Every preset, checked again.

Reference landmarks: RV in the near field; LV in length, MV, LVOT/AV and LA. The standard PLAX does not usually include the true apex. Real BSE example · page 8 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: The principal long-axis structures remain recognizable. The original plane is retained: the candidate with the brightest image rotated the LV too far toward an apical presentation. Residual shadowing and foreshortening need manual review.

Doppler: Inflow and outflow are visible in the open window; colour varies with beam angle. This view is less suited to assigning an absolute velocity from the uncorrected beam projection.

Reference landmarks: Central aortic valve/root surrounded by RVOT, right-heart structures and LA. Real BSE example · page 15 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: The plane now follows the inferred aortic-valve orientation and a revised chest contact. Root visibility improves, but the thorax still removes part of the surrounding anatomy. A clear three-cusp appearance is not consistently resolved.

Doppler: Colour is confined to sampled blood spaces, with ejection in the root/outflow region. Off-axis colour and aliasing cannot establish a normal valve jet.

Reference landmarks: Circular basal LV with anterior and posterior mitral leaflets; RV anterior. Real BSE example · page 17 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: The level follows the measured LV axis. The control image shows the leaflet region, but the thorax obscures much of the LV. The classic two-leaflet opening is not reliably identifiable throughout the cycle.

Doppler: Filling colour is present where the section intersects blood. Missing colour in the shadow is expected; it cannot be interpreted as absent filling.

Reference landmarks: Circular mid LV with two papillary-muscle sections; RV crescent anterior and left. Real BSE example · page 18 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: The revised section follows the LV axis and shows intraventricular structures in the control. A clear pair of papillary muscles is not consistently resolved. A broad thorax shadow obscures part of the LV; the RV contour remains wider than the reference.

Doppler: Mixed or weak colour is plausible with largely through-plane flow. A short-axis view should not be forced to reproduce the long-axis inflow jet.

Reference landmarks: A small distal LV cavity; the RV apex disappears in the reference apical section. Real BSE example · page 18 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: The atlas section retains a substantial RV component. Moving farther toward the tip reduces both cavities together in the inferred labels. It is now labelled distal exploration rather than claiming a verified apical PSAX level.

Doppler: Sparse colour near the apex is not itself a failure: slow motion, the wall filter and a nearly transverse beam reduce the displayed component.

Reference landmarks: Apex uppermost; RV/RA screen-left and LV/LA screen-right, both AV valves, without an intentional LVOT section. Real BSE example · page 19 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: The plane now passes through the inferred ventricular tip and both AV-valve centres. The control shows the expected chamber arrangement. The static thorax strongly obstructs the apical path; a bright complete clinical A4C is not reproduced.

Doppler: Early filling is predominantly toward the probe in the LV in the unattenuated control. Colour is strongly reduced by the thorax shadow.

Reference landmarks: LV and LA; inferior wall screen-left, anterior wall screen-right. Exclude the RV and central LVOT. Real BSE example · page 27 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: Rotation around the inferred LV axis excludes the right chambers from the principal slice. LV/LA are recognizable in the control. Anterior/inferior wall orientation is not independently verified from the unlabelled combined mesh, and the thorax window remains poor.

Doppler: LV filling has a predominantly toward-probe component. This view is not intended to contain an aortic ejection jet; remote vessel sections may appear beyond the LA.

Reference landmarks: LV/LA with MV, LVOT and AV; the aortic outflow lies to image right. Real BSE example · page 28 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: The plane uses the inferred ventricular tip, MV and AV centres. LVOT orientation agrees with the reference arrangement. The revised window gives a more recognizable outflow section, but shadowing still prevents a complete standard acquisition.

Doppler: The dominant unwrapped ejection component points away from the apical probe. At the sampled ejection phase, 67% of colour-eligible aortic samples exceed the scale, so wrapped red is prominent. This does not diagnose turbulence or regurgitation.

Reference landmarks: Anterior angulation from A4C adds LVOT/AV while retaining the four-chamber context. Real BSE example · page 23 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: Angulation now aims at the inferred AV centre. The control shows the additional outflow region. Strong apical shadowing and the exact continuity of the LVOT limit validation of this as a standard five-chamber acquisition.

Doppler: The unwrapped LV/outflow ejection component is predominantly away from the probe; the thorax suppresses much of its displayed colour. About 92% of the sampled ascending-root colour is aliased at the reviewed scale.

Reference landmarks: Liver in the near field and four chambers beyond, with septa relatively transverse to the beam. Real BSE example · page 29 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: The liver and four-chamber section remain recognizable. This is one of the more usable windows in the supplied static thorax. Breathing, probe pressure and exact septal alignment are not reproduced.

