loading...
S2 split (P2 minus A2) during breathing for the normal, fixed and paradoxical patterns
running...
How to use
Pick a preset (each also moves the stethoscope to the best site) or change any slider. The picture shows the four chambers, the four valves opening and closing, turbulent jets coloured by type, and the four auscultation spots on the chest: click one, or use the selector, to move the stethoscope; the strip below is the phonocardiogram with the ECG and the sound events labelled (S1, A2, P2, S3, OS). Tick Play the sound and keep the time scale at 1.0 to hear the synthesised heart sounds (browsers need a click first). The graphs show the sound, the pressures, a spectrum, how the aortic murmur and the gradient grow as the valve narrows, and how S2 splits during breathing. Space pauses, R restarts. Try: (1) Normal heart, then Severe aortic stenosis: the murmur grade and the Bernoulli gradient jump. (2) Aortic regurgitation: listen at the tricuspid site. (3) Mitral stenosis with the bell at the apex. (4) Split S2 at the pulmonic site, then Wide fixed split.
The science
This is an educational model of auscultation with synthetic data. It is not a diagnostic tool: real heart sounds depend on the chest wall, the patient and the instrument, and a stethoscope exam is only a screening step.
Where heart sounds come from. The first sound S1 is made when the mitral (and tricuspid) valves close at the start of ventricular contraction; the second sound S2 when the aortic (A2) and pulmonic (P2) valves close at the end of ejection. The sounds are transients of the valve leaflets, blood and the ventricular wall vibrating together, so the page draws each as a short damped oscillation, S(t) = a e−t/τ sin(2πft), at the instant the haemodynamic model closes the valve. S1 grows with the rate of rise of ventricular pressure dP/dt; A2 grows with the aortic pressure and is weak when a stenotic valve barely moves. The pulmonic closure P2 is delayed during inspiration (more venous return to the right heart), so S2 splits on inspiration; a fixed wide split suggests an atrial septal defect, and a paradoxical split (P2 first) delayed aortic closure. A third sound S3 is a low-pitched thud from the rapid early filling of a dilated ventricle (a large early mitral flow peak); an opening snap precedes the diastolic rumble of mitral stenosis.
The haemodynamics. A time-varying elastance left ventricle (Plv = E(t)(V − V0)), an arterial Windkessel and a venous reservoir, with valves modelled as Bernoulli orifices:
Q = Cq A √ΔP, Cq = 50.1 mL s−1 cm−2 mmHg−1/2 (from ΔP = ½ ρ v2)
so the jet velocity through a valve of area A is v = Q/A and the pressure drop is the simplified Bernoulli equation ΔP ≈ 4 v2 (v in m/s, ΔP in mmHg). The clinical Gorlin formula, A = Qmean / (44.3 √ΔPmean), uses a slightly different constant (it includes a 0.88 velocity coefficient) and a mean over the flow period, so it recovers the true area only to within a few percent; the continuity equation Avalve = ALVOT VTILVOT / VTIjet is what echocardiography uses.
Murmurs are turbulent jets. A jet through a narrow or leaking valve becomes turbulent and radiates sound. For free-jet turbulence Lighthill's law gives an acoustic power proportional to v8 A, i.e. a pressure amplitude proportional to v4 √A; the model uses this law with a saturating loudness, so a severe stenosis at 4 m/s is about 16 times louder than the same jet at 2 m/s while the healthy valve is nearly silent. The timing follows the flow: aortic stenosis gives a crescendo-decrescendo ejection murmur while the valve is open; aortic regurgitation a decrescendo diastolic murmur; mitral regurgitation a holosystolic murmur while the ventricle is above the atrium; mitral stenosis a low rumble in diastole. A chest piece shapes what you hear: the diaphragm passes high pitches, the bell the low ones (S3 and the rumble). Each source reaches the four chest sites with a different gain (for example the aortic murmur is loudest at the aortic area and travels to the apex).
Numerical checks run in the page. (1) The stressed blood volume is conserved with orifice valves; (2) the Gorlin area from the simulated flow and gradient against the true area of 1.0 cm2; (3) the peak instantaneous gradient against 4v2; (4) the continuity equation recovers the 1.0 cm2 valve; (5) Lighthill scaling (twofold velocity, sixteenfold amplitude); (6) severe aortic stenosis is louder than normal in systole at the aortic site; (7) the normal heart makes one S1 and one A2 per beat.
Honest simplifications. The heart sounds are synthetic (damped sinusoids and filtered noise), not recorded; the right heart is represented only by a delayed P2 and the tricuspid area as a listening site; no atrial contraction (so no S4), no valve flutter, no chest wall transfer function or lung sounds; the murmur loudness scale (grade 1 to 6) is a mapping of the normalised amplitude, not calibrated to clinical grading; audio is resampled from 2 kHz, so the sound is a plausible illustration, not an acoustic simulation. Values are round numbers, not patient data.