Audient EVO 16 vs Focusrite Scarlett 18i20 4th gen

Short answer

  • More power: the Audient EVO 16, with 87 mW into 32 Ω; the Focusrite Scarlett 18i20 4th gen has 55 mW. The Audient EVO 16 drives 5 headphones to 100 dB that the Focusrite Scarlett 18i20 4th gen cannot (HiFiMAN HE-6se V2, HiFiMAN HE6 Remastered, HiFiMAN Susvara, 2 more).
  • Hiss with sensitive IEMs: cannot be compared — neither maker publishes a noise figure.
  • Output impedance: 41 headphones are left alone by the Focusrite Scarlett 18i20 4th gen under the one-eighth rule but coloured by the Audient EVO 16 (AKG K612 Pro, Audeze LCD-3, Audeze LCD-4, 38 more).
  • Price: the Audient EVO 16 costs $51 more; what that buys on these numbers is set out further down, under What the price difference buys.

Both are audio interfaces, $750 against $699.

The Audient EVO 16 reaches 100 dB on 558 of the 567 models here, the Focusrite Scarlett 18i20 4th gen on 553.

Every published figure, next to each other

SpecificationAudient EVO 16Focusrite Scarlett 18i20 4th gen
Power into 32Ω87 mW55 mW
Power into 150Ω49 mW27 mW
Power into 300Ω32 mW27 mW
Power into 600Ω24 mW14 mW
Output impedance50 Ω11 Ω
Output noisenot publishednot published
Price$750$699
Headphones it reaches 100 dB on
out of 567
558553

Not published means the maker leaves it out, so that axis cannot be compared.

Where each power figure comes from

LoadAudient EVO 16Focusrite Scarlett 18i20 4th gen
32 ΩDerivedDerived
150 ΩDerivedDerived
300 ΩDerivedPublished
600 ΩPublishedDerived

Both makers publish a power rating. Audient EVO 16 publishes at 600 Ω; Focusrite Scarlett 18i20 4th gen at 300 Ω. The other rows are worked out from those.

So the two columns are not equally firm row by row: at 600 Ω only the Audient EVO 16 figure is the maker’s own and at 300 Ω only the Focusrite Scarlett 18i20 4th gen figure is the maker’s own. On those rows the derived number is the one to doubt first.

What the labels mean.

The headphones that tell them apart

5 headphones reach 100 dB on the Audient EVO 16 but not on the Focusrite Scarlett 18i20 4th gen — HiFiMAN HE-6se V2, HiFiMAN HE6 Remastered, HiFiMAN Susvara, 2 more.

41 headphones are left alone by the Focusrite Scarlett 18i20 4th gen under the one-eighth rule but coloured by the Audient EVO 16 (AKG K612 Pro, Audeze LCD-3, Audeze LCD-4, 38 more).

553 of the 567 models here reach 100 dB on both.

Headroom and hiss on 8 models, both sources

Headroom is at a 100 dB SPL peak.

The widest gap on the list is the HiFiMAN Susvara: Audient EVO 16 has +3 dB of headroom there against -1 dB. 4 dB is about 2.7 times the power, which matters for peaks rather than for how loud the music sits.

The hiss columns are empty for Audient EVO 16 and Focusrite Scarlett 18i20 4th gen because neither maker publishes no output noise figure.

HeadphoneImpedanceNeedsAudient EVO 16 headroomFocusrite Scarlett 18i20 4th gen headroomAudient EVO 16 hissFocusrite Scarlett 18i20 4th gen hiss
Shure SE8469 Ω0.04 mW+28 dB damping+26 dB dampingnot publishednot published
Campfire Audio Andromeda 20209 Ω4.0 mW+8 dB damping+6 dB dampingnot publishednot published
Audeze LCD-2 Classic70 Ω0.79 mW+21 dB damping+16 dB dampingnot publishednot published
Beyerdynamic DT 1990 Pro (250Ω)250 Ω0.63 mW+17 dB damping+16 dBnot publishednot published
HiFiMAN HE400se25 Ω7.9 mW+10 dB damping+7 dB dampingnot publishednot published
Audio-Technica ATH-R70X470 Ω1.6 mW+12 dB+10 dBnot publishednot published
Sennheiser HD 600300 Ω6.6 mW+7 dB damping+6 dBnot publishednot published
HiFiMAN Susvara60 Ω50 mW+3 dB damping-1 dB dampingnot publishednot published

The gap in dollars, against the gap in numbers

The Audient EVO 16 costs $51 more than the Focusrite Scarlett 18i20 4th gen. Out of the 567 models here, that difference buys 5 more models driven to 100 dB.

It does not run one way only: the cheaper Focusrite Scarlett 18i20 4th gen leaves 41 models alone under the one-eighth rule that the Audient EVO 16 does not, which is the part worth checking against your own headphones before paying the difference.

Only published figures count here; blank axes are counted separately.

Both pages in full

↑