Jump to content

Narrowband

Debayering a DSLR's Bayer matrix.


RAC

Recommended Posts

@D4N :

I compared, for each chanel (red, green, blue) the level of the debayered photosites and the levels of the intact photosites. The average gain is shown below for each chanel :

- red : approx +30%

- green : approx +30%

- blue : approx +75%

There is therefore no loss, at least on the 500D, but a real gain, in particular in the former blue photosites.

  • Like 1
Link to comment
Share on other sites

Hmmm, my D5100 project seems to be finished now and I just need to start using it. One observation about the sensor is that it's rather easy to make surface very uniform compared to Canon xxxD series.

Here are two examples of the D5100 monochrome. Both images are stacks of 4x600s / ISO800 through 12nm hydrogen alpha filter.

PA_1000.jpg

HH_1000.jpg

Or should I make active cooling to it? ;)

http://kuulapaa.com/index.html

  • Like 1
Link to comment
Share on other sites

Hi all!

I restarted the calculation of the efficiency of the CFA removal on the light gathering of a Canon EOS 500D/T1i. This time, I took a photo of a white cloud in the sky, close to mid day. The results is quite impressive :

- green chanel : +55%
- red chanel : +80%
- blue chanel : +95%

The overall gain is therefore (2x55+80+95)/4 = ca. 70%.

The older calculation was done from a photo shot using a tungsten light.

Link to comment
Share on other sites

I also calculated a flat field image. One can see that removing the CFA matrix is quite aggressive :

10293-1419684030.jpg

However all these scratches can be removed by the flatfield as shown below :

10293-1419688813.jpg

translation :

- poussières = dust

- rayures = scratches

- restes de matrice de Bayer = remaining patches of CFA layer

- avant/après application du flat = before/after flatfield correction

Fred

Link to comment
Share on other sites

  • 2 weeks later...

A few weeks ago, I outlined my proposed approach to debayering a Canon  550D. My approach will be to attempt abrasive "blasting" using an "air eraser" which is essentially an air brush that shoots abrasive instead of paint. I'll start out with simple sodium bicarbonate (often used as an effective blasting medium) and progress to more aggressive grits if necessary. The trick will be to find a grit that is hard enough to remove the CFA but softer than the underlying silicon. I'm ready to start and will report on my progress.

Today, I overcame the first major hurdle - removing the cover glass on the sensor. I have read here about the problems doing that, so I took a very careful approach. To begin, I left the sensor in its aluminum cell so I could better hold it. I then used several sharp razor blades to scrape the adhesive that had squeezed out from underneath the glass around the perimeter of the glass. This squeezeout adhesive is mounded against the edge of the glass and represents the thickest part of the adhesive. I took care to completely remove this adhesive (using several razor blades to ensure a narrow, sharp edge) and even dig out a little bit from under the glass.

The next step was to use a 60W soldering iron with a new chisel tip to heat the perimeter of the glass. I know that this has not worked well for some people here, and I didn't think it was working for me, but patience paid off. Using a magnifying light, I carefully placed the flat of the chisel tip against the glass. The important part here is the make sure that the flay of the tip is completely flat against the glass. The idea is to transfer as much heat to as much area of the glass as quickly as possible. I placed the tip against the glass for no more than 10 seconds (usually 5 or 6 seconds) - holding it in one spot for that time. I then moved the soldering iron to the opposite side of the glass so as not to heat one part of the sensor too much. Every minute or so, I stopped and waited a couple of minutes for the glass to cool. During that time, I worked the edge of the glass with the razor blade some more, being very careful not to work the glass too hard.

At first, I didn't notice any effect - even after about half an hour of doing this. But with time, I started to see some color fringing (like you see on an oil slick on water). This told me that the glass was starting to separate. Patience here is key. As I started to see the color fringing, I would place the iron at the edge of the fringed area and it would grow slightly, still moving the iron around the glass, and stopping ever minute or so to let the glass cool and to work the edges more with the razor blades.

With time, I had an entire edge showing the color fringing and I was able to work the tip of the razor blade under the glass. I resisted the temptation to pry and instead, just left the tip under the glass and went back to work with the iron. With the tip of the razor blade under the glass, things started to get easier since the tiny thickness of the razor blade tip was putting some upward pressure on the glass.

After another half hour or so, I had color fringing around about 75% of the glass perimeter and was able to get the full thickness of the razor blade under one edge. At this point, I reminded myself that patience is a virtue. I didn't want to rush things and break the glass. Shards of glass could easily damage the nefarious gold wires. So I took my time and was able to eventually remove the cover glass intact.

So, heat will work, given enough time and perseverance. It took me close to 3 hours.

More to come...

Tim

  • Like 1
Link to comment
Share on other sites

As Luis Campos and François Bernier noted, the original Canon glass cover is not good against reflexions of light when the sensor's bayer matrix has been removed. Therefore there is no need to keep it intact as it is better to replace it with a real A/R coated glass.

Anyway, this is a good job !

