Focusrite Scarlett 2i2 (4th Gen) Interface Review

[dasdoing]

I see no label on the output knob?
normally these are just attenuators afaik....and they are digital....so SINAD will get worse when you attenuate?
I might be wrong

 

[Smaestro]

The rear XLR is for output. The inputs that I used are on the back but are 1/4 inch. They are labeled as "Line Input."

Thank for the response, though you must be a bit mistaken!

Just to summarise what others have said, there is no line input label, the 1/4 inch line ins are on the front and the XLR’s are mic inputs.

The question is answered though, it’s the 1/4 inch line ins that were measured.

 

[Blumlein 88]

Seeing the faceplate reminds me this one has that auto gain circuit to prevent clipping. I wonder how it works (I don't know if it is defeatable), and if it effected the results.

 

[tccalvin]

Seeing the faceplate reminds me this one has that auto gain circuit to prevent clipping. I wonder how it works (I don't know if it is defeatable), and if it effected the results.

It's a button so I assume it is defeatable (on/off).

Measurements could potentially be affected by the Auto, Safe, Air (which has 2 modes afaik) and Inst buttons. Lots of room to mess up there, but I trust @amirm made sure to turn all of these off.

 

[amirm]

It is the other way 'round.

Ah yes, sorry. Either way, all the cabling went to the back so line in/output were used.

 

[tccalvin]

Ah yes, sorry. Either way, all the cabling went to the back so line in/output were used.

Huh. Now I'm even more confused. If all the cabling went to the back then wouldn't it mean the mic inputs (XLR) and line outputs (TRS) were used?

Sorry if I misunderstood.

 

[Blumlein 88]

Ah yes, sorry. Either way, all the cabling went to the back so line in/output were used.

That still wouldn't be right. There are no line ins on the back. Only microphone inputs.

 

[Smaestro]

It's a button so I assume it is defeatable (on/off).

Measurements could potentially be affected by the Auto, Safe, Air (which has 2 modes afaik) and Inst buttons. Lots of room to mess up there, but I trust @amirm made sure to turn all of these off.

Exactly.

The auto gain is a one-time detection period of 10 seconds, which sets the gain level based on the input signal.

The clip safe is a detector that turns gain down, but not back up, when a peak over 0 is detected. Whether that’s True Peak or PCM I don’t know.

Both are independent from each other.

 

[mdsimon2]

We don't know what the volume scale references. Maybe max is 0dB and what they say is 0 dB is -6 dB. Alternatively, it could be activating a secondary analog gain stage. This is why I said in the review I don't understand their pipeline.

Got it, thanks.

As @dasdoing mentioned, I assume the output knob only attenuates, at least that is how I read the function in the user guide. And it looks like full output is ~5 V based on 4 V output with -2 dBFS input from the analyzer and knob set to max, which matches the 16 dBu spec at 0 dBFS input.

Still doesn't explain the very odd drop in performance by using the attenuator. Any chance the attenuator engages some weird Windows audio system functionality which hurts the performance?

I'd definitely like to get an explanation from Focusrite as to the performance drop off.

Michael

 

[DanTheMan]

I have a 2i4 1st Gen that came out almost 15 years ago, and their drivers are still updated (the current ones came out less than 6 months ago).

What about the drivers and their stability though? For music production, drivers are 90% of the reason why one would buy an interface.

Mine was the series that came out before the scarlett--I think it was a Saffire... Apparently I bought at the wrong time. It was actually nice to use and very well made, but sadly not supported. I lost all desire to try again as the company didn't return my emails.

 

[ParadisDruid]

Regarding the HP output, Julian Krause measurements showed that the difference between the 2i2 and 4i4 is fairly significant. I wonder if there's any chance of testing a model from the 4i4 and up lineup.

 

[amirm]

Any chance the attenuator engages some weird Windows audio system functionality which hurts the performance?

I am using their ASIO driver which bypasses Windows processing.

 

[nagydann]

There is something totally wrong with Amir's tests here...

- Amir states that he used the back panel for line inputs. --- There are NO line inputs at the back of this interface...

- Amir never mentions what Focusrite Control Center application settings were and what state the different optional buttons were on the interface itself. --- There are at least 5 different variables that could lead to these results such as AIR mode (2 options), AUTO GAIN feature, INSTRUMENT mode for inputs, SAFE button or some routing options not to mention the digital volume controls and their state/position in the Control Center software.

- There are numerous Windows settings that could ruin the results as well. (Bitrate conversion from here to there, two or more ASIO drivers installed at the same time... and so on. Windows is windows and will always be...)

- This interface has a clear and dynamic sound audibly and objecively too: I don't think Focusrite as such would have dared to produce a product with these tested results. Half of the world uses this interface (or one of its iterations) and 2/3 of all the music/podcasts/YouTube that you hear on the planet goes through one of these or one of its predecessors. --- The tested results just cannot be like that.

A measurement of an audio interface like this needs more attention, more professionalism. For hifi/high-end stuff this kind of approach is enough, but more attentiveness is necessary when testing a professional audio product. In my opinion, at least.

 

[NTTY]

Ah yes, sorry. Either way, all the cabling went to the back so line in/output were used.

Hi Amir,

Which means you measured with Mic input, not line in. Back XLR (Mic in) have high gain low Z input (3k). Hence the early saturation and limited performance.

Front TRS/TS must be used for line input with “Inst” button selected for line input (led to show white) and not instrument (led would be green).

Cheers

Flo

 

[tccalvin]

There is something totally wrong with Amir's tests here...

- Amir states that he used the back panel for line inputs. --- There are NO line inputs at the back of this interface...

- Amir never mentions what Focusrite Control Center application settings were and what state the different optional buttons were on the interface itself. --- There are at least 5 different variables that could lead to these results such as AIR mode (2 options), AUTO GAIN feature, INSTRUMENT mode for inputs, SAFE button or some routing options not to mention the digital volume controls and their state/position in the Control Center software.

- There are numerous Windows settings that could ruin the results as well. (Bitrate conversion from here to there, two or more ASIO drivers installed at the same time... and so on. Windows is windows and will always be...)

- This interface has a clear and dynamic sound audibly and objecively too: I don't think Focusrite as such would have dared to produce a product with these tested results. Half of the world uses this interface (or one of its iterations) and 2/3 of all the music/podcasts/YouTube that you hear on the planet goes through one of these or one of its predecessors. --- The tested results just cannot be like that.

A measurement of an audio interface like this needs more attention, more professionalism. For hifi/high-end stuff this kind of approach is enough, but more attentiveness is necessary when testing a professional audio product. In my opinion, at least.

I don’t think you need to worry too much. There might be some mistakes in this particular review (no one is perfect), and we might have some communication issues. However, @amirm has been doing this for a long time, so I doubt he would be so careless as to overlook potential variables such as Windows settings and the front-panel functions of the interface.

 

[nagydann]

I don’t think you need to worry too much. There might be some mistakes in this particular review (no one is perfect), and we might have some communication issues. However, @amirm has been doing this for a long time, so I doubt he would be so careless as to overlook potential variables such as Windows settings and the front-panel functions of the interface.

In this particular case, he actually did overlook quite a few essential factors.

 

[AnalogSteph]

The mic input should still be able to take up to +16 dBu, the THAT6263 specifically implements some negative gain (-8 dB) to pull that off despite the limits of +/-5 V supplies. I almost cannot believe that Focusrite would have forgotten to implement this, it seems more likely that gain was simply set to 0 dB believing that was the minimum.

 

[NTTY]

EDIT: I updated this review to better align with the way I usually review devices, especially CD Players, as I described in the post “More than we hear”.

Hi everyone,

I grabbed a 2i2 Gen4 at my preferred shop. I wanted to know a little more about it, as I liked the Gen3 that I used some time ago.

Here below are measurements of the Focusrite 2i2 Gen4 DAC, performed with a Motu Ultralite MK5 and REW Software (for the measurements).

I tested the outputs in both balanced (TRS connector) and unbalanced (TS connector).

I used two computers, one with the Focusrite and another with the Motu, which allowed me to run their respective ASIO drivers independently.

Before measurements, I installed Focusrite Control 2 Software, and updated the firmware as per instructions of the user guide.


Focusrite 2i2 Gen - Measurements (Line Out - Balanced)

The following tests are performed using balanced connection (TRS) at the output.

The device outputs a theoretical 16dBu per the specs, I'll measure the reality later, but it seems to be very close. Most of the tests below were performed at 24bits/96kHz unless otherwise noted. Both channels are perfectly matched as per what I can measure (0.00dB).

Let's first start with my sine at 999.91Hz (without dither) at 0dBFS and with front volume knob at maximum):

SINAD = 107.7dB, ENOB just above 18bits.

Now the same but at -6dBFS:

The distorsion is much lower which demonstrates very good performances here. I obtained the same results with sampling rates of 48kHz, 96kHz and 192kHz.

And for the sake of completeness, this is a wide band capture (up to 96kHz) while the Focusrite is set at 96kHz:

We see noise shaping from the DAC starting at 45kHz. It's exactly the same if the Focus runs at 48kHz.

Amir showed the bandwidth and filter attenuation at 44.1kHz, as it is his standard measurement. So let me show the two at once but this time when the Foscurite runs at 96kHz sampling rate:

We see a flat bandwidth up to 40kHz. The out of band attenuation would have been the same as what Amir reported, but here we see the noise shaping technique of the conversion which rejects computational decimation errors (in the form of noise) at higher frequencies were they matter less. Nothing to worry about.

Multitone is identical to what Amir documented too (Volume at maximum):

If some of you read my CD Player reviews, you'll know I like to measure the THD vs Frequency @-12dBFS as a proof of good performances where it matters:

These are very good and results are better with less attenuation, as usual. I think this is important to know because it’s the way this DAC is likely to be used, by attenuating the output in digital domain, when using a DAW. At -12dBFS, which means 1.2Vrms output, you get extremely low distortion.

Other results (not shown) are:

• Crosstalk : -137dB at 1kHz and -129dB at 10kHz
• IMD AES DFD "Analog" : -99.1dB (18kHz & 20kHz 1:1 @-3.02dBFS)
• IMD AES DFD "Digital" : -95.8dB (17'987Hz & 19'997Hz 1:1 @-3.02dBFS)
• IMD AES MD : -108dB (41Hz & 7993Hz 4:1 @-1.68dBFS)
• IMD CCIF : -98.1dB (19kHz & 20kHz 1:1 @-3.02dBFS)
• IMD DIN : -104.3dB (250Hz & 8kHz 4:1 @-1.68dBFS
• DR : 110dB (997Hz @-60dBFS, no dither)

All of the above are very good.


Focusrite 2i2 Gen - Measurements (Line Out - Unbalanced)

The following tests are performed using balanced connection (TS) at the output.

I am adding this section as I think it is informative for the future users, when it comes to using unbalanced connection (TS cables or TS to RCA).

The output delivered an exact and expected 6dB less (10dBu, or 2.45Vrms).

This is my standard 999.91Hz @0dBFS (no dither):

Wow, my eyes are bleeding, distortion is over the roof! The resolution drops to 14bits only...

There are mains components being captured (50Hz and harmonics), but running the PC that drives the Focusrite on batteries (to mimic galvanic isolation) solved the problem.

Let's run my usual same tone at -6dBFS, with the PC on batteries:

The 50Hz and harmonics have indeed disappeared. If the situation improved on a resolution perspective, we still loose 1.5bits compared to using TRS connectors. At least the noise remains pretty much the same.

Let's have a look at my preferred measurement, THD vs Freq @-12dBFS:

At lower output, the situation improves a little with less than 10dB difference at 1kHz, but it's still 1.5bit of resolution lost when using unbalanced connectors.

I can only recommend to use balanced cables when connecting the 2i2 to whatever downstream device.


