Nikon D5100 Monotec modification initial tests

[fwm891]

I've just received a Monotec monochromatic camera body modification from Herra Kuulapaa.

Other postings with this mod: http://stargazerslounge.com/topic/236643-nikon-d5100-monotec-narrow-bands/ and http://stargazerslounge.com/topic/250891-nikon-d600-monochrome-images/

Cloudy since it arrived! so this is a simple daylight test using a Nikon 200mm lens at f2.8 and iso settings of 100 and 400. Inset images are a 1000% up scaling of the original image area around the tree top. Section of the full frame image spans the full height of the frame.

Lens/camera combination were static tripod mounted, captures and settings made through BackYard Nikon. Image files were then grey-scaled in PI and the scaled insets created and applied in Photoshop CS3 Extended. No levels, curves or other modifications have been applied to the images.

I will continue this thread once I get some clear skies and can mount the camera on a scope.

[fwm891]

Been playing again this morning with an old Tokina 400 f5.6 lens (great lens you can put a 1¼ inch filter in the back end of it, as its meant to fill a full frame format using it with a DX/APS-C format means theres no vignetting) - anyway I digress.

So I cut a piece of Wratten 87 gel filter and fitted that into the rear of the lens to try some daytime infra-red imaging - very effective and very sensitive. Lens was wide open at f5.6 but iso was 200 and shutter speeds ranged from 1/60th to 1/250th second.

Then I tried a short sequence of cloud shots: Intervals are 5 seconds.

Very please so far - its lighter than the older Nikon so I need to go and re-balance the scope, I also don't need the 1x field flattener as the AA 115 EDT-APO more than covers the Nikon DX format sensor.

[fwm891]

After the daytime 'play' I managed some night work.

First image is the Eastern Veil (NGC6992) comprising 12x575second subs at 800iso: (the 575s subs were done so I could fit 3 dslr exposures during one 1800s subs I was dual running with the Atik383L+.

I had to break from some cloud at this point and when things eventually cleared I tried a different target: IC1848 part of the heart and soul area:

Started here is 10x600second subs at 800iso:

Then tried same subject but iso 1600 but keeping the 600s sub duration:

I started processing these in PI but encountered a number of problems so moved to a Lightroom / Photshop combination - thinking these programs are more suited to dslr images...

In lightroom I first converted to monochrome (B&W), altered the exposure value to centre the histogram peak, the stretched the black and white points (without clipping) removed any sharpening lightroom wanted to apply and made final tweaks to the histogram before synchronising that frame with the others in the set. Files were exported as 16 bit tiff's for further work in Photoshop.

Photoshop dealt with first the alignment of the fames: I copied each image into a new photoshop document so I had a nest of layers, set the top layer to 'difference' and aligned each of the other layers to it. Turned the top layer back to normal, selected all layers and made a smart object of them under the layers tab. Once they formed a smart object they could be stacked using a median routine (under the smart object tab).

Just remained to tweak in curves, resize and save...

IC1848 is full frame, NGC6992 has been cropped.

Will go back and look at the original files and try PI again.

[Thalestris24]

Hiya

Looks like you might need to crop the edges off prior to processing. There is some info on debayering cr2 files in mono on this thread: http://stargazerslounge.com/topic/234030-processing-bw-raw-files-from-a-debayered-dslr/#entry2532250

There might be some others around too.

Louise

[fwm891]

Hiya

Looks like you might need to crop the edges off prior to processing. There is some info on debayering cr2 files in mono on this thread: http://stargazerslounge.com/topic/234030-processing-bw-raw-files-from-a-debayered-dslr/#entry2532250

There might be some others around too.

Louise

Hi Louise,

I feel I need to respond to this reply as you are quite in error:

If you look at the title of the thread it relates to a NIKON D5100. Nikon cameras produce NEF RAW files its Canon that produce CR2 raw files.

I also state in the thread that the NGC6992 image has been cropped the IC 1848 images are full frame.

These images are not designed to be pretty pictures they are test images from a newly acquired camera. The camera is a heavily modified DSLR where not only the IR blocking filter has been removed but the RGGB mosaic and the micro lens layers have also been removed giving a bare monochromatic sensor, sensitive to the whole spectrum - hence the IR landscape test image.

Processing these images is causing me the biggest headache at present as the camera seems to do all that's asked of it, the files however need to be handled differently to the normal DSLR raw files.

TBC...

[Thalestris24]

Hiya

Yes, I appreciate they are test images but you said you had problems processing in PI - I thought that might be due to the 'ragged' edges that could be got rid of by cropping. Sorry for confusing things by mentioning cr2 files. I just supposed that debayering might need to be done in the same way as with 'mono' camera cr2 files which you can do via IRIS, as I understand. I believe you can do the same with nef files. Normal software presumes you still have a colour matrix so with cr2 you need to convert to fits to allow dss, for example, to stack them as monochrome. I presume the process would be the same for nef files.

Was just trying to help

Louise

[Herra Kuulapaa]

Hi,

Sorry about this huge image. Issue Photoshop etc processing is the assumption they make; camera is unmodified. This results demosaicing process to take place, which then affects the image on a pixel level. It also makes some pixels more average, but boosts the other. I enclose a single exposure of the Crescent. Red/magenta image is how PI sees it with debayering. One on the upper right is a greyscale conversion of it. And below is a true raw treated with image calibration (dark, bias) waiting for alignment with other subs and stacking.

It's possible to make a decent image of the just greyscale converted subs, but it's a good method to enable true raw option in PI format explorer/DSLR_RAW/Edit preferences in order the get most out of the data.

These are just auto stretched examples of screen transfer function and screen caps.

Ps. Enjoy the clear skies if you have such

[fwm891]

Thanks Herra - the dslr-raw file setting has helped a lot. I've also changed some of the integration settings which has further improved the noise levels - so getting there.

NGC281 D5100 9x600s 1600 iso:

Superb mod - thanks

Francis

[Herra Kuulapaa]

Have you tested the difference between so called unity gain ISO and higher ones?

It would be really interesting to see the difference between ISO400, 800 and 1600 with same exposure time, for example 600s.

I have a theory that a lower ISO on this sensor provides the same data, but with less noise. I haven't been able to prove that yet though...

[fwm891]

Have you tested the difference between so called unity gain ISO and higher ones?

It would be really interesting to see the difference between ISO400, 800 and 1600 with same exposure time, for example 600s.

I have a theory that a lower ISO on this sensor provides the same data, but with less noise. I haven't been able to prove that yet though...

Your wish is my command...

[Herra Kuulapaa]

Ok, now I'm officially feeling a little bit weird

Anyway, we should be able to pull up the same signal in PP than what camera gain does. And in theory we should be able to do even better, if we do the signal boost selective way by increasing the signal major and the unwanted noise only a minor amount.

ISO400 is a little bit higher than what unity gain (1e/adu) is, but it is nicely in line with ISO800 and ISO1600.

What we're needing to achieve this is a sensor with good enough noise characteristics. Is the D5100 sensor good enough, I don't know, but it has given good results so far with lower ISO without increasing the total exposure time...