Remove Bayer matrix from DSLR images prior to autofocus run


#1

SG’s auto focus routine works fantastic for me with my ED 100 and Canon 60Da, but fails completely with my shiny new Edge HD 9.25. The long focal length of the Edge HD (2350mm) makes the Bayer matrix much more visible, and you have determined that this interferes with the autofocus routine. My request would be that you guys somehow work your software magic and find a way to remove the Bayer matrix from DSLR images before the auto focus runs. The thought of having to focus my new Edge HD by hand after having used SG’s wonderful auto focus is too depressing to contemplate…

Thanks guys!

Dean


#2

I don’t think there is a way to do this. The Bayer matrix is physically in the chip.


#3

By remove, do you mean debayer the image? (That’s certainly possible but uncertain if it would help.)

I haven’t had too many issues using a Canon T4i @ 2800mm with AF – the primary issues are more related to the SCT, not the camera – the only exception being narrowband filters.


#4

You wouldn’t need to debayer, just have software binning as an option for OSCs and focus. Just bin 2x2, 3x3, 4x4 without worrying about the matrix. It’s for focus, the worst that could happen is what you are now getting for OSC focusing with obstructed scopes, and it may improve things. I get reasonable autofocus performance with my SCT and a mono camera at 4x4 binning. I haven’t tried my T3 with the SCT, but if I did, I would probably run into the same issues as Dean. May be worth a go.

Frank Z…


#5

Nebulosity does on the fly 2x2 binning. It could just be as simple as summing the 4 pixels.
…Keith


#6

If anyone is interested in trying, The 2.4.1.4 beta provides software binning for DSLR cameras during auto focus (you will need to turn this on in the Canon or Nikon settings dialog).


#7

Ken your customer support is amazing. I’m freezing at the scope right now waiting for clouds to pass, but will try the auto focus with binning with my Canon 60Da ASAP.

update: Ken I installed the new version and enabled it in the Canon settings. I see that in equipment manager that the boxes for setting the binning are still greyed out. I assume you have a preset level of binning set for DSLRs when this new feature is enabled, and it is not user selectable?

I will try it using binning with the ED 100 refractor ASAP, but I had to send the Micro Touch Crayford for my Edge HD back to Starizona, as it is not compatible with the Lepus focal reducer made for the Edge 9.25. It may be a few weeks before I get the Micro Touch for the mirror adjustment of the Edge HD and can try it at a long focal length. Thanks so much for the blinding speed addressing this problem. Here’s hoping it works!

Dean


#8

This is true. If this yields better results we will consider what it will look like in a final release. Right now we are shooting for any kind of “V” and if this helps, we may provide some options to better tune this. Right now it is set at what I believe to be ideal fro the 60Da (these changes were made with the data you sent). The AFPacks, with these changes in place, while they did not make a “V”, at least started to… I don’t believe there was enough data inside of focus.

Also, just as a note, I do not get emails for updates to existing posts. You are better off replying in a new post. Just saw your update by accident.


#9

Okay, looking forward to trying it out. I’m on holidays all week so of course it is forecast to do nothing but rain. I’ll let you know how it goes ASAP.

Dean


#10

Hi Ken,

I got a chance between clouds to try autofocus with binning and my Canon 60Da using the refractor last night. It did seem to work well the few chances that I had, and gave me a good v - curve. I should have the Micro Touch for the Edge SCT in a few days and will let you know how it goes using that telescope. One question about autofocus runs using the SCT…should I try to set my step size small enough that I stay out of donut land at the furthest out of focus point of the run?

Dean


#11

Hi Ken,

Please disregard the above question about donut land. I just watched your show on the Astro Imaging Channel, and Jared answered my question near the end of the program. I will keep the step size small enough to stay out of donut land. I do have one more question though. I am working on setting up backlash compensation for the Edge SCT. I know that finding the compensation step size is going to take some experimentation, but what about compensation direction? Should I set it to in or out for autofocus runs in SGP?

Dean


#12

That’s good news. If this works out, the final release will probably be a dropdown that allows you select the level of binning 1x1 - 4x4 (as opposed to a simple on / off option).


#13

Hi Ken,

I finally got a chance to try the autofocus routine with DSLR binning with the Edge 9.25 tonight, and I am happy to say it works! The v - curves aren’t quite as symmetrical and pretty as with the ED100 refractor, but they seem to be good enough. I settled on a step size of 100. Any bigger steps than this give me donuts at the farthest out of focus frame or two at the very end of the autofocus run. Smaller steps give increasingly poorer results. I ran autofocus repeatedly after getting the proper step size, and it was pretty consistent. I then checked it with a Bahtinov mask and it was in absolutely perfect focus. Thanks very much for working on the binning of DSLR’s. SGP just gets better and better. Let me know if you want my autofocus logs if they will help you in any way.

Dean


#14

Hi Dean

I have also a C925 Edge and a canon DSLR. Did you use the focus routine at f10 or at f2.3 with hyperstar?

Could you post an image of the auto focus settings used in SGP?

Thanks in advance,

Alfredo


#15

Thanks Dean,

I guess I will make this a bit more user friendly for a formal release in 2.4.1.


#16

Hello Alfredo,
My auto focus runs have been at f10 only. Here is a screenshot of my settings. Note the low sample size. I also use a backlash compensation setting of 100 steps. You can see on the screen the Bahtinov mask confirmation of the focus.

Dean


#17

Thank you Dean

I use the microtouch/feathertouch focuser designed for this scope. Is it the same one you have?

Regards

Alfredo


#18

Hi Alfredo,
Yes I use the same focuser and motor. Let me know how the settings work out for you.

Dean


#19

Thanks for the info Dean.

I’ll have to wait until the weather here allows to do the test.

Regards

Alfredo


www.mainsequencesoftware.com