sensor analysis doesn't work on qhy 294c

A place to report problems and bugs in SharpCap
Forum rules


If you have a problem or question, please check the FAQ to see if it already has an answer : https://www.sharpcap.co.uk/sharpcap-faqs

Please also read about Troubleshooting USB Issues before posting.

*** Please do not post license keys - please report any problems with licensing to 'admin' by private message ***

Please include the following details in any bug report:

* Version of SharpCap
* Camera and other hardware being user
* Operating system version
* Contents of the SharpCap log after the problem has occurred.
[If SharpCap crashes, please send the bug report when prompted instead of including the log]
Post Reply
jasonn
Posts: 1
Joined: Sun Apr 21, 2019 1:21 pm

sensor analysis doesn't work on qhy 294c

#1

Post by jasonn »

The gain setting on this camera varies from 0-4096 where the last 800 or so values have the majority of the useful settings from what I can tell. I wouldn't even assume linear responses from these values (it looked like sharpcap is sampling the gain space linearly)

But this value range seems outside what sharpcap assumes (typically 0-100?). I tried dozens of times to do the sensor analysis with a variety of my a-spike flat box's settings on an f/7 refractor, I eventually had luck one of the times with the maximum brightness setting, which is pretty bright.

I am not the only one who had this issue with the imx 294 chipset and sharpcap:
viewtopic.php?t=1433

I've only gotten useful shots outside at night time w/ 1-2 second exposures with gains of 3200-3700 and even then I'm not sure I'm using good values or offset. This is looking at streets and sky.

Here is the result of my first and only sampling but I don't think it is a useful one:

Gain Value e/ADU Read Noise (e) Full Well (e) Relative Gain Rel. Gain (db) Dynamic Range (Stops)
0 3.98874477501322 6.75827423646607 16337.8985984542 1 0 11.2392799204715
50 3.91074866945095 6.69876058977144 16018.4265500711 1.01994403428979 0.171526841788361 11.2235506303038
100 3.66634944244231 6.41919932881018 15017.3673162437 1.08793360742945 0.73204785564718 11.1919509365338
150 3.41459545137105 6.06825390151354 13986.1829688158 1.16814563593811 1.34993981685329 11.1704332128918
200 3.32724294901436 6.04072190786113 13628.3871191628 1.1988138035411 1.57503469827192 11.1396062410976
250 3.15362069071085 5.84439982140279 12917.2303491516 1.26481437249644 2.04053583962835 11.1099542691892
300 3.15536150763239 26.291853535086 12924.3607352623 1.26411657281265 2.03574250099814 8.94126135506965
350 3.17668457180498 6.11122132795008 13011.7000061132 1.25563136183422 1.9772430902486 11.0560611135274
400 2.76855850259265 5.55062856052345 11340.0156266195 1.44072981346716 3.17165086510271 10.9964838544654
450 3.15416242721403 6.49836234842439 12919.4493018686 1.26459713697635 2.03904388487878 10.9571807647502
500 3.0494932320576 6.43428799262426 12490.7242785079 1.30800250122932 2.33217148938352 10.922789002942

The frailty of this routine is making me consider just grabbing the images and doing the sample space myself... I am a bit surprised because of all the software's I could find, sharpcap seems to be the only one offering an automated approach to this automated problem. But the insight into why it isn't working isn't there.

What leads to this "minimum exposure reached" error? I get it on the sensor analysis's third pass.

It also seems to freeze (no more script progress) when switching to 2x2 binnings, so I've been checking skip binning measurements. Program is otherwise alive.

I've seen it go tri-modal in the histogram peaks more than a few times only in ascom mode vs the native qhy driver mode, too. I kind of would expect this with a CFA (it is a color camera). I get single peak when using the direct driver, for the same exposure, offset settings.
Attachments
294_log.txt
(249.66 KiB) Downloaded 95 times
User avatar
admin
Site Admin
Posts: 13175
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: sensor analysis doesn't work on qhy 294c

#2

Post by admin »

Hi,

The problem here is that SharpCap is expecting the gain value on a qhy camera to be either a number of dB of gain (and a range from 0 to 50 or so) or the gain in tenths of the dB with a range of 0 to 500 or 600 or so. This is how things used to work, but it seems that with a lot of the new models the SDK is using some sort of raw gain value that has no true physical meaning.

Essentially this is important because some of these camera models have such a wide gain range that it is difficult to run the sensor analysis final stage correctly with a single illumination level – either the exposure required become so short at hi gain that accuracy in exposure length cannot be maintained or the exposure become so long at low gain that the whole procedure takes too long to execute. In order to avoid this situation, SharpCap understands the gain values for major brands of cameras and limits the ranges of gain that it will try to sample – typically only going from minimum gain up to about 50 dB.

The gain values you are describing for this 294C qhy model are not following the normal patterns at all, so SharpCap has no idea how to limit the gain range being tested as it normally would and this almost certainly leads to the problems you are observing. This problem seems to be unique to recent qhy cameras – all the other major manufacturers seem to be on top of this and ensure that their new models follow whatever guidelines for gain value meaning that they have established with previous models.

The tri-peaked histogram is typically a sign of a colour sensor not being identified correctly as such – when SharpCap knows that it is dealing with this colour sensor, it extracts the green channel only for the histogram. If however it is unaware that the sensor it has a colour filter array then it would use all pixels for the histogram shape and show the structure that you observed. This can either happen because the SDK is not correctly identifying the cameras are colour camera or because the 'Debayer preview' option is switch to off in the SharpCap camera controls – this effectively overrides the sensor identification performed by the SDK code.

Hope this helps, Robin
Post Reply