Cockapoo character
Model-only evidence Updated 14 July 2026
G1 Rejected v043 frozen; anatomy-first base bakeoff next

v043 bounded manual primary-form review

The sculpt moved.
The anatomy didn’t.

A focused manual sculpt pass changed local chest, neck, head, and hindquarter contours, but it did not replace the tube, pillar, and dome construction with credible canine masses. It also introduced one-sided dents and asymmetry. Gate G1 is rejected again, v043 is frozen, and this foundation will not continue downstream.

Seven fixed clay views of the rejected v043 manually sculpted nude-body candidate
Current evidence, v043. The bounded manual pass is shown in the same locked seven-view camera contract. The candidate is rejected and frozen; it is not an approved dog, likeness, or deformation mesh.
Mechanical surface
1 island
Review faces
49,624
Manual experiment
bounded
Character gates passed
0

What changed

A second attempt, with the same foundation problem

v042 proved the continuity workflow. v043 tested whether one bounded manual primary-form pass could rescue that surface. The local silhouette moved, but the construction logic did not.

Seven fixed clay views of v042 before the manual primary-form experiment
Before, v042. One closed surface after the successful union experiment, but still a rejected tube-bodied mannequin.
Seven fixed clay views of v043 after the bounded manual primary-form experiment
After, v043. Some chest, tuck, neck, and croup edits are visible. The tube, pillar, and dome read remains, with new local dents and asymmetry.
Worked

The experiment was bounded, source-preserving, mechanically auditable, and judged in the same fixed cameras.

Did not work

The manual pass did not establish ribcage, scapula, pelvis, limb chains, paws, skull planes, muzzle, or a natural neck transition.

Meaning

Freeze v043 as a failed rescue experiment. Replace the foundation through a licensed, anatomy-first candidate bakeoff.

Adversarial form review

Why Gate G1 failed

  1. 01

    Head and neck

    The pass increased separation, but the head still reads as a dome over a pipe muzzle. Stop, orbit, cheek, jaw, mouth corner, nose plane, throat, and ear roots remain unresolved.

    critical
  2. 02

    Shoulder, chest, and torso

    The chest and tuck moved locally, but the torso remains a nearly uniform tube. Ribcage, scapular plane, withers, upper arm, pectoral mass, waist, and loin do not form a coherent hierarchy.

    critical
  3. 03

    Pelvis, haunch, and hind chain

    The croup and thigh changed slightly, but the rear remains mechanical and high-folded. Pelvis, gluteal, thigh, hamstring, hock, and tail-base relationships are still not credible.

    critical
  4. 04

    Limbs and paws

    Forelegs still read as posts, hind legs as rods, and paws as tiny flat pads. The manual pass did not establish joint direction, weight support, or believable contact.

    severe
  5. 05

    Surface and overall read

    The mass distribution remains generic, while one-sided dents and asymmetry add noise without anatomical information. Cockapoo identity work still cannot begin.

    severe
v043 proves that a bounded manual sculpt was attempted and reviewed. It also proves that this construction is not worth carrying into another rescue pass.

Mechanical truth

Closed is not production-ready

v043 still has one island and no boundary or nonmanifold edges. Its uniform remesh density remains useful only as broad sculpt clay; it carries no deliberate deformation flow around the face, shoulder, hip, paws, or tail base.

This is disposable sculpt clay. It is not curated topology, rig-ready geometry, or a quality approval.

Seven fixed wire-overlay views showing the dense uniform v042 remesh
v042 wire evidence, still representative of the foundation. v043 preserved the same dense mechanical surface. No deformation-flow claim is made.

Workflow video source

The process was captured and privacy-reviewed

The v043 five-minute manual attempt and the prior v042 union work were captured in guarded Blender-window clips. The v043 review covered a sixteen-sample contact sheet plus start, middle, and end frames; v042 added forty reviewed samples. No photos, people, messages, unrelated apps, or private URLs were visible. The public package keeps only compact contact derivatives; raw recordings stay private.

Sixteen sampled Blender frames from the bounded v043 manual primary-form sculpt
v043 manual experiment. Five guarded minutes of broad form edits. The capture includes an accidental 7544-pixel brush-size entry and immediate recovery; no giant-brush mesh hit occurred. Useful as tooling-recovery footage, not a successful sculpt demo.
Sixteen sampled Blender frames covering union, continuity, masks, and the shoulder-neck stage
Part 1. Source, native union, continuity, frozen masks, and shoulder-neck stage.
Fifteen sampled Blender frames covering muzzle, croup, tail root, and the mechanical audit
Part 2. Muzzle, croup, tail root, and audit. A camera-toggle error made several late review frames too small.
Nine sampled Blender frames from the corrected locked-view recording coda
Part 3. Corrected idempotent locked front-three-quarter and right-profile coda.

Accepted use: future workflow-video editing, including an honest failed-form experiment and recoverable UI error. Not accepted as: form, likeness, topology, rig, groom, animation, or final-render approval. Raw recordings are intentionally not in this public package.

Gate truth

Everything after anatomy stays blocked

Source lineageretained evidence

v041, v042, and v043 remain frozen evidence for construction, union, and the failed manual rescue attempt.

Continuity methodmechanical pass

v042 and v043 are one closed surface with clean structural receipts. That proves continuity only.

G1 anatomyrejected

v043 still does not read as a credible canine body and will not receive another rescue pass.

G1B likenessnot evaluable

Identity cannot be judged on unresolved anatomy.

Topology and rigblocked

No deformation cage, production loops, controls, or weighting are authorized.

Groom, motion, finalblocked

No fur, animation, or polished render may conceal the failed nude-body gate.

Next bounded experiment

Bake off anatomy-first canine foundations

The manual rescue experiment has reached its stop rule. The next run compares licensed, traceable canine bases under the same nude-body camera contract before any foundation is chosen.

  1. 1

    Source and license candidates. Record origin, license, author, editable source, topology, and intended use before importing anything.

  2. 2

    Compare nude anatomy. Render every candidate in the same profile, front, rear-three-quarter, and top cameras with no fur or presentation tricks.

  3. 3

    Separate appearance from production value. Score silhouette and anatomy first; inspect deformation-critical topology and rig suitability separately.

  4. 4

    Choose or reject explicitly. A generated mesh or old model remains evidence unless it materially beats v043 and passes the same human G1 review.

Stop rule

Do not continue from v043. If no candidate establishes a credible canine nude body across all locked views, reject the set and improve the candidate pool rather than hiding defects with fur.

Rejected historyOpen the earlier prototype and retrospective

These videos remain collapsed as historical proof of an earlier, rejected workflow. They do not show v043 and they do not represent the target character quality.

Contact sheet of the earlier rejected seated prototype
Rejected curiosity loop. Reviewable delivery plumbing, unacceptable character form and motion.
Legacy retrospective. A historical stage recap, not continuous proof of the current build.

Evidence downloads

Machine-readable truth

These files are path-free or reviewed as public-safe. No reference photos, raw recording, source scene, or exported mesh is included.