Debayering a DSLR's Bayer matrix.

[pixueto]

I've managed to take some pictures with the scope tonight. Trouble is, DSS is giving me bizarre results. I converted the raw files to Tiff without demosicing using dcraw. I did the same for the darks, flats and bias. When I put the lot in DSS, I'm getting grainy pictures.

Luis, how did you stack your mono lights?

Thanks

Edited August 26, 2013 by pixueto

[Gina]

Can you please make a test photo, Gina? I wold like to see the flat result...

Sorry, but I killed this sensor getting the cover glass off, so no can do I shall certainly show images when I get to doing a working sensor

[Luis Campos]

I've managed to take some pictures with the scope tonight. Trouble is, DSS is giving me bizarre results. I converted the raw files to Tiff without demosicing using dcraw. I did the same for the darks, flats and bias. When I put the lot in DSS, I'm getting grainy pictures.

Luis, how did you stack your mono lights?

Thanks

Hi, great to hear you have a working sensor

I allways use IRIS for image stacking / callibration, this gives me best results, you can't do the "convert a CFA sequence" step, or else you will loose resolution due to interpolation, just load your light frames bias / dark and flatfield calibration, align by one or more stars and stack using you favorite method, I use #sigma median" with 2 iterations and "sigma coeficient 2".

If you need a hand with Iris please let me know.

Cheers,

[pixueto]

Hi, great to hear you have a working sensor

I allways use IRIS for image stacking / callibration, this gives me best results, you can't do the "convert a CFA sequence" step, or else you will loose resolution due to interpolation, just load your light frames bias / dark and flatfield calibration, align by one or more stars and stack using you favorite method, I use #sigma median" with 2 iterations and "sigma coeficient 2".

If you need a hand with Iris please let me know.

Cheers,

Thanks Luis.

What type of files do you use in IRIS? Do you use dcraw or other equivalent software to convert the raw file to another file, say fits without demosaicing prior to process them in IRIS?

Having a look at IRIS and I don't find it that obvious.

Thanks

[Luis Campos]

Thanks Luis.

What type of files do you use in IRIS? Do you use dcraw or other equivalent software to convert the raw file to another file, say fits without demosaicing prior to process them in IRIS?

Having a look at IRIS and I don't find it that obvious.

Thanks

Hi,

No, you just click and drag the CR2 files to the Iris window, go to "digital photo" and "decode raw files", a new window will show up, drag you CR2 files on to this window and name the files batch, say "img" for ex. and then click "CFA" and when done click..."done"

Iris will open your CR2 files and will show you a dark image, you need to adjust the levels slider just to get a correct visualization.

But before all this you must do the same to your "offset", "darks" and "flatfield" CR2 images, proceed the same way just clik and drag, it's all under "digital photo", after decoding your raw files just click "digital photo" and "make an offset" in the case of your offsets, and do the same for the others, at the end save the files as PIC and name accordingly (offset, dark and flat).

Latter when loading your light frames, Iris will automatically calibrate your image with the callibration frames you have input, load your lights, the same way (click and drag), then draw a rectangle in a starless part of your image and click "preprocessing", this will take a wile, when finished go to "processing" and "stellar registration" and choose "one star" registration or "global matching" if you want a very good registration, this option will take much longer but will offer a superior result.

When finished, click "processing" and "add a sequence" and choose "sigma median" stacking with 2 iterations and coefficient of 2, this works just fine for me, but try other values and see what you prefer

Well, this seems a lot now, but with practice it becomes so much easier, Iris is complicated for sure but it's the Holly Graal of image callibration / registration

Cheers,

Luís

[pixueto]

Hi Luis, I'm dragging the raw files to the box but it won't take them. Any idea about what the problem may be?

If I understood correctly, the lights are processed at the end; not at the start, right?

Edited August 26, 2013 by pixueto

[pixueto]

Hi Luis, managed to get the thing working. I have some questions, though. When I go to preprocessing, there are several options; I went for automatic preprocessing 1, I think.

Also for stellar registration, Iris is asking me to enter the name of the original file/files? Are those the light frames?

Obviously, I'm doing something wrong as the final image is completely white but the individual lights are OK -however, once saved and brought to Photoshop, it seems to be fine.

