ZMF Atrium: how much amplifier do you need?

The ZMF Atrium is 300Ω with a sensitivity of 96.0 dB/mW, which is 101.2 dB/V. To reach 100 dB SPL — genuinely loud, louder than most people listen — it needs 2.5 mW, or 0.87 V across the drivers.

That is a moderate load. Most USB dongles handle it without complaint; the weakest phone outputs will run out of volume before they run out of quality. 266 of 280 sources here have enough headroom. The cheapest portable source that does the job is the Generic CX31993 dongles at $12.

Sensitivity here is in dB per milliwatt. Manufacturers often quote dB per volt instead and leave the unit off, which is the single most common reason online calculators give answers that are 10–15 dB wrong. The converter explains the difference.

ZMF publishes this directly as 96 dB/mW (none stated; ZMF prints '~96dB/mW'), per milliwatt, so nothing was converted.

Source: shop.zmfheadphones.com. How this data was built.

Most people listen between 75 and 85 dB. The 100 dB row is the one worth planning around: it gives you headroom for dynamic peaks without designing your system for a volume that would damage your hearing in under fifteen minutes. The safe listening calculator puts numbers on that.

Every source in the database, judged at 100 dB SPL into 300Ω. “Headroom” is how much power is left over above what the headphone needs — 10 dB means ten times the required power, which is what you want for dynamic peaks.

Hiss. At 101.2 dB/V these are sensitive enough to reveal a source's own noise floor. The noisiest source in the database, the Schiit Mjolnir 3, would put its hiss at roughly 9 dB SPL in these headphones — faintly audible in a silent room. The hiss calculator lets you check your own amplifier's noise figure.

Output impedance. At 300Ω these want a source output impedance below 37.50Ω to satisfy the one-eighth rule. Of the 192 sources that publish one, 15 miss it, the worst being the Bottlehead Crack (OTL kit) at 120.0Ω; the other 88 publish no output impedance at all, so they cannot be checked either way. Above that threshold the source starts shaping the frequency response. Check yours here.

Worked out in advance against its own stable: vs 109 Pro · vs 99 Classics · vs AB-1266 Phi TC · vs Aeolus · vs AEON 2 Closed · vs Ananda Nano · vs Arya Organic · vs Arya Stealth · vs ATH-M40X · vs ATH-M50X · vs ATH-R70X · vs Caldera · vs Caldera Closed · vs Clear MG · vs DT 1990 Pro (250Ω) · vs DT 770 Pro (250Ω) · vs DT 770 Pro (80Ω) · vs DT 880 Edition (600Ω) · vs DT 900 Pro X · vs DT 990 Pro (250Ω) · vs Edition XS · vs Edition XV · vs Eikon · vs Empyrean · vs Expanse · vs HD 560S · vs HD 600 · vs HD 650 · vs HD 660S2 · vs HD 6XX (Drop) · vs HE400se · vs K371 · vs K612 Pro · vs K702 · vs LCD-2 Classic · vs LCD-5 · vs LCD-X 2021 · vs MDR-7506 · vs MDR-MV1 · vs MM-500 · vs Ori · vs SHP9500 · vs SHP9600 · vs SR80x · vs Stealth · vs Sundara · vs Susvara · vs T50RP MK3 · vs Utopia · vs Verite Closed · vs Verite Open. Each of those pages counts the sources that drive one and not the other.

Or open the comparison tool to put any two models side by side, or browse all 578. Or see every ZMF headphone in the database, and how much of what this site says about them came from ZMF’s own documents.