Adobe Lens Profile Creator

I thought I would share my experience using Adobe Lens Profile Creator. Since I usually only use professional high quality lenses I normally don’t worry too much about lens distortion. I profiled my Nikon 85mm f/2.8 PCE tilt-shift lens and found that lens correction is completely unnecessary when not shifted or tilted.

With astrophotography correcting lens distortions becomes a very desirable endeavor. Adobe Lens Profile Creator is normally used to profile camera lenses but I am looking at this from the perspective of evaluating the feasibility of using it to profile astrophotography lenses.

The free download from Adobe includes several targets of differing sizes and grids that must be printed, some are large so a trip to a local printer may become necessary. These targets are photographed with a particular camera, lens, and several scenarios which are combined into an extensive profile that can be used in Camera Raw, Photoshop, Lightroom, or any other imaging software that uses the Adobe *.lcp profile file structure.

Here is one set of images for a particular aperture, focal distance, and focal length:

_DSC3114_matrix

Adobe claims that a minimum of three separate images are needed but recommends a minimum of nine separate images with the target grid being 1/4 to 1/2 of the frame size for best results.

The program automatically detects the grids and calculates the distortion corrections necessary and creates an *.lcp profile file that can be used by the imaging programs. The program was created for Adobe Digital Negatives (*.DNG files) but TIFF and JPEG files can also be used, however, if a RAW digital negative file is not used for the profile creation the profile will not be accessible for lens correction of RAW files directly.

Here is a cropped and enlarged example showing the before and after shots demonstrating the correction. A line with a width of one pixel was drawn from one corner of the grid to the opposite corner, the distortion in the middle shows a displacement of nine pixels contributing to the barrel distortion of this particular lens:

_DSC3115_uncorrected_pixel

 

And here is the lens corrected shot, note the perfectly corrected geometric distortion. Note that the lateral chromatic aberration was not as successful but the vignette correction, not shown here, was corrected very well:

_DSC3115_corrected_pixel

 

The poor correction for the chromatic distortion may have been due to the fact that the target was lighted with five different color temperatures. Adobe recommends a  uniform light source.

Now how would I translate this procedure to an astrophotography lens? This may be difficult since camera rotation will shift how the correction is applied. The camera would have to be indexed to the scope in order for a profile to become useful. Also the scope would also have to focus without lens rotation, another side benefit of professional camera lenses. Adobe recommends different focal lengths and apertures and distances be used to create a complete profile, however, the astrograph would only need a single set for a complete profile.

The target would have to be in-focus for the normal infinite focus position used for astro-imaging, so the target may have to be quite a distance away, up to a mile or so for a very long focal length SCT. The resulting target would have to be large and the grid would have to be adjusted such that the individual blocks in the grid be at least the recommended minimum of twenty pixels wide.

Another concern of mine would be the assessment of how much the lens distorts optically under differing temperatures and scope orientations.

Most astro-imaging imaging software does not correct for lens distortions but instead  picks a master image for referencing how to manipulate the remaining images for alignment purposes.

If the individual images could accurately be lens corrected by a program such as Adobe Lens Profile Creator PRIOR to any other processing the resulting images might become superior, or at least the subject positions would become more accurate and possibly more scientifically useful. Less loss of very fine detail from the statistical noise reduction algorithms may also become a very desirable and useful benefit.

Added 2013-03-07:

I would also like to add that there is NO amount of processing that can replace or be superior to an accurately manufactured (distortion free) optical lens group. There is also NO amount of processing or statistical trickery that can be applied that would be equivalent or superior to a high quality (read non-amature) telescope mount.

Unfortunately for the amateur crowd this translates into very expensive professional equipment that for most is financially unobtainable. Time sharing on existing remote equipment then becomes a very desirable option. This also has an added benefit of the remote locations of these time-share set-ups being located in superior high altitude dark and relatively clear skies. In addition these setups are normally equipped with top of the line laboratory grade instrumentation that will nearly always result in vastly superior imagery, both aesthetically and scientifically.