Thank you for your help

Edited August 27, 2013 by pixueto

[Gina]

I've bought a "not working" 1000D body for £70 with delivery expected tomorrow so I should soon have one working and one non-working 1000D to play with. I thought that was a reasonable price compared with around £40 plus carriage and import duty from the USA for a used sensor alone. If I succeed in producing a good result with the first 1000D sensor I can still see if I can get the non working one fixed or failing that sell the parts. I've started collecting the 1100D parts together ready to see about selling - there's a lot of them.

Edited August 27, 2013 by Gina

[Gina]

The non-working 1100D hasn't come yet - maybe tomorrow...

[Coco]

The non-working 1100D hasn't come yet - maybe tomorrow...

Good luck Gina...

[Gina]

Thank you That 1100D hasn't arrived yet. But the non-working 1000D has Of the ones I've bought in the last couple of weeks, the last one has arrived first! Funny that

In order :-

Canon EOS 1100D EF-S 18-55 111 kit (spares and repairs) - 16 Aug, 2013

Canon EOS 1000D Rebel XS 10.1 MP Digital SLR Camera with 18- 55 lens - 25 Aug, 2013

Canon EOS 1000D Digital SLR Camera Body, only Spares or Repairs - 26 Aug, 2013

[alistairsam]

good luck Gina. the glass for the 1000d is the easiest to remove. comes off with slight sideways pressure with a knife. turns white very easily.

cfa is a bit harder to remove with the wooden tool but not too hard.

I just got my spare 600d sensor and the heat desolder station.

I'll test the sensor in the camera first before I attempt debayering

anyone thought of using aluminium tape on the glass with the heat method to hold it in case it cracks?

I will post progress

Alistair

[Gina]

Here's the blurb about the sale

This sale is for the camera body only as shown above - there are no accessories (i.e battery / charger / cf card / BODY CAP ) included in the sale

This item is being sold as Faulty Spares / Repairs. We have tested this item and found the following fault The Camera will not power up and so describing this item as FAULTY / SPARES / REPAIRS. We have not tested any further once we have found this fault, so we cannot state with certainly if there are other faults.

Now I have options which differ from what I would have done if the working 1000D had come first. One option would be to wait for that, of course but that will try my patience

1. Wait for the working 1000D
   
2. Strip this camera down to the sensor and work on that
   
3. Partly strip the camera down, connect some power (no battery supplied) and see if I can get it working or get some idea of what the fault is. I already know that it won't power up so can start there.
   

I think I favour no.3.

[Luis Campos]

Hi Luis, managed to get the thing working. I have some questions, though. When I go to preprocessing, there are several options; I went for automatic preprocessing 1, I think.

Also for stellar registration, Iris is asking me to enter the name of the original file/files? Are those the light frames?

Obviously, I'm doing something wrong as the final image is completely white but the individual lights are OK -however, once saved and brought to Photoshop, it seems to be fine.

Thank you for your help

Hi,

The "preprocessing" I use is under "digital photo" and then "preprocesing", I guess it's different from that option you used

Please try that and let me know how it goes.

On "stellar registration" it's about your light frames, you must name them with an entry name for ex. "A" and an exit name for ex. "B", or 1 and 2 and so on if you prefer

Cheers,

Cheers,

[Gina]

good luck Gina. the glass for the 1000d is the easiest to remove. comes off with slight sideways pressure with a knife. turns white very easily.

cfa is a bit harder to remove with the wooden tool but not too hard.

Thank you Alistair

I just got my spare 600d sensor and the heat desolder station.

I'll test the sensor in the camera first before I attempt debayering

Good luck to you too then

anyone thought of using aluminium tape on the glass with the heat method to hold it in case it cracks?

I will post progress

Alistair

Is that adhesive aluminium tape? Don't think I know of that I'll look forward to reading your progress

[Gina]

I've started stripping down the 1000D. I've taken the power board out ready to solder on some wires for an external power supply. Only problem is that there's no indiction of which way round the battery connections are and nothing on the power board to use as reference for use of an ohmmeter to check. I could assume that the battery terminals are the same as the 1100D but they might not be and I culd do damage by applying volts the wrong way.

Could someone with a 1000D battery please tell me what the connections are? - they're usually marked on the battery as - T +

[JamesF]

Does the large image on this page help, GIna?

http://www.ebay.co.uk/itm/2-Lot-Battery-Canon-LP-E5-EOS-450D-500D-1000D-Rebel-T1i-Xsi-Kiss-X3-X2-UK-/281004305104

James

[JamesF]

As they're apparently the same as for the 450D I've just checked mine and the terminals are in the same orientation as the picture linked above. That is, contacts uppermost and on the left the terminals read - T + from top to bottom.

