Karmin Syncing Process

Post Reply
KieranMITDUT
Posts: 12
Joined: Sun Jun 09, 2019 2:15 am

Karmin Syncing Process

Post by KieranMITDUT »

Nerian,

are there any resources available that explain the process for time syncing the karmin camera to the SSP? And also approximately how accurate it is?

-Kieran
KieranMITDUT
Posts: 12
Joined: Sun Jun 09, 2019 2:15 am

Re: Karmin Syncing Process

Post by KieranMITDUT »

just to clarify, I'm curious as to how this is actually done (not how the user enables it)
k.schauwecker
Posts: 118
Joined: Mon Mar 25, 2019 1:12 pm

Re: Karmin Syncing Process

Post by k.schauwecker »

Hi Kieran,

which synchronization are you referring to exactly? There are two different things that are happening:

1. The cameras exposure times are synchronized, such that both cameras record a frame at the same moment in time. This happens through the trigger cable, where SceneScan outputs a pulse signal. For each pulse, the cameras acquire a frame.

2. When SceneScan outputs a frame, it is annotated with a time stamp. There are two options how this time stamp can be determined. By default, SceneScan takes the timestamp from its internal clock clock, at the time a frame is received. This can be 2 to 3 ms behind the image acquisition. If you activate the "use trigger time as timestamp" option in the trigger settings, SceneScan will take the time where the trigger pulse was output. This is the most accurate way to get a timestamp, but you cannot use a different trigger source in this mode.

SceneScan's clock can itself be synchronized using different methods. I recommend using the PTP protocol. With this method, a sub-microsecond accuracy can be achieved.

Regards,
Konstantin
KieranMITDUT
Posts: 12
Joined: Sun Jun 09, 2019 2:15 am

Re: Karmin Syncing Process

Post by KieranMITDUT »

Hey Konstantin, what I'm curious about is how are the trigger pulses from the trigger cable (which are time stamped in terms of the SSP's clock) associated with a frame that returns via USB.

Another way to ask the question would be, how are images that come immediately after a dropped frame time stamped, and why are they not time stamped with the trigger pulse from the last frame (which was dropped).

-Kieran
k.schauwecker
Posts: 118
Joined: Mon Mar 25, 2019 1:12 pm

Re: Karmin Syncing Process

Post by k.schauwecker »

Hi Kieran,

the cameras themselves have a clock, and we synchronize this clock to SceneScan's internal clock. We also keep a timestamp history of the most recent trigger pulses. Hence, we can then match a trigger pulse to a received frame.

Best regards,
Konstantin
KieranMITDUT
Posts: 12
Joined: Sun Jun 09, 2019 2:15 am

Re: Karmin Syncing Process

Post by KieranMITDUT »

Thanks Konstantin, is it an actual synchronization process? or more of a time translation process (i.e. is it an actual clock on the cameras or is it a counter or something)?

are there any details of how this is done over USB3?
k.schauwecker
Posts: 118
Joined: Mon Mar 25, 2019 1:12 pm

Re: Karmin Syncing Process

Post by k.schauwecker »

Hi Kieran,

the cameras have an actual clock, but that clock is free-running. We estimate the time offset between the cameras' clocks and SceneScan's clocks. For each captured frame we add this offset and then check which trigger pulse time stamp matches the frame time stamp best. This is the time stamp that is then selected.

Regards,
Konstantin
Post Reply