Page 1 of 1

ASCOM and Offset -- a call for expertise

Posted: Fri Mar 13, 2020 12:23 am
by iamhondo
Recently, there's been a discussion on the ASCOM standards forums (in Groups.io) concerning the offset values for cameras.

Some users want the ASCOM designers to add an Offset property to the Camera interface. The offset value seems to be commonly available and controllable from many camera vendors' API. So the pro-offset team suggests that that if the vendors offer access/control to a property, then ASCOM should expose a new property on the Camera interface to support it. That's how things work with the Gain property of the Camera interface.

The anti-offset team says that offset is an obscure, arcane property with little tangible benefit to common users. The anti's contend that adding a property to the interface will trigger unwelcome ripples of development effort by ASCOM driver writers and the apps that use ASCOM. So unless the offset value has significant value, it isn't worth the collective development effort.

That's a synopsis of a "vigorous" series of technical wrestling matches. The whole kerfuffle could be quelled by appropriate expert testimony on the intrinsic value of offset for astrophoto imaging. A quick/easy contribution would be to clarify how offset influences image capture and processing.

Volunteers are welcome at the ASCOM group site: https://ascomtalk.groups.io/g/Help/topics .
The ASCOM group has separate sections for developers and users. This discussion is in the user section ("Help") because a user initiated the request. The topic is "CMOS Camera ASCOM Driver Offset".

Oh, and how does this tie to SharpCap? The SC Sensor Analysis feature needs to march through a variety of camera offset values to refine the data we use in the Brain. With current ASCOM drivers, the offset cannot be adjusted (or read). AFAIK, SC is currently only able to do SA on non-ASCOM driver cameras. If ASCOM implements an Offset property in the Camera interface then SC can do SA for "any" camera with an ASCOM driver.

Regards,
Joe

Re: ASCOM and Offset -- a call for expertise

Posted: Fri Mar 13, 2020 8:53 pm
by admin
Hi,

The value of an adjustable offset control is fairly simple and is very much tied to having an adjustable gain control.

One of the things that you should always try to achieve when imaging is to separate the histogram from zero ADU point so that there are no pixels where the readout value is zero (and which potentially perhaps should have had a negative readout value except that the value is limited to 0).

When cameras have no adjustable gain control, it's easy for the manufacturer to pick a single value for the offset that is just high enough to ensure that the histogram remains unclipped without wasting a significant amount of ADU value between zero and the mean level for a bias frame (which is essentially controlled by the offset).

When the camera has an adjustable gain control, the width of the histogram peak bias dark frames will vary massively as the gain is changed. An offset value that separates the histogram from zero ADU for minimum gain will be totally insufficient when used at maximum gain. An offset value that separates the histogram correctly at maximum gain will be wasting a considerable range of ADU and will effectively be reducing the dynamic range and full well depth available if used at minimum gain.

Here is a typical histogram taken at high gain (50 dB), which is just about adequately separated from zero ADU by increasing the offset control for the camera
50db.PNG
50db.PNG (39.35 KiB) Viewed 2912 times
Setting the same camera back to minimum gain but leaving the offset alone gives you this histogram, which makes it clear just how much of the ADU range would be wasted if there was no adjustable offset control
0db.PNG
0db.PNG (12.04 KiB) Viewed 2912 times
Note that SharpCap's sensor analysis can run quite happily on a camera that doesn't have an offset control – in fact it makes things simpler from the measurement point of view. The reason the sensor analysis has to make multiple measurements with different offsets is so that it can be aware of the ADU level expected for a bias frame at any offset control value that the user might choose. If there is no way for the user to change the offset then SharpCap only needs to make a single measurement for the ADU value of a bias frame.

Cheers, Robin

Re: ASCOM and Offset -- a call for expertise

Posted: Sat Mar 14, 2020 1:14 pm
by mAnKiNd
Can the smart histogram analysis use this to suggest minimum "viable" offset value? Because I've seen it suggest values that would be too low for a dark frame resulting in clipped data.

Re: ASCOM and Offset -- a call for expertise

Posted: Sat Mar 14, 2020 6:00 pm
by iamhondo
Robin,

With your permission, I'd like to post your statements and graphics into the discussion on the ASCOM group site.

I don't have much expectation that this will convert any of the adversaries and produce a tidy consensus. You've given each side a little more ammo for their particular interpretation and valuation of offset. And they'll congratulate themselves for the expert validation of their respective positions.

Bolstering each side of an argument is one of my favorite interventions. So I'm perfect delivery person for this message.

Thanks!
Joe

Re: ASCOM and Offset -- a call for expertise

Posted: Sat Mar 14, 2020 9:06 pm
by admin
Hi Joe,

Feel free to quote this. In general I think I would be in favour of this being directly controllable via the Ascom camera interface as it is a significant limitation for CMOS cameras.

Cheers, Robin

Re: ASCOM and Offset -- a call for expertise

Posted: Sun Mar 15, 2020 7:59 pm
by iamhondo
Robin,

Thanks much. I just posted on the ASCOM group page.

If there is noteworthy response, I'll share with you.

Clear skies,

Re: ASCOM and Offset -- a call for expertise

Posted: Tue Mar 31, 2020 2:55 pm
by iamhondo
Robin,

The issue of ASCOM and the proposed offset property in the ASCOM Camera interface is still under debate. Your contribution to the discussion was well very received.

Your insight is so valued that Bob Denny, the father of ASCOM, would like to get in contact with you. ASCOM is still evolving and he's very eager to discuss upcoming advances with the best app writers. It seems like a valuable chance for each of you to benefit from the other's experience and knowledge.

I have Bob's contact information. What's the best way to get that to you privately? (You're welcome to pinch my email address from the forum member database.)

Joe

Re: ASCOM and Offset -- a call for expertise

Posted: Tue Mar 31, 2020 9:09 pm
by oopfan
Robin said:
When cameras have no adjustable gain control, it's easy for the manufacturer to pick a single value for the offset that is just high enough to ensure that the histogram remains unclipped without wasting a significant amount of ADU value between zero and the mean level for a bias frame (which is essentially controlled by the offset).
In my opinion astro-camera manufacturers dropped the ball when they introduced CMOS sensors to the market. They could have easily written their firmware to adjust Offset as a function of Gain. It doesn't need to be a complex continuous function, instead a step-function is more than adequate.

Of course it is too late to put the genie back in the bottle: an orphan child of short-sighted thinking.

Brian