by jhayes_tucson » Wed Nov 02, 2016 6:05 pm
Ann (and others,)
Thanks for your questions. Narrowband imaging is quite different from RGB imaging. First, by looking at the object with narrowband filters peaked at different wavelengths, it is possible to pick up detail in the object that might be barely visible with broadband filters. Remember that a red filter lets through a wide range of spectral lines from all of the different elements in the object. If you restrict the view to within about 5 nm around the Ha line (for example,) you greatly increase the contrast of features that exist in the hydrogen clouds only. Second, narrowband imaging always maps the result into a false color palette. The normal Hubble palette (HSO) maps OIII into blue (which is roughly correct,) Ha into green (which isn't even close,) and SII into red (pretty good.) Since Ha is normally the dominant element, its emission almost always swamps everything else so an additional step has to be taken to normalize all of the signals to the weakest signal (usually sulfer.) Even in the best processed Hubble images, the final colors don't do a good job of showing the actual distribution of gasses in the object. If you want to understand gas distribution, you need to look at the raw images for each channel. What the HSO palette helps to do is to paint a stunning scene of detail that is not otherwise seen with broadband imaging--with maybe just a little bit of information about how the elements are distributed. The colors are mostly an artistic choice.
When you use only two filters (as in this image,) you have to figure out how to map the result into something that is visually revealing. M27 has very strong signals in both OIII and Ha and they overlap quite a bit. If you simply combine the images without proper pre-processing, you end up with a screaming violet image that shows little detail. The most important part of processing a bicolor image is to come up with a synthetic green channel that best splits the contributions from each element. I spent over four months and made over a dozen attempts to process this data before I realized that the trick is to spatially separate the color components using something called local histogram equalization. This approach can essentially separate the high spatial frequency components from the lower spatial frequency components for each element, allowing the underlying colors to come through. As with any narrowband image, the goal isn't to exactly map the location of the contributing elements; it's to bring out the underlying detail using a pleasing false color palette that approximates how the image detail maps back to the elements themselves. Clearly, the result is something striking and at the same time, a bit unfamiliar. That's because you are seeing features in the object that aren't always so immediately obvious. In this case they are cosmic shockwaves in interstellar hydrogen and oxygen.
I hope that explanation helps to better understand what you are seeing in this image.
Best regards,
John
PS For anyone who is interested, this image was taken with a Celestron C14 Edge on an AP1600 mount. The camera is a FLI ML16803 and focus was held in real time (with the shutter open) using Optec FocusLock and an IFI ONAG guider. The total exposure was 6.3 hours @-25C. Processing with PixInsight and a bit of PhotoShop. Location: "Fly by Night" Roll Out Observatory, Bend Oregon
Ann (and others,)
Thanks for your questions. Narrowband imaging is quite different from RGB imaging. First, by looking at the object with narrowband filters peaked at different wavelengths, it is possible to pick up detail in the object that might be barely visible with broadband filters. Remember that a red filter lets through a wide range of spectral lines from all of the different elements in the object. If you restrict the view to within about 5 nm around the Ha line (for example,) you greatly increase the contrast of features that exist in the hydrogen clouds only. Second, narrowband imaging [i][u]always[/u][/i] maps the result into a false color palette. The normal Hubble palette (HSO) maps OIII into blue (which is roughly correct,) Ha into green (which isn't even close,) and SII into red (pretty good.) Since Ha is normally the dominant element, its emission almost always swamps everything else so an additional step has to be taken to normalize all of the signals to the weakest signal (usually sulfer.) Even in the best processed Hubble images, the final colors don't do a good job of showing the actual distribution of gasses in the object. If you want to understand gas distribution, you need to look at the raw images for each channel. What the HSO palette helps to do is to paint a stunning scene of detail that is not otherwise seen with broadband imaging--with maybe just a little bit of information about how the elements are distributed. The colors are mostly an artistic choice.
When you use only two filters (as in this image,) you have to figure out how to map the result into something that is visually revealing. M27 has very strong signals in both OIII and Ha and they overlap quite a bit. If you simply combine the images without proper pre-processing, you end up with a screaming violet image that shows little detail. The most important part of processing a bicolor image is to come up with a synthetic green channel that best splits the contributions from each element. I spent over four months and made over a dozen attempts to process this data before I realized that the trick is to spatially separate the color components using something called local histogram equalization. This approach can essentially separate the high spatial frequency components from the lower spatial frequency components for each element, allowing the underlying colors to come through. As with any narrowband image, the goal isn't to exactly map the location of the contributing elements; it's to bring out the underlying detail using a pleasing false color palette that approximates how the [u][i]image detail[/i][/u] maps back to the elements themselves. Clearly, the result is something striking and at the same time, a bit unfamiliar. That's because you are seeing features in the object that aren't always so immediately obvious. In this case they are cosmic shockwaves in interstellar hydrogen and oxygen.
I hope that explanation helps to better understand what you are seeing in this image.
Best regards,
John
PS For anyone who is interested, this image was taken with a Celestron C14 Edge on an AP1600 mount. The camera is a FLI ML16803 and focus was held in real time (with the shutter open) using Optec FocusLock and an IFI ONAG guider. The total exposure was 6.3 hours @-25C. Processing with PixInsight and a bit of PhotoShop. Location: "Fly by Night" Roll Out Observatory, Bend Oregon