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

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

#1

Post by ramdom »

So according to my email, my license expires on Sep 6, 2019 (which includes a two month grace). For the first time today, I discovered the new histogram with "brain" feature in the "learn more" section but there is no brain icon. Do I need to renew ahead of time since I'm on the 2 month renewal bonus period?

(I normally use Sharpcap just for capturing but I'm having trouble with balancing stuff on a new QHY247C I got which I'll post about in the QHY specific subforum).

Thanks.

--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

#2

Post by admin »

Hi,

A couple of things that come to mind here

1.you need to be in raw mode for the brain to show up as in RGB mode or mono mode on colour camera the raw data from the pixels has already been lost to the colour processing algorithm

2.you may need to run a sensor analysis on your camera so that the brain has sensor data to work with.

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

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

#3

Post by ramdom »

I am doing the sensor analysis on the 247C now - does the temperature of the sensor matter?

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

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

#4

Post by ramdom »

Either I did something wrong, or this camera is even more sensitive than I thought. I measured the sky quality (1.2 e/px) and then it recommended I use a gain of 0 (and an exposure length of 47s) for the sensor data I had captured earlier, with a black level of 12. Can that be right?

This was my sensor analysis data (only did the RAW16 measurement), which didn't seem to go over 500 gain.

And I've been using a gain of 1000 and an exposure length of 2 minutes for the most faint of my targets with this camera.

Also, when I use the mono 163M camera to do this (which thankfully already has the brain icon), do I need to measure the sky quality with specific narrowband filters or does it not matter? Thanks.

--Ram

Gain Value e/ADU Read Noise (e) Full Well (e) Relative Gain Rel. Gain (db) Dynamic Range (Stops)
0 0.0322887811234017 2.50298310080898 529.019389925813 1 0 7.72352824104906
50 0.0319519641093551 2.49402054464288 523.500979967673 1.01054135554528 0.0910818269442328 7.71357507808344
100 0.0316024271093626 2.48198958183458 517.774165759797 1.02171839560499 0.186624254775366 7.70468211406274
150 0.0309972045268741 2.4642655286772 507.858198968306 1.04166751861148 0.354582443108446 7.68712420541363
200 0.030612197401781 2.45604721205033 501.55024223078 1.05476848654854 0.463142916881399 7.67391212462138
250 0.0302046163698905 2.44221902381429 494.872434604286 1.06900153036172 0.579566538734509 7.66272028285141
300 0.0296955478065295 2.41848700171342 486.531855262179 1.08732734394285 0.727206194816346 7.65228567314886
350 0.0292740565037814 2.41092588906626 479.626141757955 1.10298281070923 0.851374885752151 7.63617918023441
400 0.0288536687230011 2.4009040664143 472.73850835765 1.11905288139883 0.977012196287488 7.62132081982809
450 0.0283267082335371 2.39255176769665 464.104787698271 1.13987057222462 1.13711083445347 7.59975662950654
500 0.0278958219126293 2.36675487104991 457.045146216518 1.15747731773351 1.27024978421546 7.59328257793781
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

#5

Post by admin »

Hi,

Those aren't good readings as they are showing a full well depth of only 500 electrons and ridiculously low value for the E/ADU figure. What bit depth did the analysis routine claim to find for your camera? An incorrect bit depth reading could cause both of those figures to be out.

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

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

#6

Post by ramdom »

Well, I redid it and my brain and cooler issues are perhaps converging. The bit depth it reported was 14 (for sure this time and I believe last time also). It stopped at 500 gain and this time I used a different light source (natural light). This time I had also turned on the logging so here

As far as the cooling issue, it may have cropped up in this round too but I am not sure. What happened is that the power wasn't connected tightly so I connected it back and the cooler started working properly (which is how it should be). But during the sensor analysis, it went from on to off. Don't know if this is due to the sensor analysis resetting things or if this is part of the bug, but the cooler definitely acted up during the sensor analysis.

Attached is the log to the session where the sensor analysis stopped at 500 gain. Link here also: http://www.ram.org/images/space/downloa ... 7-4140.log

--Ram

PS: Also I know that the instructions stated that the settings should be set to "neutral" before the sensor analysis but last time I had indeed done that and it reset everything anyway (I'm using a different white balance, etc.) to neutral so this time I didn't bother but it did appear to reset the white balance, etc. this time also. So just in case you see that in the logs...
Attachments
Log_2019-08-11T17_47_57-4140.log
(709.61 KiB) Downloaded 137 times
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

#7

Post by admin »

Hi,

Well I can see one thing which is that the range of the gain control on your camera is up to 4095, which I've never seen before on a qhy camera. Previously I've seen QHY cameras with gains that go up to about 60 and are measured in dB and gains that go up to about 600 and are measured in tenths of a dB and gains the go up to 255 and are linear and gains the go up to 255 and have no physical meaning whatsoever. Sigh… If only they could be consistent…

Anyway, so one part of the problem is that the code sees a maximum gain that is above 255 and interprets that in the way that I expected to see qhy gains go above that level – i.e. measured in tenths of a dB and limits the total gain range to measure to what it thinks is 50 dB, but isn't.

