SharpCap Timestamp accuracy

Anything that doesn't fit into any of the other forums
SteveInNZ
Posts: 39
Joined: Tue Feb 14, 2017 6:56 pm

Re: SharpCap Timestamp accuracy

Post by SteveInNZ » Mon Dec 03, 2018 9:17 pm

Would AstroFlashTimer (https://sites.google.com/site/astroflashtimer/home) be of any use to you ? It's intended for occultations and has been tested against the IOTA VTI.

Steve.

Furgus
Posts: 19
Joined: Fri Mar 16, 2018 9:27 am

Re: SharpCap Timestamp accuracy

Post by Furgus » Sat Dec 08, 2018 5:22 pm

Thanks Dave,

Yes, I've thought about how things might run more reliably....I do indeed turn off wifi and bluetooth before every session.
I also change the priority (Task Manager, Details) in W10 to 'RealTime' for SharpCap (it defaults back to standard after a re-boot).
W10 generally manages time better than W7 I am told.
The NTP programmes seem to be RealTime priority anyway. (NTP from local GPS server that is, not internet!)
Thats a good point about the antivirus.

Having said that though.....there really is no problem with whats happening ie the stamping seems correct nearly all of the time. The mystery is why that should be so, as Robin has checked that the timestamp is allocated after the exposure is complete, not before! So there should be a one frame error time....which there is not! (see attachment).
I'd be happier knowing why this is working...but don't want it to change!

Furgus
Posts: 19
Joined: Fri Mar 16, 2018 9:27 am

Re: SharpCap Timestamp accuracy

Post by Furgus » Sat Dec 08, 2018 5:24 pm

Thanks Steve,
I'll have a look at Astro Flash timer.

Furgus
Posts: 19
Joined: Fri Mar 16, 2018 9:27 am

Re: SharpCap Timestamp accuracy

Post by Furgus » Mon Dec 10, 2018 11:12 am

A small clue, possibly, as to whats happening:
I find that when using a USB 2 cable, both ASI224 and ASI174 cameras produce frames that have a 'late' timecode, (very roughly 1 frame duration late).
When using USB 3 (which I always do).....the timestamps appear 'correct' nearly always (ie the timestamps match the 1PPS LED, within the limits of frame duration).
I can confirm that at least for long exposures (several seconds), that frames are given the timestamp value corresponding to the end of each exposure.
Its therefore very odd it appears correct at normal 'video' type frame rates!

Furgus
Posts: 19
Joined: Fri Mar 16, 2018 9:27 am

Re: SharpCap Timestamp accuracy

Post by Furgus » Fri Dec 28, 2018 12:18 pm

Hi,

Had just one more thought on this,
Would there be any circumstances where SharpCap could allocate the timestamp of the previous frame (allocated when its exposure finished) to the new frame, just starting its exposure? (that too would fit what I'm seeing).

Thanks

And have a great 2019!

User avatar
admin
Site Admin
Posts: 2599
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: SharpCap Timestamp accuracy

Post by admin » Fri Dec 28, 2018 3:38 pm

Hi,

I honestly don't think so! I have just tested saving some frames to an SER file when the exposure was set at 15 seconds making it really easy to spot whether the timestamp was captured for the beginning or the end of the frame and in my testing it behaves exactly as I'd expect – the timestamp saved for each frame in the SER file is the time that the frame finished being captured.

Similarly, if I capture single 15 second frames to still image files the timestamp stored in the capture settings.txt file is also the end of frame time.

Cheers, Robin

Furgus
Posts: 19
Joined: Fri Mar 16, 2018 9:27 am

Re: SharpCap Timestamp accuracy

Post by Furgus » Fri Dec 28, 2018 9:52 pm

Thanks Robin,

I agree with you....I tried the exact same trial and it does indeed stamp as expected, on longer exposures (ie the time stamp represents the end of the exposure.) I have investigated several capture programs in this way....they do vary a bit.

But....the anomaly, for me, comes at 'normal' capture rates (say, 15-80 fps) where it seems to stamp the beginning of the frames. (Like the example I sent....and I have hundreds of others! ) I cant see why this would be....as a last gasp I wondered if there was so much going on at the faster frame rates that the 'previous' timestamp was used, (as if in some sort of buffer) if you see what I mean!

Many thanks!

User avatar
admin
Site Admin
Posts: 2599
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: SharpCap Timestamp accuracy

Post by admin » Sat Dec 29, 2018 2:56 pm

Hi,

there are some bit different bits of code that kick in or drop out as the frame rate and exposure length change, but I can't see any that would have the effect that you are measuring. As far as I can tell SharpCap doesn't even know when the frame starts for most brands of camera when they are running in high speed mode! If you do happen to find a changeover exposure time where the behaviour changes from the expected end of frame timestamp to the odd apparently the beginning of frame timestamp then maybe that would give me something to use as a starting point in another dig through the code to see if I can work out what is going on.

Cheers, Robin

Furgus
Posts: 19
Joined: Fri Mar 16, 2018 9:27 am

Re: SharpCap Timestamp accuracy

Post by Furgus » Sat Mar 09, 2019 7:11 pm

Hi,
I wasn't really going to take this any further, as I am happy with how the system works (ie its reliable and accurate), but have found out a couple of bits of extra info:

Firstly, when testing my usual laptop with my GPS server, I found that with an exposure time of 400ms or shorter, Sharpcap was stamping at the start of a frame, while an exposure setting of 450ms or above was stamped at the end of the frame.

Secondly, I was configuring a 'new' different laptop to be used as a 'Capture Laptop' only.....ie no emails, or unnecessary programs...a real slimmed down machine, running on the latest W10 64bit (Version/update 1809). The odd thing is, Sharpcap always stamps every frame at the end ie the time of when the exposure finishes, irrespective of exposure duration. This was unexpected and odd.

The original laptop is a quad core machine also on W10, but not 1809. The 'new' laptop is a dual core machine, not quite as fast but with actually more RAM than the other original one. Don't know if that is relevant.

Can anyone offer thoughts or a possible explanation??!

Thanks,
Rgds
S

PS, My Desktop PC Sharpcap stamps at the beginning of frames also, but as yet have not checked the 450ms 'cut-off' as yet

User avatar
admin
Site Admin
Posts: 2599
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: SharpCap Timestamp accuracy

Post by admin » Sun Mar 10, 2019 9:00 am

Hi,

For ZWO cameras, SharpCap runs the camera in video mode for exposures of less than 500 ms and in still camera mode for exposures of more than 500 ms. However as far as I can see in both cases the code sets the timestamp to the current computer high-resolution time after the frame has been fetched from the camera SDK.

Cheers, Robin

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest