Brain icon doesn't show - I have a valid license

Anything that doesn't fit into any of the other forums
ramdom
Posts: 41
Joined: Wed Jul 05, 2017 4:03 am

Re: Brain icon doesn't show - I have a valid license

#11

Post by ramdom »

Also not sure if it matters but the Windows I am running is Windows 10 (educational). Not NT 6.x whatever... just wanted to make sure these kinds of things weren't interfering with the proper functioning of the camera.

If you recall the white line in the middle error I talked about when starting in raw16 mode - I have the log of that also since it just happened when I was testing the histogram spiking and then I have to change something like exposure, gain, go back to raw24 and then raw16 to make that white line to go away.

--Ram
User avatar
admin
Site Admin
Posts: 13173
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Brain icon doesn't show - I have a valid license

#12

Post by admin »

Hi,

Windows 10 education edition shouldn't be a problem at all.

I think what concerns me is the fancy tricks you are having to go through to get the camera into raw 16 mode – what I think is problem happening is that the camera isn't really going into that mode properly for some reason, but it works enough for SharpCap to get images. Until we work out the reason behind that I don't think we will get valid results from the sensor analysis or from any imaging that you do to be honest.

One thing that might help is to put the camera into still mode rather than live video mode – this should be a control on the right that allows you to do this. Maybe that would help give correct image data stop

Cheers, Robin
ramdom
Posts: 41
Joined: Wed Jul 05, 2017 4:03 am

Re: Brain icon doesn't show - I have a valid license

#13

Post by ramdom »

You mean the file selection mode - FITS instead of SER? Normally I do have it in FITS when I image but it always resets the default to SER and usually when I image I explicitly set it to FITS and go from there. I left it under SER thinking it won't matter for the sensor analysis. I will check out the histogram behavior under FITS also.

But if what you're saying is correct, then the histograms of the single frames I've produced with this camera must also look spiky right? But they do not. Here's an image of a calibrated and debayered frame in PI (no other changes). It's nice and smooth, as is the mono FITS image histogram.
foo.jpg
foo.jpg (42.92 KiB) Viewed 2554 times
bar.jpg
bar.jpg (36.31 KiB) Viewed 2554 times
--Ram

PS: Even when I stretch it, it looks a bit spiky but still continuous but unstretched, it doesn't look spiky at all and I realised in imaging I've been taught to not stretch during capturing.
ramdom
Posts: 41
Joined: Wed Jul 05, 2017 4:03 am

Re: Brain icon doesn't show - I have a valid license

#14

Post by ramdom »

Also, a couple of images I've produced with this camera thus far (there are a few others here: https://www.astrobin.com/users/ramdom/) :

https://www.astrobin.com/418204/
https://www.astrobin.com/414567

These are the final stacked images so obviously I could be stacking 8 bit data to produce a final 16 bit image but they don't look any different from the ones produced from the other mono camera.

As I showed, the single frame histograms look even better. When I said things looked spiky, something like the first image above is what I meant.

I am waiting for the sun to go down so I can check the histogram on the camera with the FITS setting and also do it outside in the dark.

--Ram
ramdom
Posts: 41
Joined: Wed Jul 05, 2017 4:03 am

Re: Brain icon doesn't show - I have a valid license

#15

Post by ramdom »

Okay, I figured it out - there is a "force still mode" on/off option. So when I set that on, it causes all kinds of problems. The histogram appears static (regardless of whether I take exposures or not). Then it repeatedly crashes Sharpcap (with a memory error).

Nonetheless, once I reset that to off, the histogram appears is quite different - it is a combination of spiky and continuous but not like what it was before (the bottom part is solid whereas before it was only a bunch of spikes clumped together). I reran the sensor analysis but got the same result as before...

Also, under force stlll mode set to "on" I couldn't run the sensor analysis...

I mentioned the Windows 10 since the log says Windows NT 6.x...

Let me know if you want the Logs for the session where setting the force still mode to on caused problems.

--Ram
Last edited by ramdom on Thu Aug 15, 2019 7:45 am, edited 1 time in total.
User avatar
admin
Site Admin
Posts: 13173
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Brain icon doesn't show - I have a valid license

#16

Post by admin »

Hi,

This is all getting very confusing :-(

Let's try to sort out whether the histogram is working correctly or not first – please can you capture a flat frame (or something like a flat frame – basically expose for 50% histogram peak) at the highest depth into fits format and share that file with me please.

