30 replies to this topic

[PeterLaubscher]

Has anyone been brave enough to try rmoving the Bayer matrix layer from the sensor in their one-shot colour camera - be that a DSLR or a dedicated OSC astro camera?

 

What were the results?

 

Asking for a friend...  

[pyrasanth]

Brave to even think this- never mind the actual task!

[Jim Waters]

Don't attempt the mod.  Get a real CMOS mono camera.

[triplemon]

There are several threads on this here on CN. The search function or even google will bring them up readily.

 

A  good search phrase is "monochrome mod".

Edited by triplemon, 02 June 2025 - 07:10 PM.

[whwang]

Don't attempt the mod.  Get a real CMOS mono camera.

 

second to this.  get one from ZWO or QHY before the tariff hits.

[bignerdguy]

If this is a DSLR then there is no need to mod the camera for this, simply set the camera to shoot in mono.  You will lose a bit of resolution when you do this since it may not use all of the sub/color pixels to take the image but it will do it and wont look terrible. i have taken photos in mono on my canon camera on occasion and it was ok.

[PeterLaubscher]

Many thanks, everyone for your advice and comments. I will leave well alone and will start to save my pennies - it is going to take a l-o-n-g time  

[whwang]

If this is a DSLR then there is no need to mod the camera for this, simply set the camera to shoot in mono.  You will lose a bit of resolution when you do this since it may not use all of the sub/color pixels to take the image but it will do it and wont look terrible. i have taken photos in mono on my canon camera on occasion and it was ok.

 

I suggested the OP not to do this because of the difficulty/risk of the mod, compared to the cost of alternative solutions.  But if you really want to compare an OSC and its mono mod, then the mono mod does not just offer a resolution advantage. The main advantage is sensitivity, and the difference can be quite substantial according to examples we saw on this forum in the past decade or so.  Indeed, the resolution advantage can be rather easily offset by doing Bayer drizzle in post processing, which should offer you essentially the same resolution as a mono one.  So I never view resolution as a disadvantage of an OSC.

[bignerdguy]

I suggested the OP not to do this because of the difficulty/risk of the mod, compared to the cost of alternative solutions.  But if you really want to compare an OSC and its mono mod, then the mono mod does not just offer a resolution advantage. The main advantage is sensitivity, and the difference can be quite substantial according to examples we saw on this forum in the past decade or so.  Indeed, the resolution advantage can be rather easily offset by doing Bayer drizzle in post processing, which should offer you essentially the same resolution as a mono one.  So I never view resolution as a disadvantage of an OSC.

True but you're talking about an OSC dedicated camera. I was referring to a DSLR which there is no need to mod to get a mono image from.  Most modern DSLR's come with a mono mode built in, its a simple menu selection.

[triplemon]

Well, the idea of modding a bayer matrix sensor is to increase the resolution when used with narrow band filters, like Ha. The green and red pixels don't do any good otherwise. Or for near UV or IR imaging. For visual luminance its not so much of an advantage - as the increased sensitivity for each pixel (roughly 3-4x) is mostly wasted by the loss of the micro lense array.

 

I also second that if you can buy a monochrome sensor instead, its almost always preferable. But if you're deep into modding you likely have some supply of near free old DSLR, then its a great thing to try.

Edited by triplemon, 04 June 2025 - 02:46 PM.

[vidrazor]

You can pick up the overpriced Pentax mono camera if you want a mono only camera. However ASCOM and INDI drivers for it are apparently hit and miss.

Much better to simply pick up a mono astro camera.

[whwang]

Both you guys have some fundamentally wrong understanding about mono mod.  But since we have the same conclusion (don't do it), who cares?

[erictheastrojunkie]

True but you're talking about an OSC dedicated camera. I was referring to a DSLR which there is no need to mod to get a mono image from.  Most modern DSLR's come with a mono mode built in, its a simple menu selection.

Lol, setting a DSLR to "monochrome" doesn't take the bayer array away, light still has to be passed through and filtered by each bayered pixel which reduces efficiency. That's why mono cameras exist, if you could simply just set a OSC to "mono" and get the same performance as a true monochrome camera then those bayer-less cameras would not exist in the first place. OSC astro cameras actually shoot in "mono" and the images have to be debayered in post processing. 

 

I'm going to take a slightly different path than the "don't do it" crowd, I'd say if you have a DSLR sitting around that's not being used otherwise...go for it, what do you have to lose? You'll probably learn a little more about cameras, sensors, the technology, assembly, etc and maybe it'll scratch some itch. If you're thinking about actually buying a DSLR/MILC to try and perform this mod on to be used for astro imaging....don't do it, just buy a mono astro camera. 

