new optimal exposure calculator should use latest image when in sequence

[Former user]

Hi,

Checking out at the new optimal exposure calculator, I noticed it no longer uses the latest captured image. Why is that? Essentially, it could auto fill in the gain (and offset!), the filter, the exposure duration etc. The only thing should be left to fill in is the BIAS mean, FWC in e, and the read noise which are essentially constants.

Regarding read noise, for modern cmos cameras the read noise is some function of the gain. Perhaps it would make sense to enter a table of read noises per gain. So users can input their preferred gain settings (most of us settle on a few that we use), and the calculator could pick it up by itself.

This will essentially require the user to populate the require attributes only once, and the calculator could show the optimal duration based on the previously taken photo (as it was before). When not in a sequence, an option to take a measurement shot would be useful, even then all the info will already be present.

It would be also very cool to have preloaded values for commonly used cameras, populated automatically when a recognized camera is connected.

Am I making sense? Am I not getting something about the new design?

Thanks!

[Shlomi Vaknin]

Forgot to login before posting. Surprised it allowed me to post an issue without logging in, pretty cool though!

[Stefan B]

Hey,

the reason the exposure calc was moved into a separate module is due to the fact that it is a tool to determine the exposure time prior to starting the sequence. There is not much sense in having this tool run during exposures, as it requires a specific amount of exposure time and settings.

For the other ideas to store the values per gain, yes i thought about it and it might come in a future build.

[Shlomi Vaknin]

Hey Stefan,

I agree that the option to pre-calculate the exposure is important. However, as the sequence progresses, after some time, the position of the objects change and the exposure requirements may change accordingly. For example, last night, the mean background started at around 1100 ADUs, but by the end of the session it went all the way down to around 800 ADUs.

Another interesting option would be to adjust the exposure times automatically, so instead of sequencing on a exposure duration, we could sequence on the background mean. To make darks easier to manage, perhaps we could specify the different durations that we would like to have (60s, 100s, 120s, 150s, 200s, etc) so we can built a dark library prior and have the sequencer adjust.

Thanks for all your wonderful work!

‌

[Niccolò Coli]

Can you imagine the complications in processing/calibrating an imaging session with X different random exposure times throughout the night? The difference in swamping the RN are marginal enough to keep your exposure constant unless your sky goes from a B8 to a B5 during the course of a session

[Shlomi Vaknin]

‌

Can you imagine the complications in processing/calibrating an imaging session with X different random exposure times throughout the night?

Of course, random exposure length will be a nightmare to handle. That is exactly what I meant by:

To make darks easier to manage, perhaps we could specify the different durations that we would like to have (60s, 100s, 120s, 150s, 200s, etc) so we can built a dark library prior and have the sequencer adjust.

So exposure length will not be selected “randomly”, rather, out of the set of exposure-lengths that we have defined and prepared master darks for, the closest exposure length to the calculated optimal would be selected. In any case, you have a good point about the added value may being marginal.

To my original point though, I do think the option to at least chose the last exposure as the basis for optimal exposure calculation could be very useful, as well as allowing for multiple gain entries.

[Stefan B]

Even with having matching darks the stacking process will be really difficult as you cannot just stack different exposure times together in one run.

Anyways - what i will take out of this issue is the proposal to maintain the read noise and other parameters per gain in some table to be used, when trying to get an exposure time for different gains to not having to re-enter these values over and over again.
For the other topics, this is not in scope and also not intended for this tool.

[Shlomi Vaknin]

Stefan, I don’t know, APP does not seem to have any problems stacking different exposure lengths at one go. But regardless, that was just a wild idea.

Anyways - what i will take out of this issue is the proposal to maintain the read noise and other parameters per gain in some table to be used, when trying to get an exposure time for different gains to not having to re-enter these values over and over again.

What about the OPTION to use the last image taken?

[Stefan B]

exposure calc is reworked in 1.10 beta. for multiple gains the sensor analysis from sharpcap can be used