Tool Comparison

What is being compared

All three tools read the same characteristic curve and report the same kinds of numbers: where the film starts responding (speed), how steeply it climbs (contrast), and how far it runs (range). They differ in resolution and in a few definitions. The point of this document is to show, on real data, exactly where those differences land and which number to trust for each job.

Dataset and the 31 to 11 reduction

The tablet is sampled at 1/3-stop steps (0.10 log H). ZoneLab works in whole stops (0.30 log H), so to feed it an 11-zone test we take every third step. Zone 0 maps to base plus fog (log H 0), Zone X to maximum exposure (log H 3.0). The full curve and the 11 sampled points are plotted below: the rings sit on the lines, so the reduction itself loses almost nothing of the curve's shape.

Net negative density against relative log exposure. Lines are the full 31-step measurement at each development time; rings mark the 11 whole-stop zones ZoneLab actually evaluates. Longer development lifts and steepens the curve.

The reference columns in the analysis were checked against the compiled Swift engine on these reduced arrays. Contrast Index and gamma matched to three decimals at all five times, so what follows reflects the real engine, not an approximation of it.

Where they agree: contrast and speed

Contrast Index is the headline agreement. ZoneLab's 11-zone arc CI is identical to the Analyzer run on the same reduced 11 points: ZoneLab's CI is the concentric-arc method, not a lookalike. Compared against the full 31-point fit, the only daylight is at 4 minutes, and even there the gap is 0.027. By 8 minutes it is exact.

Contrast Index by development time. The delta above each pair is ZoneLab minus Analyzer (full 31). The whole-stop reduction costs at most 0.027, at the lowest-contrast end.

The speed point agrees just as tightly. At the Lambrecht threshold (fog + 0.17), all three tools land within 0.017 log H of each other — about six hundredths of a stop — across every development time.

Speed point (log H at fog + 0.17). The three series sit almost on top of one another at every development time.

Where to be careful: gamma on steep development

Gamma tracks well through moderate contrast and then breaks at the steepest development. At 16 minutes the 11-zone gamma overreads by +0.288 compared to the full-resolution straight line. This is a resolution artifact, not an engine error: the same method on the full 31 points behaves.

Straight-line gamma by development time. Agreement holds through 11 minutes, then the coarse 11-zone fit overreads sharply at 16.

The cause is visible on the curve itself. At 0.30 log H spacing, the best four consecutive zones can land entirely on the steepest part of the climb. The slope of that short segment is far steeper than the honest straight-line fit across the whole usable curve. Contrast Index, which integrates the arc rather than chasing the steepest segment, does not have this problem.

The 16-minute curve. The four filled zones are the steepest 4-zone run; their slope is what ZoneLab's auto-gamma reports. The dashed line is the full-resolution straight-line fit — visibly shallower.

Average gradient: a definitional drift, not a fault

ZoneLab's gradient field is the slope across the fixed Zones II to VIII window. Lambrecht's average gradient is rangeD / rangeH over the useful curve. Those are two different spans, so the numbers match mid-range and drift apart at the toe-heavy and shoulder-heavy extremes — by about a tenth.

Fixed-window gradient (ZoneLab) against average gradient (Lambrecht). Matched mid-range, off by about a tenth at the 4 and 16 minute ends.

Effective ISO: the Delta-X shadow speed

A speed point read at a fixed density (base + fog + 0.10) is only a true speed point at the ISO-standard contrast. As development pushes the curve steeper, that fixed point slides left faster than the film's real speed gains, so a naive fixed-density reading overstates how much speed you pick up from extended development — and understates it when development is cut short. This is the well-documented limitation behind the Delta-X criterion.

ZoneLab corrects for it. Instead of reading speed straight off the 0.10 point, the engine locates the fractional-gradient (Jones) speed point — the exposure where the curve's local slope first reaches one third of its average gradient — and references it to the standard ~0.29 log H offset. At normal development the correction is essentially nil, so ZoneLab still agrees with fixed-density tools there; away from normal, it tracks the real speed.

The uncorrected fixed-density shadow speed swings 2.5 stops across these five development times (EI 11 to 64). The Delta-X correction compresses that to under one stop (EI 25 to 47) — the film's speed is far more development-independent than the fixed-density point alone suggests, exactly as the fractional-gradient method predicts.

The dual-speed model still holds: ZoneLab reports both a Delta-X shadow speed and a Zone V speed and recommends the lower of the two. At 4 minutes the curve is too flat for the Zone V grey-card target to fall on the measured range, so the Zone V speed craters and recommended = min correctly defers to it. Everywhere else the recommendation now follows the corrected shadow speed rather than the inflated fixed-density figure.

These are real measured TMax-100 densities. The absolute ISO numbers are nominal only because the exposures were placed via step tablet rather than an in-camera meter, so box speed stands in as the metering reference — read them as relative behavior between methods, not as a film-speed verdict. The point is the shape: Delta-X flattens the development dependence of shadow speed, which is what a correctly located speed point should do.

Definitional differences (not bugs)

Two divergences look like disagreements but are just different definitions of the same word. Dmax splits for the 11 and 16 minute curves because ZoneLab reports the literal density maximum the curve reaches, while Lambrecht caps Dmax at the useful range. Where the curve maximum equals the useful maximum — at 4, 6, and 8 minutes — they agree exactly.

Maximum negative density. Identical through 8 minutes; past that, the curve runs beyond Lambrecht's useful cap, so the two definitions part by design.

One more note for the record. An earlier comparison document reversed the curve, mapping Zone 0 to maximum density instead of base plus fog, which inverts the whole characteristic curve. That document's conclusions are invalid. The analysis here uses the correct orientation throughout.

Bottom line

On this film at five development times, ZoneLab's 11-zone engine reproduces the reference tools almost exactly for Contrast Index, speed-point location, fog, and Dmax. The reduction from 31 steps to 11 costs at most 0.027 CI units — less than a third of a stop's worth of contrast error.

It also makes one deliberate departure: the Delta-X correction on shadow speed. A naive fixed-density speed point exaggerates how much film speed moves with development; ZoneLab references the fractional-gradient point instead, which flattens that swing from 2.5 stops to under one. This is ZoneLab choosing the more accurate speed over the easier one — and it's why its shadow EI will differ from a tool that reads speed straight off the 0.10 (or 0.17) point at non-standard development.

Two things still read as in-tool values rather than cross-tool comparisons: auto-gamma overreads on steep development (a resolution artifact worth up to about 0.29 at 16 minutes), and the fixed Zones II–VIII gradient drifts roughly a tenth from Lambrecht's average gradient at the extremes. The rule: