I was wondering what is the efficiency of human muscle? For example, how many kilocalories does one burn in order to, say, increase the potential energy of a bar while bench pressing. Of course, simple PE calculations wouldn't quite work--you'd really have to look at an integral of acceleration (~force) over a range of motion to calculate the actual work done, but you get the idea.

Similarly, when doing cardio, you are essentially working against a brake dynomometer. Is the number of calories burned reported by the machine a direct representation of the work disipated by the dyno, or is it modified by some average efficiency value?

Yes, I was bored, and yes, I am an engineer.