design.vecreal.com / workshop / V valley · round 10
Brand reference · workshop · round 10 · 2026-05-12
V valley · round 10 — final showdown + V overshoot review
Two finalists: R4-Bw (valley y=33, stroke 12) and R4-Bhm @ y=32.75
(valley y=32.75, stroke 12.14). Both rendered at the locked spacing operator picked:
V-to-e -0.12em + dot 0.23em (margin-left unchanged at calc(0.03em+2px)).
Plus V overshoot review — is translateY(0.025em) still right given R4-Bhm's new viewBox proportions?
V overshoot math — translateY(0.025em) in context
Optical-overshoot rev-8 lock (0.025em) was calibrated for R4-A's viewBox. R4-Bhm has different proportions — same translateY produces a different V-apex position relative to baseline.
| Variant | viewBox | Empty space below curve | Apex position at translateY(0.025em) |
|---|---|---|---|
| R4-A | 10 14 44 38 | 2u × (0.66em/38) = 0.0347em | 0.01em ABOVE baseline (slight lift) |
| R4-Bhm | 9 14 46 39.27 | 1.5u × (0.66em/39.27) = 0.0252em | ~AT baseline (0.0002em above) |
| Dot (any V) | n/a | n/a (round shape) | 0.02em BELOW baseline (overshoot) |
What this means. At the same translateY, R4-Bhm's V apex sits 0.01em LOWER than R4-A's. The rev-8 history shows operator went 0.0325em → 0.025em because "feels too low at 72-96 px" — suggesting preference for V slightly above or at baseline, not below. R4-A at translateY(0.025em) sat 0.01em above baseline; R4-Bhm at the same translateY sits right AT baseline. The position has shifted closer to "at baseline" with the new geometry.
Plus: R4-Bhm has a more rounded apex than R4-A. The "pointed-shape optical floating illusion" that drove the overshoot decision is less pronounced. Less compensation needed for the rounder shape.
Section A · Final showdown — R4-Bw vs R4-Bhm @ y=32.75 at locked spacing
Both finalists with V-to-e -0.12em (tightened, operator pick) and dot 0.23em
(unchanged). Plus R4-A baseline for reference. At 80 px first, then 140 px close inspection
in section B.
Section B · Close inspection at 140 px
Same pair at display size. Red horizontal line marks baseline so you can see exactly where V apex and dot bottom sit relative to it.
R4-Bw ⇄ R4-Bhm @ y=32.75 at 140 px (baseline indicator)
Section C · V overshoot test — translateY 0.015 / 0.025 / 0.035em at R4-Bhm
Three test renders showing R4-Bhm @ y=32.75 with three V translateY values. Red baseline indicator. The middle option (0.025em, current locked) sits the V apex right at baseline with R4-Bhm's geometry — likely correct.
V translateY test at 140 px
Section D · V overshoot test at 80 px
Same three overshoot variants at 80 px for cross-size sanity check. At smaller display the difference between values is sub-pixel — overshoot decisions should be tuned for display sizes.
Spec summary — final candidates
| Spec | R4-Bw @ y=33 | R4-Bhm @ y=32.75 ★ |
|---|---|---|
| Outer x | 9 | 9 |
| Apex y | 50.27 | 50.27 |
| Valley y | 33 | 32.75 |
| Inner top (L, R) | 21, 43 | 21.14, 42.86 |
| Stroke (constant) | 12.00 | 12.14 |
| Curve dips to | 32.35 | 32.10 |
| V silhouette | 46 × 39.27 · aspect 1.171 | 46 × 39.27 · aspect 1.171 |
| V-to-e margin | -0.12em (operator pick) | -0.12em |
| Dot size | 0.23em | 0.23em |
| Dot margin-left | calc(0.03em + 2px) | calc(0.03em + 2px) |
| V translateY | 0.025em | 0.025em |
Honest verdict on V overshoot
Keep translateY(0.025em). The math shows R4-Bhm at this value
sits V apex right at baseline (vs R4-A's 0.01em-above-baseline). The rev-8 history shows
operator preferred "above or at baseline" over "below baseline" — R4-Bhm's at-baseline
position is closer to that target than R4-A ever was. The 0.01em-above on R4-A was a
calibration artifact of its viewBox, not an explicit preference.
Plus the optical-illusion reasoning. The original overshoot decision was to compensate for the "pointed-shape floating" illusion. R4-Bhm has a rounder apex than R4-A — the illusion is less pronounced, so less compensation is needed. The math and the perception both converge on keeping 0.025em.
The dot independently overshoots by 0.02em below baseline — that's unchanged (dot geometry hasn't shifted). The dot/V system remains as locked: dot *descends* slightly, V *grounds* at baseline. Asymmetric but each calibrated for its shape. The rev-8 "symmetric" framing was always approximate.
My verdict: V translateY = 0.025em locked. Don't reopen this — the math validates the existing value for the new geometry.
Final lock decision — pick the V geometry to commit:
R4-Bw @ y=33 (stroke 12, valley dips to y=32.35) — fully Bw silhouette + valley depth
R4-Bhm @ y=32.75 (stroke 12.14, valley dips to y=32.10) — same silhouette, slightly
shallower valley, slightly thicker legs
Both at locked spacing: V-to-e -0.12em, dot 0.23em, dot margin-left calc(0.03em+2px),
V translateY 0.025em.
Call R4-Bw or R4-Bhm and the rev-11 SVG regen runs in a single clean pass.
Vecreal · brand workshop · 2026-05-12 · v-valley round 10 (final + overshoot review)