Mouthpieces and masks change breathing, and distract the subject. Accepted non-invasive methods avoid this problem, inductive plethysmographs and respiratory magnetometers, but are expensive and unusable in magnetic resonance imaging (MRI) scanners. Because changes in ventilation affect arterial gases, and thus cerebral blood flow, measurement of breathing is desirable during many functional MRI studies. Using an old principle, we constructed an inexpensive, non-invasive device unaffected by magnetic fields. We adapted a simple calibration method to reduce error and make the method accessible to more users. 'Pneumobelts' consist of flexible corrugated silicon tubes worn around the rib cage (RC) and the abdomen (AB). Changes in RC and AB are determined from pressure changes within the 'pneumobelts'. Estimates of tidal volume are generated from the sum of the RC and AB changes. We empirically determined the appropriate RC weighting as 1.3:1 (RC:AB). Volume estimation was tested (n = 9) in different body positions and during different breathing maneuvers. The weighted sum of the two signals gave an accurate estimate of tidal volume with tidal volumes less than 1200 ml (mean error = 6-7%). Breaths over 1900 ml produced larger errors (mean error = 11-16%). Our results are generalizable to any linear circumference measuring device.