Link to comment
Share on other sites

I got infected, by an idea to cool down my Nikon D5100 monomod. For that I did some temperature tests with ISO1600/10min dark frames. An average temperature during the exposure was measured with 0,1°C precision meter attached directly to sensor and noise was analyzed with Pixinsight noise analysis tool.

Here is the first part of the test:

Heat_2.jpg

Noise starts to increase faster when sensor exceeds 20°C.

And second part is the zoomed in table of low temperatures:

Heat_3.jpg

Data was dark current included and no subtractions were made. Conclusion is that I'll aim to -5°C sensor temperature with D5100, practically same results Fred got.

Link to comment
Share on other sites

Looks like there's little difference between 0C and -5C.  I think I shall go for around freezing whereas I was aiming for -10C to -15C.  Aiming for 0C means a much smaller Peltier TEC is required :)

Link to comment
Share on other sites

Exactly what I been finding with the cooled cameras I'm using, there's nothing to gain going sub zero, the ideal is to maintain at around 5Cº to 0Cº, this not only requires a smaller TEC and heatsink but also much less power consumption and less likely to get frost on the sensor ;)

Cheers,

  • Like 1
Link to comment
Share on other sites

Funny thing about astrophotography in Finland is that it usually is quite cold during winter and it makes cooling quite a bit easier than for example in Portugal. You may have some other climate benefits there though ;)

Anyway, what I mean is that it's not uncommon to take exposures somewhere between -10 and -20°C. We all know what does to noise :) During my temperature tests I noticed that D5100 sensor stabilizes at ~12°C above ambient temperature during long session. This means that the sensor actually is already at optimal temp during cold weather. But what fun would it make to let nature do the cooling?! Nothing, so I'm working on TEC D5100 monomod :)

  • Like 1
Link to comment
Share on other sites

Some unfortunate news for a change. For the last couple of months I’ve been trying to modify a full frame Nikon D600 sensor and now I have to admit that it cannot be done.

Issue is the resin between the cover glass and sensor frame. It seems that the heat needed to release the resin exceeds tolerance of the sensor by a great margin. I have used nearly ten hours with different heating methods but it doesn’t even give the normal color change. Also the resin is so hard that there is no way putting anything between the glass and the frame. Sensor frame on the other hand is so hard that you can only affect it with a Dremel cutting disk and that shatters the glass in process. Even though there is a way to Dremel a passage under the glass and pry glass from there, it doesn’t help.

There may be a mechanical way to remove the glass, high precision cnc or something, but it cannot be done by the methods described in this thread.

 

Brave ideas are welcome though..

Link to comment
Share on other sites

There may be a mechanical way to remove the glass, high precision cnc or something, but it cannot be done by the methods described in this thread.

 

Yes, CNC with high speed spindle and diamond coated tool should cut through the glass easily. I may try that some time...

Link to comment
Share on other sites

Hi!

Now I have a B&W EOS 500D, I can use OIII, Ha, SII filters. But the filters I have are 1"1/4 and I can't afford buying 2" filters.

I called 3D-clever guys to design a 1"1/4 filter holder to be put inside a Canon DSLR like an EOS Clip filter. Below is the 3D Printer view of 4 stacked filters holders :

17224-1420527751.jpg

They arrived to a design that works quite well, as shown :

10293-1421261824.jpg

Now, I just have to wait for the weather to calm down and let the sky going on !

  • Like 4
Link to comment
Share on other sites

Salut Fred,

That look like a cool inexpensive gadget for sure, I also did something of teh kinf do use my existing 1.25" Ha filter when I debayered my 350D, it works but you'll have quite some vigneting, sure you can crop out the bad peripheral areas and end up with a nice clean image :)

Abou the filter, when I can buy them go for the 36mm unmounted versions, they are so much cheaper then the 2"! Cheack the filter drawer system by TS that replaces the T-adaptor and allows you to use the full range of 36mm filter, that what we all use around here :) Also this filter drawer allows you to use all accesories like raducers, coma correctors, etc, as it maintains the same distance as a convencional T-adaptor, very clever stuff indeed!

Cheers,

Luís

  • Like 1
Link to comment
Share on other sites

See here: http://www.teleskop-express.de/shop/product_info.php/info/p5741_Filter-Quick-Changer-incl--1x-filter-drawer-for-unmounted-36mm-filters---low-profi.html

Baader 36mm 7 Nm Ha filter: http://www.teleskop-express.de/shop/product_info.php/info/p4367_Baader-H-Alpha---7nm---36mm-w-o-cell.html

If you can also build an adaptor of some sort you still can use the filter only for now while you can't afford the drawer system, I did mine with a foam ring and place it inside the T-adaptor ;)

Link to comment
Share on other sites

Yes Luis, this is it !

But still at 120 EUR+port, which is quite significant. The EOS Clip filter holder costs only a few EUR... this is a "solution for the poors ;)"

I've got another question, but it should better be adressed in a separate topic : how do you set up Deep Sky Stacker to work with Black & White RAW files made from a CFA removed DSLR ? There are 4 options :

13521235017825623.jpg

and none of them is generating a pure B&W image.

Tell me if a new topic should adress this subject.

Fred

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.