Doppler: Inflow colour remains visible through the liver window. Hue varies along the oblique beam; the near-field liver must remain grayscale.

Reference landmarks: Ascending aorta screen-left, transverse arch above, descending aorta right; branch vessels above and RPA below. Real BSE example · page 31 ↗

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: Arch and branch anatomy are partially visible. The ascending limb is substantially shadowed and the exported descending vessel ends limit continuity. An alternate fitted plane did not improve this reliably, so the existing plane is retained.

Doppler: The reference expects toward-probe flow in the ascending limb and away-probe flow descending, with weak signal near perpendicular flow. The sampled ascending and descending components have the expected unwrapped signs, but about 93% and 100% respectively exceed the colour scale. The model has mixed/aliased colour and limited distal coverage; the complete reference pattern is not verified. Added thorax vessels have acoustic geometry but no solved flow field.

Reference landmarks: No assigned standard echocardiographic equivalent.

Previous preset · depth compensation 0.30.

Updated preset · depth compensation 0.45. Rib and lung attenuation retained.

Same reviewed probe and settings, with thorax acoustic effects disabled. Pericardium retained.

Anatomy and visibility: The default contact moves from a nearly opaque posterior lung path to the anterior side of the same free plane. Some cardiac anatomy is now visible; a broad shadow remains. It should not be labelled or interpreted as a standard echo view.

Doppler: Colour follows the local beam projection only where the inferred blood-flow field exists. There is no reference hue pattern for this arbitrary section.

What the Doppler comparison can establish

In the reference convention, red means toward the probe and blue away. Apical AV inflow is predominantly red and LV outflow predominantly blue before aliasing. Flow perpendicular to the beam can appear weak or absent. The ascending and descending arch should have different signs; a uniform colour arch would be misleading. These principles are described in ASE section 5.2, page 38.

The actual shared GPU velocity functions are sampled in every plane at filling and ejection. The measurements include signed beam components, colour fractions, aliasing and rejection of samples inside tissue. Those numerical probes precede acoustic propagation. The displayed captures include the thorax and show how much colour is subsequently lost to shadowing.

Percentages describe colour-eligible samples in the current synthetic field before thorax attenuation, not agreement with measured patient velocities. Arch bins are approximate reference-coordinate segments. At the current scale, aliasing reverses much of the displayed aortic colour. Raise PRF within the depth limit or lower frequency to reduce wrapping; do not interpret a red patch alone as reversed flow. No colour was produced in the sampled tissue points above the mask threshold.

This is still an approximate velocity-to-colour display. It lacks pulse-ensemble estimation, RF/IQ data, measured flow or pressure, probe motion clutter and a complete peripheral circulation. Grain and aliasing can resemble clinical colour without demonstrating an accurate healthy velocity field. Added surrounding vessels do not yet have their own flow solution. We cannot claim quantitative agreement with real Doppler velocities from these pictures.

Changes and remaining work

The apical sections use the inferred LV tip and valve centres; short-axis levels follow the measured LV direction. The aortic-valve plane and selected skin contacts are revised. Depth compensation is modestly increased for the added soft-tissue path. Free oblique starts from the anterior side. Distal PSAX is explicitly labelled exploration because the RV persists.

A brighter candidate was rejected when it required an implausibly rotated PLAX or an abdominal contact for an apical view. The existing arch section was retained because its candidate remained incomplete. Bone and lung occluders were preserved. The blood-particle transport, valve transfer, ECG, conduction and sound are unchanged.

The most useful next model improvement is a verified common thorax/heart pose with an expiratory or left-lateral acquisition state, plus authored LV apex, valve-ring and wall labels. These would support true surface probe placement and independent checks of foreshortening and wall orientation. A sonographer should review the resulting windows before they are used to teach standard acquisition.

Capture method

These are offline captures of the actual WebGPU shaders through native Dawn/Vulkan, not browser screenshots, patient recordings or phone benchmarks. All 12 presets are captured at two mechanical phases and at early filling/ejection, with the animated pericardium included. Before and reviewed captures use High sampling, 3 MHz, gain 0 dB, 60 dB display range, full-sector colour, 4 kHz requested PRF and a 2.5 cm/s wall filter. PRF is depth-limited by the renderer; the interactive controls additionally round their maximum down to 0.1 kHz. The reviewed/control pair uses identical geometry, probe, phase and image settings; only the thorax acoustic effects are disabled in the control. Before/after also change the selected alignments and depth compensation, so they are not an isolated test of either change.

No copyrighted reference panels are embedded in this review. No pass rate or visual resemblance score is presented as clinical validation.