Some words about the Volume knob

If I lower the output digitally (playing my usual test tone but @-2dBFS instead of 0dBFS), with the front volume knob at max, I get:

Now, If I use a 0dBFS test tone and lower the output using the front volume knob (which is not digital) to get -2dB output, which is 4.1Vrms as Amir did in his first dashboard:

That is even worse that what Amir got. So the issue is with the front manual volume knob, not the DAC.

If I lower the volume to 2Vrms output (that means roughly -8dB), still using the volume knob and still with a 0dBFS test tone being played, I get:

So, just leave the front knob at maximum, and lower the output from your DAW or whatever software you use to play music. This is the exact same recommendation you can read in the user guide (page 19) about the output control, by the way. Now you know why.


Conclusion (Line out)

With front volume knob at max and when using balanced cables at the output, the 2i2 Gen4 nearly delivers per its published specs (THD+N = -109dB).
I found a DR of 110dB though, not 120dB as published, but it could be a limitation of my measurement interface in use here (it's not an AP).

Lowering the output in digital domain ensures high level performances for monitoring.

Just avoid touching the front knob volume or be warned that you risk not to get 16bits of resolution. Same with TS connectors, don't use them, else be ready to suffer decreased performances.

Cheers

 

[tccalvin]

Hi everyone,

I grabbed a 2i2 Gen4 at my preferred shop. I can review per my standard protocol when I review CD Players.

In the meantime, this is Line out of the Focusrite (at 48kHz/24bits) measured by my Motu Ultralite at the input (TRS cable).

Before measurements, I installed Focusrite control center 2, and updated the firmware as per instructions of the user guide. Measurements below are with the ASIO driver of the Focusrite.

First a sine at 999.91Hz (without dither) at 0dBFS and with front volume knob at maximum (which gives a theoretical 16dBu per the specs, I'll measure the reality later, but it seems to be very close):

View attachment 409891

SINAD = 107.7dB, ENOB just above 18bits.

Second, if I lower the output digitally (same tone but @-2dBFS), and still volume knob at max, I get:

View attachment 409898

Now, If I use a 0dBFS test tone and I lower the output using the front volume knob (which is not digital) to get -2dB output which puts me in roughly 4.1Vrms, as Amir did in his first dashboard:

View attachment 409901

That is even worse that what he got. So the issue is with the front manual volume knob, I'd say.

Last, if I lower the volume to 2Vrms output (that means roughly -8dB), still using the volume knob and still with a 0dBFS test tone being played, I get:

View attachment 409903


Partial conclusion (Line in)

With front volume knob at max, the 2i2 Gen4 nearly delivers per its published specs (THD+N = -109dB).
I found a DR of 110dB though, not 120dB as published, but it could be a limitation of my measurement interface in use here (it's not an AP).

I can run more tests and I'll also have a look at the inputs as well, but later (lunch time for me).

Cheers

Wow. Thank you so much for taking the time and patience to gather this data. This confirms that the volume pot is behaving weirdly by design.

Forgive me if I’m asking too much, but would you be able to take a look at these test signals on the new 2i2?

https://reference-audio-analyzer.pro/en/demo-cs131.php#gsc.tab=0

The test file is supposed to reveal the presence of clicks in a spectrum analyzer, such as the one found in Audacity.

Alternatively (AFAIU), it’s possible to detect the clicks by running a multitone test and slowly decreasing the input level. If the harmonics start "bouncing around" at some point, it means clicks are present.

In the video below, the right graph illustrates the "bouncing" starting from the 0:25 mark.

I’m only asking because the gen4 Scarletts utilize the CS43198 as a DAC chip, and I’ve personally verified that clicks are audible on its "sibling" (CS43131—see post #21). I think having this issue in a professional product would be very concerning.

That being said, I'll stop asking now since people don't seem too concerned about this issue.

 

[mdsimon2]

I found a DR of 110dB though, not 120dB as published, but it could be a limitation of my measurement interface in use here (it's not an AP).

I’ve found for dynamic range measurements it is best to use the UL Mk5 front panel XLRs. I’ve been able to measure ~120 dB DR at 4 V output. This seems to work better because the Mic input has a much lower input sensitivity than the TRS line inputs which is perfect for -60 dBFS DR test.

Michael