However the bigger problem is the incorrect measurement of the full well depth and the E/ADU value. I can see where that's coming from in the log.

Code: Select all

Info:	17:53:29.8480864 Thread:Image Process Thread#8 SharpCap.CameraAnalysis.SensorDepthMeasurementEngine.CalcBitDepthFromHistogram(Int32[] histogram) :: Histogram length is 65536
Info:	17:53:29.8530834 Thread:Image Process Thread#8 SharpCap.CameraAnalysis.SensorDepthMeasurementEngine.CalcBitDepthFromHistogram(Int32[] histogram) :: Significant Gaps are 256
Info:	17:53:29.8560815 Thread:Image Process Thread#8 SharpCap.CameraAnalysis.SensorDepthMeasurementEngine.CalcBitDepthFromHistogram(Int32[] histogram) :: Highest power of two in all gaps is 128
Warning:	17:53:29.8560815 Thread:Image Process Thread#8 SharpCap.CameraAnalysis.SensorDepthMeasurementEngine.OnFrameImpl(MathGrid frame, Int32 frameBitDepth, Int32 frameEffectiveBitDepth) :: Measured bit depth as 9, but camera reports as 14, using 14
What's interesting here is that although the maximum pixel value coming through is correct for a 16 bit value, the camera only seems to be sending images in a bit mode, just stretched to 16 bits. Basically this log code says that most of the gaps between histogram points that have significant counts are of size 256. I'd expect you would see this effect if you showed the normal SharpCap histogram – rather than a smooth curve you would expect to see a number of spikes if the data shown above is a true representation of what is going on. Then the question comes of why on earth we're not getting proper data from the camera – if you do see spikes in the histogram I would suggest working through all the different available controls to see if any have the effect of making the spikes go away and giving you a smooth histogram curve.

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

#8

Post by ramdom »

Yep, the range goes up to 4095 and according to the Web, the unity gain is around 2200. I've been trying to use other software to find it myself and one of the reasons the brain analysis interested me was that it would relate the gain scale they had to something more mathematical in terms of the e/ADU ratio.

I'm not sure what you mean by spikes - if you mean near complete separation at the bottom so it's a multimodal instead of unimodel distribution, then no, I don't think I've seen that often while imaging or even while doing this test. The single curve histogram itself has looked a bit spiky while doing this test and with my latest settings but in the past I've gotten smoother curves. The spikes started after I reduced the white balance to 32,32,32 and the offset to 12. Previously at offset of 16 and white balance of 48,48,48 it didn't seem to be as spiky and it was more to the right so I wanted to bring it back down since that's what I normally do - find the settings the keep the histogram as close to the left as possible but also spread it out as much as possible.

I can try this out with a uniform light source right (t-shirt covering artificial light)? Also thoughts about the cooling issue or do we do one thing at a time?

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

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

#9

Post by ramdom »

I had this interesting bug when I started using this camera and Sharpcap for the very first time. I couldn't get RAW16 to work - whenever I put it on it would either give me a completely black image or an incorrect image (like a thin white line in the middle of it). So I kept using RGB24 (just to test things out) and it would give me this warning about using 8 bits of data only, etc.

I then wanted to make sure that the RAW16 mode could work so I went to EZCAP and it worked fine there so then I came back to Sharpcap and RAW16 started working. Since then I've not had any problems actually using RAW16 mode. I have these intermitment issues like that - this one appears to be thankfully resolved (unless it is contributing to the brain issue) but the other one is the cooler. That also works eventually by either restarting the PC or by disconnecting and reconnecting everything.
ramdom
Posts: 41
Joined: Wed Jul 05, 2017 4:03 am

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

#10

Post by ramdom »

So I went back to check the histogram as you suggested and the red and green are definitely far more spiky than the blue which is shifted to the left of the other two. The more I shift the histogram to the right (or spread it) the more spiky it appears. Yet the spikes themselves are one long curve (even the green and the red), so the blue looks close to being spike free since the spikes are all so close together. None of the settings I tried (primarily gain and exposure) could get rid of the spikes entirely. Just make it bunch up so close that it doesn't look that bad but that means not using the full 16 bits available to me right?

This is under natural light with the scope out of focus and a cloth covering the front.

--Ram
Post Reply