Exposure time range is returned in different units for different camera types

[PB-Ole]

Hi,

I hope someone can help me with this.

We have different camera types in use (DFK 33GX236 and DFK 37AUX290 for example) and are using the tisgrabber library to control the camera settings in our C++ application.
We want to allow the user to change the exposure time within a range limited a) by the camera itself and b) we set an upper limit of 33 ms to allow for a minimum of 30 FPS.
We use a call to

IC_CameraPropertyGetRange(grabber, CAMERA_PROPERTY::PROP_CAM_EXPOSURE,  &camExposureRangeMin, &camExposureRangeMax)

to obtain the min and max exposure time of the camera device. Now the problem is that for the DFK 37AUX290 this gives us

// unit not clear? 2^-15s to 2^4s?
camExposureRangeMin=-15
camExposureRangeMax=4

and for DFK 33GX236 we get

// unit is micro seconds
camExposureRangeMin=20
camExposureRangeMax=30'000'000

Is there a possibility to obtain the range in the same unit independent on the device? Or a way to know which unit is being returned by IC_CameraPropertyGetRange?

I have found the IC_GetExpAbsValRange(grabber, fMin, fMax) method which gives me the exposure range in micro seconds for the DFK 37AUX290 type, but for the DFK 33GX236 the application is crashing, even though (or because?) I check IC_IsExpAbsValAvailable(grabber).

I would like to avoid hard-coded camera types and different handlings depending on them..

[TIS-Stefan]

Hello

The -15 to +4 value range is the old style Physical Values interface. Do you have the correct driver version 5.1 installed for your camera?

Also I suggest to use the Absolute Values interface of exposure instead of range. Then you get the exposure times in fraction of seconds from both cameras.

Additionally, I suggest to switch to IC Imaging Control 4 and discontinue using tisgrabber, if possible.

Stefan

[PB-Ole]

Hi @TIS-Stefan

many thanks for your very quick reply. Switching to IC Imaging Control 4 is at the moment unfortunately not possible for us. We do have the 5.1 driver version installed.

I would like to switch to the absolute values interface, but then I get problems with floating point precision.

If I do

float fMin, fMax;
IC_GetExpAbsValRange(grabber, &fMin, &fMax);

Then fMin is supposed to be 2e-5, as the minimum exposure duration for that camera model is 20 us. But this is not representable by a float, so it is in fact something like 1.999999995e-5.
So if I subsequently call

IC_SetExpAbsVal(grabber, fMin)

My application crashes, as fMin is outside the allowed range. I am already surprised that it crashes and does not simply ignore the invalid value, but that is probably another issue.

I will try to use double precision, but I am not sure how the tisgrabber library will handle this. Or do you have another suggestion what I should do?

Cheers,
Ole

[TIS-Stefan]

Hello Ole
"1.999999995e-5." is a very nice float number, but it is displayed in the exponential way.
What does "crash" mean? Is there an error message?

[PB-Ole]

It is a very nice float number, I agree, but tisgrabber does not seem to handle it well :)

There is no error message displayed. I have gdb attached and get until the line

IC_SetExpAbsVal(grabber, fMin)

where fMin is a float which was set by the preceeding call to IC_GetExpAbsValRange(grabber, &fMin, &fMax); and not changed by me after that. And there the crash happens. I see that fMin is 1.99...e-5 before the call and that I believe is the problem, because it is less than the allowed minimum value. Or am I missing something?

My current workaround is simply to add 1 mico second to fMin, since I don't really care about 1 us and then the application seems to be stable.

[TIS-Stefan]

Hi Ole

Which compiler do you use?

Stefan

[PB-Ole]

Hi Stefan,
we use MinGW with GCC 12.2 to compile our application on Windows 10.
Cheers,
Ole

[TIS-Stefan]

Hi Ole
I wonder, why your application ends without error message, when setting a value out of range. The function in the tisgrabber.dll checks for the range:

	/////////////////////////////////////////////////////////////////////////////////
	// Set the current value of a given property and element combination. 
	// AbsoluteValue interface is used
	// Property and Element names are not case sensitive
	int setPropertyAbsoluteValue(DShowLib::Grabber &Grabber, char* Property, char *Element, double Value)
	{
		int iSuccess = IC_SUCCESS;
		tIVCDAbsoluteValuePropertyPtr pAbsval = NULL;

		iSuccess = GetAbsValPropertyPtr(Grabber, Property, Element, pAbsval);

		if (iSuccess == IC_SUCCESS)
		{
			if (pAbsval != NULL)
			{
				if (pAbsval->getRangeMin() <= Value && Value <= pAbsval->getRangeMax())
					pAbsval->setValue(Value);
				else
					iSuccess = IC_INDEX_OUT_OF_RANGE; 
			}
		}
		return iSuccess;
	}

Maybe there is an error message in the the Windows EventViewer.
However, if it is working with your workaround, you should stay with that.

Stefan