Focusrite Scarlett 2i2 Gen 3 Performance Measurements

[mediatechnology]

I've been wanting to try the Focusrite Scarlett 2i2 particularly since it runs well with Ubuntu 16.04 Linux.

I'm impressed with this little box as a test and measurement interface.
It sounds good too.

The AudioTester results here are with the Scarlett 2i2 running in a Win 7 x64 environment.
The loopback connections are combo-TRS balanced. (The Scarlett's XLR inputs are the mic input.)
The sample rate is 96 kHz 24B depth.
FFT 65 kpts 10 sample avg.

The input gain of the Scarlett 2i2 was calibrated by inputting 2V RMS at the line input and then adjusting the input level until 0dBFS/clipping was observed.
2VRMS (+8.23 dBu) was chosen to approximate unity gain from the TRS input to the A/D modulator.
Virtually every 5V-powered A/D modulator's differential inputs overload at 2V RMS.
By using 2V RMS at the inputs and setting the level for 0 dBFS the overall net gain/attenuation from input to modulator is normalized to unity.
The level controls when set at this point also conveniently line up with the "1" and "2" marks on the front panel around the 11 o'clock position.
The DAC output "Monitor" gain was set to maximum. (Unity internal gain, DAC 0 dBFS is +15 dBu owing to 6 dB gain in the balanced output.)


Focusrite 2i2 Gen 3 Loopback Balanced TRS-TRS 0dBu 1kHz


Focusrite 2i2 Gen 3 Loopback Balanced TRS-TRS 0dBu 10kHz


Focusrite 2i2 Gen 3 Loopback Balanced TRS-TRS -10dBu


Focusrite 2i2 Gen 3 Loopback Balanced TRS-TRS -10dBu 10kHz


Focusrite 2i2 Gen 3 Loopback Frequency Response TRS-TRS -3dBFS


Focusrite 2i2 Gen 3 Loopback Balanced TRS-TRS Idle Channel Noise

Added 12/07/2019.

I checked the Scarlett 2i2 just below clipping and it's still excellent.
In the following two tests the output was raised to +10 dBu (+15 dBu max) and the input gain lowered to just below input clipping at +8 dBu.


Focusrite 2i2 Gen 3 Loopback Balanced TRS-TRS Output +10dBu Input +8dBu 1kHz


Focusrite 2i2 Gen 3 Loopback Balanced TRS-TRS Output +10dBu Input +8dBu 10kHz

The unweighted DR of the A/D looks to be about 104 dB.
Though its about 20 dB shy of the PCM4222's 122 dB the THD performance is outstanding.

The output clipping level (balanced) is about +15 dBu.
The Zout is about 220Ω per leg and is not cross-coupled or OutSmarts® gain-compensated.
The line input overload point is around +9 dBu.

The Scarlett 2i2 is going to be a valuable addition to the test bench and its capable of doing everything well except very low-level noise measurements.

FWIW Linux installation was painless: Plug in the USB connector and the Scarlett 2i2 pops up in Pulse Audio. You're done.

Win 7 requires installation of media pre-loaded on the Scarlett 2i2 as well as web downloads.
Win 7's forcing sample rate conversion required me to set the Scarlett 2i2 to 96 kHz even when playing back 44.1 and 48 kHz material and have those sources. presumably, up-sampled to 96 kHz. If I didn't do this, and just set everything to 44.1 to play back 44.1 files, it produced clicking.

ASIO on Win 7 with AudioTester worked well.
Haven't tried ASIO in Linux yet.

[billshurv]

That does look rather handy for the price as a bench tool.

[JR.]

Nice, thanx.

Where was this stuff back when I was assembling Heath kit test equipment....?

Once again modern technology makes life better and easier...

JR

[mediatechnology]

The Scarlett 2i2 Gen 3 is an incredible value at $159 US.

[mediatechnology]

Found the high level THD measurements done at +10 dBu out and +8 dBu in.
Posted the FFTs in the OP.

