Blooms, Pulse Pours and Pre-infusion

Almost nobody brews in one motion. A pour over is a bloom and then two, three or four pours; an espresso is a pre-infusion and then a ramp; a French press is a steep, a stir, and a skim. The calculator's basic model treats a brew as one contact time at one temperature, so this article does two jobs: it lays out what the research says about each stage, honestly tiered, and it explains the small corrections the calculator now applies for them.

The headline: the mechanisms behind blooming, pulsing and pre-infusion are well established, the measurements are mostly practitioner-tier, and there is no peer-reviewed sensory study of pulse pouring at all. So the calculator's stage effects are deliberately modest.

Why stages exist: CO₂, wetting and channelling

Freshly roasted coffee holds carbon dioxide that escapes over days to weeks¹. When hot water hits a dry bed of fresh grounds, that gas expands and pushes water away from the particles, so parts of the bed wet late or not at all; water then finds the easy paths, and you get channels that over-extract next to pockets that under-extract². Every multi-stage technique attacks that problem in one of three ways: get the gas out before the main brew (bloom, pre-infusion, pause), re-agitate the bed so trapped water is replaced by fresh water (pulses, stirs, swirls), or slow the wetting down so it happens evenly (low-pressure pre-infusion, gentle pours).

Pour over: the bloom

What it is. Wet the grounds with two to three times their mass of water, wait 30–45 seconds, then start the main pour.

Evidence. The mechanism is described from physics by Gagné². Barista Hustle ran a series of measured experiments: brews with larger blooms reached higher extraction, though the authors could not separate the bloom itself from the slower drawdown a larger bloom causes; and stirring the bloom extracted about 7 % of the coffee's mass during the bloom alone, far more than swirling or leaving it³. With small (15 g) doses the choice of bloom method made little difference to the final TDS; the effect grew with dose³. On freshness: the bloom matters most when there is gas to release, roughly the first three weeks after roasting¹². A 2025 trade survey notes that some competition brewers now skip a distinct bloom and use several small pours instead, treating each pour as a partial bloom⁴.

What we conclude. Blooming is a wetting and evenness technique with modest, dose-dependent effects on extraction and a probably larger effect on consistency. On coffee more than a month off roast it does little. Agitating the bloom is the strongest lever inside it.

Pour over: pulse pouring

What it is. Splitting the main water into two to five pours with pauses between them, instead of one continuous pour.

Evidence. Each pour is an agitation event and a fresh-water event, and both raise extraction⁵²; pulses also lengthen total contact time, which the time term already captures, and they let the slurry cool between pours². UC Davis's drip brewing studies used a water-on/water-off pulsing duty cycle as one of the controls to hold extraction fixed, which shows pulsing moves extraction, but the group did not publish pulsing as a studied variable⁶. Beyond that, the comparison of pulse pouring with continuous pouring exists only in trade articles and personal experiments⁷.

What we conclude. More pours means a little more extraction and, more importantly, more control: you can adjust the pace to the drawdown you see. The size of the pure "pulse" effect is unmeasured; the calculator uses small, diminishing increments.

Espresso: pre-infusion, pause and pressure

Pre-infusion is a few seconds of low-pressure (roughly 1–3 bar) water before full pressure, so the puck saturates before the pump loads it. The 2020 espresso modelling study identified channelling in fine, tightly packed pucks as the dominant source of shot-to-shot variation⁸; pre-infusion is the standard mitigation. The most careful measurements we could find are one experienced home brewer's controlled comparisons: pre-infused shots showed visibly fewer channels, shorter pre-infusions gave slightly higher TDS and extraction, and longer ones tasted marginally better⁹. That is tier-3 evidence from a single machine, so treat the direction as reliable and the magnitude as unknown.

Blooming espresso adds a pause with the pump off after pre-infusion, letting the wet puck degas and soften; shots then run faster and are usually ground finer to compensate. This is technique documented by practitioners⁷, not by a study.