Edited by erictheastrojunkie, 04 June 2025 - 10:15 PM.

[AndyMcKay]

True but you're talking about an OSC dedicated camera. I was referring to a DSLR which there is no need to mod to get a mono image from.  Most modern DSLR's come with a mono mode built in, its a simple menu selection.

The sensor is still a standard color sensor. It still has the bayer matrix and each pixel is still only sensitive to red, green, or blue. 

 

A mono camera isn't as simple as taking the resulting color image and discarding the color info. 

[erictheastrojunkie]

This thread has reminded me a recent experience at work, I do biomedical research and as part of my job we use fairly basic microscope setups for imaging of brightfield and fluorescent labeled cells. We decided to purchase a new microscope for better resolution and sensitivity than the one we had previously, got some bids, and ended up deciding on a Nikon microscope which is "decent", it comes with all the fluorescent filters and....a 24mp full frame camera for the imaging. This....caught my eye, to say the least. Microscopy imaging is not unlike telescope based imaging, just the "opposite" in a sense. All the same limitations apply, you have optical objectives, you have field of view, you have resolving limits, you have illumination circles and reducers, etc etc. The very first thing I wondered is whether or not the microscope could even fully illuminate a full frame sensor, after digging way too deep into their technical specs and diagrams for the various scopes....not even close. In fact with their optical layout and their 1x "flattener" they are barely fully illuminating an ASI533 sized sensor. How much are they charging for the camera? A cool $10k, L.O.L. I about fell out of my seat. So I started doing more digging, what's the sensor? They advertise a monochrome mode for use with the fluorescence imaging and as far as I know there's no commercially available 24mp full frame sensors that are actually monochrome. Of course it's not an actual monochrome sensor, it's just the same old sensor from their Nikon D610 series cameras, they just use the same nonsensical trick of stripping away the color information in mono mode while tricking their customers with verbiage. Of course mono mode is even more limited, it's 10-bit while in color mode the image output is 12-bit, plus resolution is limited to only ~9mp total since even with the flattener attached they can only full illuminate roughly a third of the total sensor area.

 

In the end it didn't matter because the true power of the system for our purposes was mostly in their software, but that didn't stop me from laying down some thick questions for their rep when he came to do a demo and ultimately install the scope once we actually purchased it. Of course the rep was absolutely mystified that I even had those questions and absolutely could not answer any of them coherently, just more hand waiving nonsense. Once of these days I am actually going to bring down my mini-pc and hook up my ASI533mm to compare with Nikon camera. 

[whwang]

Once of these days I am actually going to bring down my mini-pc and hook up my ASI533mm to compare with Nikon camera. 

But you say the true power is the software.  Does the software accept images from other cameras?

[erictheastrojunkie]

But you say the true power is the software.  Does the software accept images from other cameras?

No, unfortunately it's limited to a couple Nikon cameras, but I may be able to convert the file format from xisf to something the software can use in it's editor mode. It would mostly be to scratch a curiosity itch of whether the image quality would be better with the ZWO camera vs the larger pixels of the Nikon. Plus some of the fluorescent imaging we do is on low surface brightness cells with dim protein targets which means sensitivity needs to be very high. 

[bignerdguy]

Lol, setting a DSLR to "monochrome" doesn't take the bayer array away, light still has to be passed through and filtered by each bayered pixel which reduces efficiency. That's why mono cameras exist, if you could simply just set a OSC to "mono" and get the same performance as a true monochrome camera then those bayer-less cameras would not exist in the first place. OSC astro cameras actually shoot in "mono" and the images have to be debayered in post processing. 

 

I'm going to take a slightly different path than the "don't do it" crowd, I'd say if you have a DSLR sitting around that's not being used otherwise...go for it, what do you have to lose? You'll probably learn a little more about cameras, sensors, the technology, assembly, etc and maybe it'll scratch some itch. If you're thinking about actually buying a DSLR/MILC to try and perform this mod on to be used for astro imaging....don't do it, just buy a mono astro camera. 

 

The sensor is still a standard color sensor. It still has the bayer matrix and each pixel is still only sensitive to red, green, or blue. 

 

A mono camera isn't as simple as taking the resulting color image and discarding the color info. 

Yes to both as this is what i said basically in my first response.  It will reduce the resolution since you lose some of the pixels as a color sensor has a 4 color Bayer pattern.  What i am confused about is how you can strip the Bayer pattern off the sensor.  the sensor has hardware that detects certain frequencies per sub pixel.  this is not something you can remove unless you physically replace the sensor.  So i don't see how you can strip the Bayer matrix from a OSC sensor.  Its not really a software fix since the pattern is baked into the hardware.  Yes setting mono mode in software does the software thing by only accepting image data from like 1 subpixel, not all of them. It does not strip the Bayer matrix since that would require a new sensor.

[erictheastrojunkie]

Yes to both as this is what i said basically in my first response.  It will reduce the resolution since you lose some of the pixels as a color sensor has a 4 color Bayer pattern.  What i am confused about is how you can strip the Bayer pattern off the sensor.  the sensor has hardware that detects certain frequencies per sub pixel.  this is not something you can remove unless you physically replace the sensor.  So i don't see how you can strip the Bayer matrix from a OSC sensor.  Its not really a software fix since the pattern is baked into the hardware.  Yes setting mono mode in software does the software thing by only accepting image data from like 1 subpixel, not all of them. It does not strip the Bayer matrix since that would require a new sensor.

You can literally scrape it off utilizing tools and chemicals:

https://www.cloudyni...n-sony-sensors/

[michael8554]

You do not "lose some of the pixels as a color sensor has a 4 color Bayer pattern."

 

You are merely removing the Bayer Matrix to expose all of the pixels to the whole spectrum.

 

There is an old 106 page post on the UK "Stargazers Lounge" forum:

 

https://stargazerslo...s-bayer-matrix/

[FrankieT]

(...)

 

Yes setting mono mode in software does the software thing by only accepting image data from like 1 subpixel, not all of them.

That's also not how "mono" mode works with a DSLR either. Basically, to produce an image, including grayscale, the camera  debayers the raw sensor data to interpolate full colour information for each pixel. In grayscale mode, the camera converts the colour values to luminance values (often using a weighted average, like 0.2126R + 0.7152G + 0.0722B) to create a monochrome image. Of course, actual implementations will differ, but in cameras with Bayer filters, the monochrome image typically comprises data from all subpixels (red, green, and blue) and retains the full resolution of the sensor.

Edited by FrankieT, 08 June 2025 - 01:51 PM.

[AndyMcKay]

So i don't see how you can strip the Bayer matrix from a OSC sensor.

As noted, you physically remove the substrate the Bayer matrix filters are on. The sensor is a OSC camera IS a mono filter. It is filtering each individual pixel to a specific wavelength range, and then applying the algorithm to de-Bayer, which just infers "true" colors based on the values from a pixel and its neighbors. A bit oversimplified, but that's the principle.

 

Now if this actually going through the process of actually removing the filter matrix is worth it or not is a completely different discussion. 

[xonefs]

The problem with mono conversions is you also strip off all the microlenses. So much of the efficiency that is gained from a mono sensor is offset by efficiency lost by having no microlenses. It is still a net gain, but not as dramatic as having a proper factory monochrome sensor with microlenses intact. 

Edited by xonefs, 08 June 2025 - 02:41 PM.

[whwang]

The problem with mono conversions is you also strip off all the microlenses. So much of the efficiency that is gained from a mono sensor is offset by efficiency lost by having no microlenses. It is still a net gain, but not as dramatic as having a proper factory monochrome sensor with microlenses intact. 

 

This is absolutely correct. The removal of the microlenses partially cancels the increase in photon throughput going from any of the R,G,B filters to an effectively L filter.  As Jason said, it's still a net gain, but not worthwhile at all.  The real huge gain appears when one images with narrow-band filters, where one goes from only having 1/4 or 1/2 active pixels to 100% active pixels.  But still, it's not worthwhile in my opinion.  It's a fun experiment.  Do it only when you have a camera that you will never use otherwise.

[NuovaApe]

The camera firmware wont know its colourful RGB sunglasses have been stripped off, so will still return its type as "colour" to any USB software and still do binning based on R/G/B photosites instead of immediate mono neighbours.

 

For 2x2 binning, it'll R+R, G+G, B+B from these:

 

RG RG
GB GB

 

Above is simplified. Better description + pretty pictures here:
https://skyinspector...xplainer/#Bin2x

 

For a fun technical challenge I'd say go for it!

 

But you'll get lower res in-camera binning compared to true mono, have to select "ignore Bayer" in all apps, and potentially (almost certainly) ruin a perfectly good camera you could have sold and put the money towards a mono cam.

 

Best, Ed.