Histogram Movement

[kaymann]

This is not a bug report but a What Do I Do Now...

So I was live stacking and for the first few exposures things looked really Okay then as new frames were added to the stack the histogram slowly slid to the left until it was touching the left and I could not figure out how to move it out again. I have not seen this before and could not figure out what I was doing wrong. This was on over 100 15s frames and then over 100 30s frames using dark and flat-bias frames...

C9.25, ASI533MC Pro, 0°C, .62 Reducer.

Brightness 10, R 50 B 50

Log_2024-05-30T09_38_29-14632.log

(102.54 KiB) Downloaded 62 times

[admin]

Hi,

sadly I don't think that's the right log file, since it doesn't seem to have any information related to live stacking in it - in that log you have opened the 'Deep Sky Test Camera' and used the histogram...

In general, you would only expect that sort of thing to happen in live stacking if the live stacking suddenly gets sent lots of images that are much darker in the background than usual. If you catch it happening again, it would be worth closing live stacking and looking at the main histogram to see how that looks - is the peak pushed right to the left hand side?

Of course, that just takes the trail back a step - what might cause the over dark images? Well, the usual candidates would be

* Over subtraction by dark frames - different brightness used in the dark frames to the light frames, for instance (or different exposure/gain)
* Background subtraction playing up (but I believe previous issues with this fixed in the version you are using).

Flats are unlikely to impact this.

Sorry not to be able to provide any definitive answer, but hopefully the pointers above will prove helpful is the problem happens again.

cheers,

Robin

[kaymann]

I was using Folder camera to try and figure out was happening...

I think this might be the right log file:

Log_2024-05-29T22_29_38-9592.log

(376.53 KiB) Downloaded 70 times

BTW I am currently in the process of rerunning the Darks for a band new set and I never change gain from 105 nor Brightness from 10.

[admin]

Hi,

yes, that looks like the right log - I can see the use of live stacking and the individual captures happening in it. Unfortauntely the live stack doesn't record a great deal in the main log unless something goes badly wrong, so the main log will not reveal much here.

One more thing did occur to me - if the telescope shifted during stacking (something slipped perhaps) then it could move most of the original frame out of view, which would lead to a big part of the stack getting darker as it would no longer be receiving new image data. That would be revealed by a big offset showing up for the alignment of some frames if you replay the stack in the folder monitor camera.

cheers,

Robin

[kaymann]

One telling part of this situation is that last night when I started I was using Hot&Cold Removal and all look good. When I switched to my new darks (created by SC) master from 60 darks the slow shift started up again. With the live stack still going I shifted back to H&C removal and the histogram slowly started to shift back to normal... The darks were 60x 30s, 105 gain(I use 105 to absolutely guarantee I have fallen off the cliff), 10 Bright, 50 R & B. the same as the current live stack... I forgot to mention I used a Antilia Ultra RGB first night and a Optolong L Pro the last two nights and my flats are created just before going outside with the same settings except exposure. I did do a Ctrl Camera start to ensure no weird settings.

Given you comment on drift I will blink through the lights to look for anomalies.

Well I from the fits individual fits I can see no discernable drift (I use a PHD2 guide scope)

EDIT: I just stacked last night using ASTAP with the new Darks, Flats-Bias, and Lights (all FITS). I then threw the master light into Siril and everything looks great. So for now I will EAA with H&C Removal and post process with darks and things look great!

[admin]

Hi,

sounds like you have done most of the things I would have suggested to try to work out what is going on. The only thing left would be to open the master dark and some of the light files in FitsLiberator and have a look at the average pixel values of the two. If the average pixel value of the master dark is higher than the light frames then that would be the cause of the leftward histogram drift .

cheers,

Robin

[kaymann]

Ok so you are correct Left Value is Master Dark Right Value is a light exposure:

Screenshot 2024-06-03 112500.png (12.7 KiB) Viewed 1305 times

So how did that happen? What do I do now? BTW I used SC to create the Master Dark set to the same gain, temp, exposure time, etc as the light ...

[admin]

Hi,

check the capture settings files saved with the master dark and the light frames to see if any of the camera settings are different when you expected them to be the same (colour balance for instance - they act as additional digital gain).

cheers,

Robin

[kaymann]

All exactly the same...

[admin]

Hi,

all very odd then. The only similar thing that springs to mind was someone a couple of years back who had a seemingly similar problem - they were fairly sure that the ZWO camera in use seemed to be changing its settings without being asked to (similar sort of symptoms), but we never worked out why or how...

SharpCap normally doesn't ask the camera for values of the various settings frequently - it will send setting changes to the camera and refresh the values from the camera at certain points (like changing resolution), but otherwise it trusts the camera to stick with the latest value it has been sent. However, if the camera for some reason has changed the brightness value on its own, SharpCap's display won't update to show the change (and the capture settings file will not show the change).

So... if you see it happen again, try changing the binning/resolution of the camera, which should trigger SharpCap to refresh the settings values, which might just suddenly show a different value for brightness. On the other hand, if it doesn't show a change, that doesn't really prove anything, since maybe the camera changed settings on its own due to some bug, but the ZWO SDK doesn't know that has happened...

cheers,

Robin