Pressure. A 2002 controlled study brewed espresso at 7, 9 and 11 atm and found 9 atm gave the most consistent foam and the highest share of fresh and fruity key odorants, with 11 atm losing them¹⁰. A 2026 study instrumented a café machine and found that above roughly 5 bar, more pressure no longer increased flow, because the puck compacts under the load¹¹. And a 2023 TUM study that collected ten consecutive fractions during shots found flow rate had the strongest effect on what ended up in the cup, more so at finer grinds and higher temperatures, and showed extraction is heavily front-loaded: the first fractions carry most of the dissolved mass¹². Put together: pressure between about 6 and 9 bar is a plateau, very low pressure makes shots slow and under-extracted, very high pressure buys nothing and may cost aromatics, and the interesting variable is flow.

Immersion: stirring, breaking the crust

Barista Hustle measured stirring during brewing and found it raised extraction, with the size of the effect depending on dose⁵; the mechanism is the same replacement of saturated water at the particle surface¹³. Breaking the crust on a French press and skimming the floating grounds, then decanting without plunging, is a widely used technique⁷; it releases trapped gas and keeps fines out of the cup, and its extraction effect is small.

What is still unknown

• No peer-reviewed sensory study compares pulse pouring with a continuous pour at matched extraction.
• No peer-reviewed study isolates the bloom's effect on flavour at matched TDS and EY; the temperature study that held those fixed found little independent effect of temperature⁶, and the same design would settle the bloom question.
• Pre-infusion has one careful single-machine dataset⁹ and the modelling rationale⁸, but no multi-machine controlled study.

If you log brews with and without these stages on the same coffee, with a refractometer, the analytics page will accumulate exactly the data that is missing.

How the numbers are computed

The calculator adds a schedule term and a freshness term to the base model. Every coefficient below is a Cup Engineering working assumption: the sign follows the sources above, the size was chosen to be small because the measurements are practitioner-tier.

CO₂ remaining   g = e^(−days / 8)                       mirrors the degassing curve

bloom           +0.8 × (0.3 + 0.7 × g)                   even wetting; most on fresh coffee
                × 0.5 if bloom water < 1.5× dose
                × 0.4 if bloom < 20 s; −0.1 per 15 s beyond 60 s (slurry cools)
                + min(0.3, 0.15 × (bloom multiple − 2.5)) larger blooms extracted more (Barista Hustle)
                + 0.3 if swirled, + 0.6 if stirred
no bloom        −0.6 × g   when the coffee is ≤ 5 days off roast

pulses          +0.30, +0.25, +0.20, +0.15, +0.10   for pours 2, 3, 4, 5, 6

pre-infusion    +0.4 (under 10 s) or +0.6 (10 s and over)
pause           +0.5 if ≥ 15 s
peak pressure   −0.3 at ≤ 4 bar; −0.2 at ≥ 11 bar; 5–10 bar: 0

stirs           +0.3 each, capped at +0.8
break and skim  +0.2 (French press)

freshness       −0.4 at ≤ 2 days off roast; −0.2 at 3–4 days; 0 after

Worked example: 20 g at 1:16.7, grind 6, 94 °C, 3:20 total, coffee 7 days off roast, with a 40 s swirled bloom at 2.5× dose and three pours. g = e^(−7/8) = 0.42, so the bloom gives 0.8 × (0.3 + 0.7 × 0.42) = 0.47, plus 0.3 for the swirl; three pours add 0.30 + 0.25 = 0.55; total schedule term +1.32 points of EY. The time term adds 1.4 × log₂(200/180) = +0.21 for the longer contact time the pulses cause. Against a single-pour brew at 3:00 the prediction rises from 20.0 % to about 21.5 %, which is the right order of magnitude for what Barista Hustle measured between a plain bloom and a stirred one, and it should be read as "a bit more and more even", not as a precise number.

The pour plan (the table under the calculator's schedule controls) spaces the pours evenly across the first 70 % of the brew time after the bloom and reserves the last 30 % for drawdown. That is a sensible default rather than an optimum; competition recipes vary the spacing deliberately⁴.