[crin]

i'm a bit late for the thread, but still … bought a 4i4/3rd gen last december as mic preamps for voice recording. after two days, the unit stopped working; it just wouldn't power up. (i used the supplied USB A > USB C cable.)

the unit was replaced by the vendor without further ado

replacement unit started powering up and down erraneously after a week or so.

the same thing happened to me a couple of years earlier with a RME fireface 400.

the culprit was the same in all 3 cases – bus power.

i returned the second 4i4 and said i wanted an interface that was NOT bus powered instead. i received a scarlett 8i6 which is wallwart-powered (the upgrade was free; i assume i was not the first customer with this kind of issues.)

no bus power, no issues.

[mediatechnology]

I haven't checked the USB current draw of the 2i2 but haven't had that problem.
I can imagine the 4i4 pulls a lot more current.

How does the 8i6 sound?

[crin]

i think it's not so much a question of current draw but of accidentally shorting the A connection. i used the scarlett with a USB 3.1 PCIe card in my old mac pro; the card features four ports, each with it's own, independent supply. however, the connectors are located very close to the metal dividers between the four PCIe slots and it's very easy to short the USB bus or damage the connector when inserting the plug. and the higher the available current (1500 vs 480 mA), the bigger the damage. but then again, i'm just no fan of bus power…

i can only compare the scarlett's mic preamps with those of my metric halo ULN-8 and my RME fireface 400. the 8i6's preamps offer ≈20 dB IIRC less gain vs the ULN-8 which can be an issue; i used a samson condenser mic and had to crank the gain almost all the way up. with the same gain, the 8i6 is quite a bit noisier than the ULN-8. with the same mic, the 8i6 sounds less transparent and darker than the ULN-8. the scarlett has an AIR titled effect which tilts the FR above 2k and adds some 3rd (focusrite does not explain how the effect works, i found out with an FFT…). with AIR engaged, voice recordings sound more lively, more open and less dark. no wonder…

i'm not complaining, quite the contrary: with 2 mic and 6 line inputs, 4 line outputs and two separate (!) headphone amps, the 8i6 offers A LOT for $200 or so! (i assume the price also reflects economy of scale; very likely the 2i2, the 4i4 and the 8i6 all use the same circuits.) i'm using the 8i6 on my work computer with a pair of focal CMS 40 and occasionally a HD 650. i have not used the line inputs.

the only thing i really don't like is the focusrite software that's neccessary to control the settings. it is completely counter-intuitive. but then again, i'm probably spoiled by metric halo's console software… the focusrite control app is pretty vague in terms of headroom (the gain halo around the knobs changes from green (good) to orange (pre-clip) to red (clip) but doesn't say where at how many -dBFS pre-clip starts and if red really means clip or close to…and it also lacks an essential feature such as a simple 80 hz highpass. which means that i first have to run the recording through a DAW software in order to eliminate the A/C hum… (i use a very simple and easy to use app for recording that offers no plug-ins, filters etc whilst recording)

[AnalogJoe]

I bought a Scarlett 2i2 Gen 3 around 3 days ago, it was $148 with shipping on e-bay. I will be using it for measuring audio equipment and to measure acoustic instrumentation via MATLAB for my PhD work, I will have to use the ASIO4ALL drivers in order for it to work properly with MATLAB, I'll report back on this. Im also thinking of buying a Cordell Distortion Magnifier to extend the range of distortion to measure really low THD like opamps and so on.

[AnalogJoe]

Got my Scarlett a few days ago, I haven't tested it to the fullest but I must say I am impressed, the noise level is around -150dBV with minimum gain, I've measured it using ARTA. The next step will be to test THD, THD+N and IMD performance. I also own a ultra low distortion oscillator, the THD+N spec of the oscillator is 0.0004% whilst the THD is 0.00005%, I'll be using it with the Scarlett to check for distortion, I wonder what the residual THD on this interface is. Either way, for $148 shipped, it is a bargain.