As an example, consider making a photograph of the following scene:
It’s after dark and you’re standing out in an open valley out in the great American West — say, perhaps, in eastern Utah — on a clear summer night, with zillions of stars and a full moon so bright that you can make out details of the lunar terrain.
You’re not sure how you got here or how you’ll get home, but you watch as a long passenger train (the California Zephyr) zooms by at 65 mph (100km/h), the individual passengers visible in the windows, illuminated by the little reading lights over their seats.
To create the most “accurate” image of what you see, you’d have to combine at least four different exposures:
1. one [long] exposure to show the details in the moonlit terrain;
2. one exposure for the train (for the moving train not to be a blur, it would probably have to be a daylight photo, severely darkened so it looks like night — a darkening that would by itself be highly deceptive);
3. one exposure for the passengers who are visible through the windows (good luck with that if the train is moving);
4-5. one or two exposures to show the stars in the sky and the detail in the moon.
Viewers of that composite image may agree that it is a good representation of what the scene looked like to someone who watched the train go by, and the image could be deemed “accurate” in that sense.
But any viewer of the photograph who even suspects that the train image is a composite of multiple very different photographs — taken at different times of the day, in very different lighting — would not bother to apply their usual curiosity of how much they should trust what they’re seeing.
They would regard the composite very differently than they would regard a TCQ photograph.