Does SharpCap switch to Still Mode at => 1 second exposures (occultation use)

[graems]

I am using SharpCap with a ZWO ASI432MM camera (global shutter) for occultation work. I use two "goalpost" GPS flashes into the telescope to determine timing. This method relies on the camera being in video mode (i.e. the camera is the master timer outputting frames at a constant rate). The camera has buffer memory which prevents the computer from affecting the frame rate unless the computer is very slow.

I normally see the frame rate slightly less than expected compared to the exposure time (about 80 usecs). I attributed this to the "dead zone" in which data is being read out of the sensor before starting the next exposure. This is not an issue.

But the other night I used a longer exposure than usual (1 second). The dead zone appeared to increase to 23msec. I experimented tonight and found the following:
1) 999msec exposure, dead zone was about 80usec as expected.
2) 1000msec exposure, dead zone was about 31msec.
3) 1001msec exposure, dead zone was about 31msec, and now there is a progress bar display at the bottom right of the screen in which SharpCap is showing time left for the exposure.

This makes me think that SharpCap has switched the camera from video mode to still mode (in which SharpCap now requests each frame from the camera). Can you confirm this is happening? I have seen posts that mention the switch over occurs at 2 second exposure, but has that changed to 1 second exposure now? Is this determined by ZWO's drivers or SharpCap?

Thanks
Graem

[admin]

Hi Graem,

internally, SharpCap switches ZWO cameras over to still exposure mode when the exposure is 2s or more (exception - not when auto exposure is turned on). If the exposure is less than 2s then video mode is used. This is primarily because it is easier to cancel a still mode exposure, meaning that you don't get stuck waiting for the camera to finish a long exposure if you have - for instance - reduced the exposure from 30s to 5s.

Additionally, if SharpCap itself is in 'Still Mode' then the ZWO SDK still mode is always used.

The progress bar being shown for exposures of >1s is separate to this - that is just a usability thing as it turns out that the progress bar is distracting if it keeps flicking across on shorter exposures, but fairly useful for exposures of 1s or above.

What I can't really help with is what is happening inside the ZWO SDK - maybe it changes modes at about 1s exposure - it's not something that you are likely to notice without the sort of timing equipment that you are working with.

One thing I will add is that the frame rate that you see is determined by the longer of

* The exposure in use
* The time taken to transfer a frame's worth of data from the camera to the computer.

The latter depends on the 'Turbo USB' setting and the height of the image. While this is normally more relevant at high frame rates, you will get some dead zone if the transfer time is longer than the exposure - typically the camera will wait until it knows that the USB transfer will end just before the next exposure ends before starting that next exposure. It's just possible that the USB transfer rate is reduced for longer exposures - that might happen at the 1s mark.

hope this helps,

Robin

[graems]

Hi Robin - thanks for the reply. I can see that for a typical user, switching to still mode for longer exposures would give more responsive operation with no drawback.

But for occultation work it is desirable to have understandable, consistent operation. That way we can verify timing accuracy of our data without worring about a number of different scenarios. I was thinking of putting in a request for an "occultation" tick box, that when selected, would hold off switching out of video mode until at least 10? seconds exposure. But the issue I am having sounds like it is out of your hands.

It appears the SDK is switching out of video mode at 1 second. I will need to experiment more. The extra dead zone/jitter between frames from still mode operation might not be an issue for 1 second exposures if it is small enough. But it seems to be quite variable on my computer and that needs to be understood. I tried changing the Turbo USB value and increasing the priority of SharpCap with no luck. Maybe I have to increase the priority of some lower level routine for USB or the ZWO SDK.

Cheers
Graem

[admin]

Hi,

if the SDK is switching the camera to still mode at 1s exposure then I don't think that you have much scope for tinkering successfully on the PC side unfortunately. The ZWO SDK runs inside SharpCap, so you have already been adjusting the priority for that. The driver for these cameras tends to be very simple, just passing USB data back and forth between PC and device, so there isn't any significant room for tinkering there.

My guess is that at 1s exposure or more, instead of the camera starting each exposure automatically on completion of the previous, the next exposure is started by the PC, so that the sequence at the end of the exposure is that the PC first needs to notice the old frame is completed, then send a signal to start the new one. That will be causing your additional latency.

Other brands of camera may handle this in different ways, but unfortunately I can't be sure, since I don't have the sort of GPS timing device needed to test. You can also now get GPS equipped cameras from both QHY and Moravian instruments with potentially new GPS models coming in the future from ToupTek. Not all of these are global shutter, but at least in the case of the ToupTek models you can recover the line delay and therefore workout the exact exposure timing of any line in the image. Of course, these are not going to be cheap...

cheers,

Robin