Lenshacker Posted March 15, 2015 Share #1 Posted March 15, 2015 Advertisement (gone after registration) A new Color to Monochrome conversion process. Coded up on Pi Day. Within each 2x2 Bayer cell, compute and Scale the ratio of Red:Green and Blue: Green. Convert to a Complex number and then use that in the Mandelbrot algorithm. Output stored as Photoshop RAW. L1015634 by fiftyonepointsix, on Flickr FRACTAL by fiftyonepointsix, on Flickr What did everyone do to celebrate Pi day? Link to post Share on other sites More sharing options...
Advertisement Posted March 15, 2015 Posted March 15, 2015 Hi Lenshacker, Take a look here Where no M8 has Gone Before. I'm sure you'll find what you were looking for!
jaapv Posted March 16, 2015 Share #2 Posted March 16, 2015 Interesting, but I fear the example does not convince me... Link to post Share on other sites More sharing options...
Lenshacker Posted April 1, 2015 Author Share #3 Posted April 1, 2015 This is using a Red filter (600nm) with the M8, 1934 5cm F2 Sonnar, wide-open. M1015846 by fiftyonepointsix, on Flickr Full-res jpeg uploaded, click and hover over the image to see the detail. Green and Blue have enough response in the red region, need to be boosted ~ a factor of 3 to match red. This is basically making use of the monochrome nature of the sensor, and the color bleed in the dye used in the Bayer mosaic filter. Blue and Green channels have been combined, only 1 pixel of the 2x2 Bayer cell is interpolated for 1 "green-Blue" layer, this is added with the interpolated red layer. As response is reduced on the Blue/Green channels, the "white" saturation is ~3x the 14-bit single-channel maximum. I have "White" set to 45K on these, and set it up a notch from there on the latest cut. M1015861 by fiftyonepointsix, on Flickr Link to post Share on other sites More sharing options...
VIEWSTER Posted April 1, 2015 Share #4 Posted April 1, 2015 I have know idea what you are saying. It looks to me like a out of focus low resolution inferred photo. The M8 is capable of shooting some of the finest IR photos of any camera made. Link to post Share on other sites More sharing options...
Lenshacker Posted April 1, 2015 Author Share #5 Posted April 1, 2015 Most people don't understand what I am talking about, probably the reason I've spent 40 years writing code for computers. These images are taken with a Red "R60" filter with a cutoff of 600nm. Infrared starts at 720nm, and an R72 filter is used for Infrared with the M8. The Blue and Green channel are picking up a good bit of Infrared, but it is about 1/10th as strong as the Red channel. This is what the image looks like out of camera, RED dominates the image. L1015853 by fiftyonepointsix, on Flickr This is what I am doing by boosting the Blue and Green channels to the intensity level of the Red channel. C1015853 by fiftyonepointsix, on Flickr Note the 1/125th second shutter speed on the EXIF data, ISO160, F5.6 used. This is the image converted directly to Monochrome, using the same linear-Raw format used by the M Monochrom. M1015853 by fiftyonepointsix, on Flickr The second and third images are DNG (Version 1.1 compatible) files output by the FORTRAN code. I'm having fun with this, I wrote this software decades ago- added the routines to interpret and write DNG files. I will try with the R72 Infrared filter. Link to post Share on other sites More sharing options...
pgk Posted April 2, 2015 Share #6 Posted April 2, 2015 This is what the image looks like out of camera, RED dominates the image. Lenshacker It would really help to also shoot and post an image of the scene in colour, without the filter. as this would give an indication of the actual scene. Otherwise its not so easy to determine what you have achieved as there is no colour reference. Link to post Share on other sites More sharing options...
silvanoamato Posted April 2, 2015 Share #7 Posted April 2, 2015 Advertisement (gone after registration) Dear Lenshacker, Trovo interessante il tuo approccio; ho sempre desiderato entrare più in profondità ai file raw (in questo caso i DNG della M8) ma... come accedo ai valori dei tre canali? Silvano Link to post Share on other sites More sharing options...
silvanoamato Posted April 2, 2015 Share #8 Posted April 2, 2015 Dear Lenshacker, Trovo interessante il tuo approccio; ho sempre desiderato entrare più in profondità ai file raw (in questo caso i DNG della M8) ma... come accedo ai valori dei tre canali? Silvano Sorry, I pasted text in italian... I find it interesting your approach; I always wanted to get deeper into the raw file (in this case the M8 DNG) ... but how do I access the values of the three channels? Link to post Share on other sites More sharing options...
Lenshacker Posted April 2, 2015 Author Share #9 Posted April 2, 2015 Thankyou- Truly wish I knew Italian! I use M8 in it's service mode, and M8RAW2DNG to convert to DNG files. Arvid stores a single image in the file, so it was easier to begin with. I have been posting details of the software here: Experiments in Monochrome Conversions | Leica Place - Leica User Group The DNG file stores the image data row-by-row, with the Bayer pattern RGRGRGRG.... GBGBGBGB.... The number of rows, number of columns, and "offset" to the image data is given in the file header, the "Image File Directory". This is a table of "Tag Values and Descriptors" for the file. It's not hard to parse, was more tedious to type in. I listed the subroutine that processed the "IFD" for this file at leicaplace. I think in Fortan. I will put up some more detailed notes for columns and rows when I get back home. When I get a free weekend day, will take the M8 and the M Monochrom on an outing and get some comparison shots. Will also remember to use the M8 without the filter to show the original scene. Link to post Share on other sites More sharing options...
Lenshacker Posted April 3, 2015 Author Share #10 Posted April 3, 2015 I used a B&W 090 filter, 5x RED. This is without the filter, L1015873 by fiftyonepointsix, on Flickr This is the in-camera JPEG, L1015874 by fiftyonepointsix, on Flickr This is the output from the Fortran program which equalizes the blue and green channels with Red. These are all done using Arvid's M8RAW2DNG and the custom post-processor. Having the full 14-bit uncompressed image allows this process to work, DNG-8 simply does not store enough intensity values to allow the curves to be equalized. I will clean up the routine and post it on Leicaplace. I1015874 by fiftyonepointsix, on Flickr The Blue channel is almost all Infrared, Green channel is about evenly split with Red and Infrared. This next image is done by rotating the Hue in Photoshop. I used to shoot Infrared Ektachrome, which was visible+infrared. In EIR, Infrared used a Red Dye, visible light was Blue and Green. I1015874_2 by fiftyonepointsix, on Flickr With the filter, I could shoot at F4 and 1/250th~1/500th on a Sunny day. 28mm Elmarit used. You can of course do this all in photoshop by using a red (R60, B&W 090) filter and using curves on the individual channels. My algorithm simply takes that step out of doing in photoshop, and automates it. The ability to use a standard camera (as opposed to converted VNIR) Visible+Infrared is unique to the M8 as the IR cut filter is weaker in it. A visible-blocking R72 or B&W 092 will cut out visible-Red and leave just IR. The image shown above does not cut out visible-Red. Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.