|In Search of Solar Nirvana
(RGB levels in a monochrome H-a signal)
The Image on the left is the RAW unstretched image as delivered by the camera (downloaded in 16/48bit TIFF
format via Images Plus Canon conversion utility) and the fully processed image is to the right. Surface detail is
somewhat washed out, as I used a lenghtened exposure prominence shot for this sequence (to illustrate that even
at increased exposure levels, with a low-leakage camera the B and G channels will appear dark). RAW images are
preferred over JPEGs, as most cameras automatically stretch the latter, giving a false indication as to the true
RAW images appear darker and muddier than the data that has actually been recorded, as the only channel we are
interested in (Red) is mixed with an equal amount of the Blue and twice as much from the Green channel signals
(RGB digicam pixel allocation commonly is R=25%, G= 50%, B=25%). Since the Blue and Green channels should
not have any signal and are essentially black, the resulting composite RGB image is:
R+G+B or SIgnal + Black + Black = mud!
Also appearing black in the RAW version, the Blue channel has an extremely low signal level (intensity range = 0 to
684), with the stretched version displaying severe clipping and posterization.
The exposure level could be increased so as to exploit the higher pixel density of the Green channel, a somewhat
effective technique for maximizing image potential with those cameras experiencing high leakage. The downside
though is that longer exposure settings allow more time for atmospheric disturbances to smear the details, as well as
bleeding of the now saturated adjacent Red pixel wells.
The ideal H-a imaging device will properly confine the H-a signal to a single channel, but reality dictates we find the
methods that work best with our own unique equipment.
For more information on this particular image, and H-a imaging in general visit the10-Sep-03 page, or for the entire
Solar section visit the Solar Imaging section.
I am often asked what a typical raw H-a image looks like and what steps are used to process it. I wish to stress that I
am not a solar guru, but merely a fellow traveller sharing some insights he has gleaned along the way. Nor is this
intended to be a tutorial, as many splendid examples already exist. Successful methods for gathering and processing
of H-a images vary greatly and are largely a procedure of trial and error, fine-tuned for your specific equipment and
needs. Your path to Solar Nirvana may be quite different from mine!
If you have not already done so, you may want to peruse the primer "Variations in H-a Solar Exposures" and "In
Search of Solar Nirvana" Part1, and "In Search of Solar Nirvana" Part2, as they inter-relate to this page.
| Original RAW Image Completed Image
The image has been split into its RGB components to analyze the contents of each channel. The left image is the R
component as directly extracted from the RAW image and the right image has been auto-stretched. Note the
histograms which show that the RAW Red has captured almost the complete 64K intenstity range without clipping
(intensity range = 689 to 61,301). The image requires very little streching to fill the range completely, meaning
maximum detail has been captured and preserved.
RAW Blue Channel Stretched Blue Channel
|The above snippits were gleened while continuing my search for Solar
Nirvana, and may not necessarily mark the path you should take!
Solar H-a filters are precision devices, designed specifically to allow only signals at 656nm to pass. This is solidly in
the Red portion of the spectrum, thus no signal whatsoever should be found in the Green or Blue channels. This
does not present a problem when gathering images with a dedicated astro-CCD, as they are typically monochrome
(grayscale) devices. Likewise, film accurately records the level of detail, albeit in a singular Red-tone.
It is not uncommon for problems to occur however, when a digital camera is used to gather the images, as the RGB
pixel layout/filtering system may allow leakage... loads of leakage in some instances! That topic is covered
elsewhere though; the purpose of this section is to illustrate the signal content in a low-leakage system... which in
this case is a Canon 10D digital SLR employing a CMOS detector array.
RAW Red Channel Stretched Red Channel
In stark contrast, the Green channel appears black, and the histogram attests to the low signal level present
(intensity range 31 to 2213). When expanded to 64K, the stretched version is somewhat coarse, with mild clipping
and posterization (areas of distinct demarcation).
RAW Green Channel Stretched Green Channel