above, with different pixel corrections. I think "None" looks the best.
Mars, PC chip only (U2H50403T)
Of course, there are exceptions to the PC-only-and-no-pixel-correction rule. For example, the planet may not be centered in the PC chip (Saturn almost never is). In fact, a lot of images don't have a PC-only preview- most notably, most of the preview images from 1994's collision of comet Shoemaker/Levy-9 with Jupiter. (Usually, only one or two of the WFPC2's four chips were read out, to prevent the rapid succession of images from overloading the on-board tape recorders.)
Also, it's not always the planet's image you're after. Here's an example of a "Comet Crash" image, through the narrow-bandwidth 7200 Å filter. By turning up the pixel correction to "Brighter", you can clearly see the plume of a comet fragment collision on the limb of the planet. You can still see the plume at lower pixel corrections, but you can't see the planet too well. The "Brighter" pixel correction brings them both out. (It also brings out the other impact sites on the planet in this image.)
Comet Shoemaker-Levy Colliding with Jupiter (U2FI0A09T)