Biosensor Array Respiratory Volume/Frequency
My current approach to this problem is to mount a small muffin fan in place of one of the filter cartridges on a painting respiration mask. I'll read the voltage produced by the fan being spun by the flow of breath. That voltage will be translated to a volume. An improvement will be to read the voltage in both directions, to measure flow rates of both inhalation and exhalation. I think this will require some additional circuitry, to deal with the negative voltage in one direction.
According to wikipedia the maximum typical total lung capacity is about six liters of air and the typical range of respiration rates is between 12 and 20 breaths per minute. Minimal, at rest or tidal volume per breath is about one half liter of air. Maximum inspiratory capacity appears to be about 4 liters of air.
So, the sampling rate must be at least once every second, in order to hope to capture the maximum rate of one breath every three seconds. The sensor used to capture this data must be sufficiently sensitive to record a minimal flow of 0.1 liters per second and a maximum flow of 1.4 liters per second. There's also a need for identifying any pauses in between breaths where the subject is neither inhaling nor exhaling.