Forums / The Science / Photometry / Nikon D200 with MaxIm DL and a question about Bayer planes
Nikon D200 with MaxIm DL and a question about Bayer planes
Background: I am an experienced CCD imager and photometer. This is, however, my first work with DSLRs. So... some questions.
Tonight is night-1 for testing. Equipment:
Nikon D200 on tripod
Laptop
MaxImDL: I intend to keep using MaxIm DL as it is the standard for my students. I know there are other fine (and free) products out there.
Lenses
No external filters
My intention is to use MaxIm DL to automate the imaging in monochrome RAW mode. Takes darks and biases and go from there. Bayer plane extraction is easy enough through a provided command in MaxIm. The result leads to 4 planes: Red, Blue, Green-1 and Green-2.
In processing this data, I can get through the use of biases and darks easily enough. What to do with both green planes though? Do these get averaged? Summed? My bet is that I should take a slew of raw images, letting the star track through the image frame from image to image, then average all the greens for a better response.
With any luck, I will get my transformation coefficients and and some useful data tonight.... as well as freeze a little bit.
All comments and responses welcome.
~johnb
johnb,Very cool! There is an undergraduate at the University of Denver who has been taking data on eps Aur using his Nikon camera for several months now. With his help, we have assembled a "how to" processing tutorial for MaximDL that will probably make it online this week after a few final revisions.If I recall correctly, his camera (D300) defaults to a 12-bit RAW file, but there is an option for a 14-bit RAW. You might poke around on the camera's settings to see if you can change it on your camera too.Brian
Hey Brian,That would be most awesome to be able to get the D200 up to 14 bits raw. I have combed the menus for any hope, but perhaps it is not an obvious entry there. I look forward to seeing what the U-Denver student has to say here. I will keep looking in the mean time.Went out tonight and did MUCH better with a 60mm portrait lens. The data are cleaner and certainly more easy to frame than a 300mm telephoto. Oh - and it is a balmy 10F outside tonight (not 2F like last night ;-).There is an interesting side to all this. A couple of years ago, I had a student who wanted to try to use DSLRs for bright star photometry. I let him do it as a senior project. The result was, well, less than stellar, largely because we didn't have the ability to separate the Bayer frames out properly. It was a slight nightmare. Now, with all the available software, it has become a doable thing. Times change, and our abilities improve. I love technology.~john
Hi! Sorry, while the D300 has a 14 bit A/D converter, the D 200 only has a 12 bit ADC. But many people here are doing photometry with 12 bit RAWs, so that shouldn't stop anybody. CS Heinz
Hi all,Agree with Heinz, 12 bits is absolutely no problem, with some defocusing we usually accumulate 100k ~ 200k ADUs into the integration circle, that's far enough resolution for making mag measurment better than 1% ! What is truly important for the SNR (and uncertainty reduction) is the number of electrons being accumulated on the imager itself not its rescalling to 12 or 14 bits (said gain) at level of the ADC.Yours truly,Roger
12 bits - no problem. A few years back I used an 8 bit modified webcam to detect this 0.02 mag depth exoplanet transithttp://www.threehillsobservatory.co.uk/astro/Photometry_SatN_step6.htm(Just take care not to saturate and keep stacking frames until you get enough total counts for the precision you are looking for)Robin
Aaron has asked me to share my blog posting here as well, so here it is. I spent last night clicking slews of images through the Nikon D200 and came up with a very basic process. Maybe this will help others who are starting out.Well, the data are in and looking ok. I was surprised to see just how easy it was to overexpose the star on a D200 chip: the chip goes to 12 bits not 16, so things ramp up and then peak at pixel values of 4096! Not good really, but I got the hang of it. I opted to select low iso values and then ended up closing down the aperture as well. Tonight (if clear) I will use a shorter focal length lens (60mm portrait) to make centering the field a bit easier. I used the 300mm telephoto last night… way too shaky for a cold night at 6F. No thank you.The data flow looks something like this: * Take images: both lights and equal time darks. Images need to be monochrome RAW Nikon files which contain the raw, unmeshed Bayer plane data. They look black and white, but the color data is really there. Honest. * When back inside open the light frame. * Subtract the averaged dark frame from it. * Check to see that the star images have ADU or pixel counts below 4096. * Extract the Bayer plane for the color of interest: MaxIm DL allows for this in an easy menu pick. Note that the colors are not easily determined: you need to test this out on a Macbeth color chart before going out. 1 = red, 2 and 3 are the green planes which then need averaging, and 4 = blue. If you want to play, you can color combine these planes to see the full color image. It’s neat. * Perform differential photometry on the extracted single color images. Be sure to use the correct comparison stars for the correct color. * Report the data. * Done.At this time, I am averaging a slew of images together to increase S-to-N.