Cheers, Robin
ramdom
Posts: 41
Joined: Wed Jul 05, 2017 4:03 am

Re: Brain icon doesn't show - I have a valid license

#17

Post by ramdom »

Okay, here's a single frame:

http://www.ram.org/images/space/downloa ... 00001.fits

and the data for the single frame (histogram, camera settings) is also here:

http://www.ram.org/images/space/downloads/

--Ram
User avatar
admin
Site Admin
Posts: 13173
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Brain icon doesn't show - I have a valid license

#18

Post by admin »

Hi,

So there are lots of gaps in the histogram of that image (and in the histogram CSV file too). Those are definitely going to be a problem for the sensor analysis and also going to be losing you data your imaging. The underlying cause of those gaps clear yet though.

A couple of things I can suggest is putting white balance controls and the gamma back to their default values. As far as I can tell the defaults for the white balance 64 and the default for the gamma should be 1.0 - that certain seems to give neutral results for my 178C camera. Of course other cameras may have different defaults :-(

What does the histogram look like in SharpCap? When I show the histogram for my 178C you can see that I get smooth histogram curves in the 16 bit mode. This camera has a 14 bit sensor just like yours I believe. Note that I have enabled the logarithmic vertical scale in the histogram stop.
Capture.PNG
Capture.PNG (21.99 KiB) Viewed 2525 times
Given the data your fits file and in the histogram CSV I would imagine that your SharpCap histogram will be really spiky lots of gaps in it. Watching this while you adjust various controls might help you determine if a rogue setting on one controller and another is causing the problem with the gaps in the data. If you can get a smooth histogram like the one above then I think everything will work a lot better.

Hope this helps, Robin
ramdom
Posts: 41
Joined: Wed Jul 05, 2017 4:03 am

Re: Brain icon doesn't show - I have a valid license

#19

Post by ramdom »

Thanks for posting your image. Yes, none of my images look like that exactly - they are variations and some less spiky than others, but they're all more spiky than not (but no gaps though - at the bottom there does appear a continuous line). The only exception is for the white line which does look similar. What my RGB histograms look like throughout is like how your corner of the white and blue lines look like.

I put the controls back to normal and it didn't change the spikiness and I tried a bunch of things but nothing really seemed to work (debayering off/on, etc.). Just made the bottom more solid. Here're the images:

RGB24, default settings: http://ram.org/images/space/downloads/t ... 00002.fits
RAW16, default settings: http://ram.org/images/space/downloads/t ... 00003.fits

In the TLD for each of those images (i.e., http://ram.org/images/space/downloads/t ... T17_52_05/) you can see the settings, etc. and also the histogram directly if you don't wish to download the FITS files. Even the RGB24/RAW8 seems to show the the spiky histogram.

For comparison purposes, I used EZCAP to take some shots. It's hard to see in PI the gaps in the histogram since everything is clumped together but I
can see that the images from EZCAP are different from those produced by Sharpcap. Those images are here (capture1.fit, etc.): http://ram.org/images/space/downloads/testing/ - I was just playing around with EZCAP to generate the other files - you just need one of those files to see the difference.

Unfortunately, another bug EZCAP introduces is that the axis of the images is set to 6088 x 6052 (which doesn't seem to be correct based on the images themselves) but the histogram seems to be correct here (smoother, more clumped together). So for now I should be capturing with EZCAP right? Sigh...

--Ram

PS: This is really a problem with the way the dynamic range is sliced up right? It doesn't change the range itself, or the values - it's how those values are separated. While I can see it can cause issues with getting ideal images it should be useable.
User avatar
admin
Site Admin
Posts: 13173
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Brain icon doesn't show - I have a valid license

#20

Post by admin »

Hi,

Here's a snapshot of what's the fits file that you captured in EZCap looks like when you open in fits Liberator.
Capture.PNG
Capture.PNG (86.75 KiB) Viewed 2511 times
You can see that the Bayer grid visible in the zoomed in portion of the image is even and that the histogram has three solid peaks in it (the peaks come from the different colour channels). The fact that their solid indicates that the histogram is not spiky.

Here's the same scrap shot of one of the 16 bit images you captured in SharpCap
capture2.png
capture2.png (90.18 KiB) Viewed 2511 times
Not only do we have a very spiky histogram, the Bayer pattern grid is not even (this could be down to the image that you used to capture this not being a flat frame I suppose).

Something very weird is happening to the image that is coming through SharpCap meaning that the pixel values are distorted to give the spiky histogram effect. You haven't enabled any of the image processing features in SharpCap like flat or dark correction or the row noise reduction, so the data saved file will be exactly what came out of the qhy SDK, so that must be kicking out the weird data for some reason.

If you can somehow get the SharpCap histogram as shown in fits Liberator to look like the one from the EZcap then you've got the problem sorted but exactly how I have no idea.

Cheers, Robin
Post Reply