James

[Gina]

Thank you James - yes, very helpful - just what I wanted to know

[Gina]

Connected camera to bench PSU and it's drawing current but the current doesn't increase when the ON/OFF switch is set to ON so I've removed the contacts and cleaned them with IPA. Back together but still no change with ON/OFF switch. The contacts operated by the battery cover DO work. If the circuitry is silimar to the 1100D (and with multi wire connectors there's no way of knowing) the battery cover contacts operate power control circuitry on the main board. I'll do more testing...

[pixueto]

Hi Gina, as Alistair has said, the glass covering the sensor in the 1000D is the easiest one to remove. Proceed with caution, though. You should be able to remove it intact.

As for the removal of the CFA, you may want to wait until I finish processing the subs with my debayered 1000D (Luis is giving me a hand doing this with IRIS as DSS is no good for that). I took those the other day with an astronomik light pollution CLS clip filter and the SW200P. My results may help you avoid some of the mistakes I've made. You'll see in my pictures that the scratched areas do come up but the uneven illumination should be sorted with the flats which the computer seems to be able to process.

I used the wooden tool and then fine car windscreen polish and isopropanol to remove and clean the CFA leftovers. I made some scratches and left the sensor a bit uneven in the middle central band but there are intact areas so there must be a way of producing an entire clean debayered sensor. Tonight, if it's clear, I'll take some subs with the 7nm Baader H-alpha CCD filter which just arrived today.

[Gina]

Hi Gina, as Alistair has said, the glass covering the sensor in the 1000D is the easiest one to remove. Proceed with caution, though. You should be able to remove it intact.

Yes, good

As for the removal of the CFA, you may want to wait until I finish processing the subs with my debayered 1000D (Luis is giving me a hand doing this with IRIS as DSS is no good for that). I took those the other day with an astronomik light pollution CLS clip filter and the SW200P. My results may help you avoid some of the mistakes I've made. You'll see in my pictures that the scratched areas do come up but the uneven illumination should be sorted with the flats which the computer seems to be able to process.

I used the wooden tool and then fine car windscreen polish and isopropanol to remove and clean the CFA leftovers. I made some scratches and left the sensor a bit uneven in the middle central band but there are intact areas so there must be a way of producing an entire clean debayered sensor. Tonight, if it's clear, I'll take some subs with the 7nm Baader H-alpha CCD filter which just arrived today.

Yes, I'll do that - I'm in no great hurry. I might remove the cover glass and apply epoxy resin to the sensor when I get it out but I'll leave removing the CFA until I have a working camera to test it in.

[Gina]

As to fixing the 1000D, I've taken the top apart to see what connects to where. The two pairs of wires from the power board connect to the flash unit and have nothing to do with the power on circuit. The ON/OFF switch, function dial and push buttons all connect through a ribbon cable that goes to the main board. I have visually checked the ribbon cable, contacts and connections and can see nothing wrong. Things are pointing to a main board fault or possibly the power board. The usual way of diagnosing the fault would be to swap major components with working ones but I can't do that until I get the working 1000D, of course.

[Gina]

I've removed the sensor and now removing the cover glass. I carefully worked around the glue joint with the fine tip of a craft knife. Gradually I was able to produce a white area and then gradually increase this until it went right the way through. At that point I thought further working carefully around applying gentle pressure with the knife it would eventually lift off.

OH NO!! Not on your life - it seemed to be as secure as ever! I gradually applied more pressure until first of all, a corner broke off. Then with a "ping" about a quarter of the glass shot off. I tried again with a bit less force. Again, "ping" - another piece sprang off. This is where I'm at now. I think I'll stick some sticky foam on before tackling the rest. No glass has so far got on the sensor itself.

[Gina]

It's off - well near enough. With the pad stuck on I was able to break off pieces of glass while all the bits were held clear of sensor and all those fine wires. When I thought it was all parted from the frame I turned the assembly over and the pad plus glass fell off. All except one little bit still attached to the frame. Since the inner edge of this is clear of the imaging area, I'm leaving it where it is.

That's it for today - I've put it carefully into a sealed plastic box and put that out of the way where nothing will touch it - until tomorrow.

Edited August 28, 2013 by Gina