How Precisely Can ExifTool Report Time?

[kbellis]

The fascinating thread started by Doug Kerr back in June 2011 has me curious as to what might have been developed since and in particular, how precisely can ExifTool report time; e.g., 00hr:00min:00.0001sec.

From the following can I assume that the actual value of the exposure was 6.94 seconds?


If this assumption is correct, could one write a tag something like:
(integerExposureTime+(SubSecTime/100))

[Phil Harvey]

ExifTool will report whatever precision is stored.  In theory, the limit is about 65000 decimal places in JPEG images, or about 4000000000 digits in TIFF/RAW images.  Here is a quick example:

Code Select Expand

> exiftool a.jpg -datetimeoriginal=now -subsectimeoriginal=123456789012345678901234567890123456789012345678901234567890
    1 image files updated

> exiftool a.jpg -subsecdatetimeoriginal -S
SubSecDateTimeOriginal: 2014:01:12 20:03:34.123456789012345678901234567890123456789012345678901234567890

- Phil

[kbellis]

What would be the proper call / switches to see the extent of what value was stored for ExposureTime?

Code Select Expand

Warning: Tag 'subsectimeexposuretime' does not exist

[kbellis]

Phil - After looking at your example, I should reword the question:

Any ideas on how precisely ExposureTime is recorded and what is the proper command line switch to get ExifTool to report EXPOSURE time?

Ideally, at least for these bulb ramping experiments, 0.0001 second (one millisecond) would be determinable.

Thank you for any reply.

Kelly

[Phil Harvey]

Hi Kelly,

Sorry, I misunderstood.  This shows the precision of ExposureTime:

Code Select Expand

> exiftool a.jpg -exposuretime=0.12345678901234567890
    1 image files updated

> exiftool a.jpg -exposuretime
Exposure Time                   : 1/8

> exiftool a.jpg -exposuretime -n
Exposure Time                   : 0.1234567901

With the -n option, you see the full resolution stored in EXIF.  This resolution is limited by the precision of the 8-byte rational storage format.

- Phil

[kbellis]

Thanks Phil.

Guess that's the end of the road ... for now, as full resolution values (invoked through the -n switch) still appear to be expressed to just the tenth of a second; no different than without it. However, Kerr's discussions still has my interest and that recordings of exposure time in the millisecond range might be possible.

This resolution is limited by the precision of the 8-byte rational storage format.

My understanding of this is limited, but I think this means we have the potential afforded through the 8-byte rational storage format to record a much more precise value than is being recorded at a 0.1 sec. Is this correct?

Kelly

[Phil Harvey]

I think this means we have the potential afforded through the 8-byte rational storage format to record a much more precise value than is being recorded at a 0.1 sec. Is this correct?

Yes.

- Phil