Audio Interface advice and recommendations

[Andy N]

Reaching out for some collective wisdom.

I’m thinking about upgrading my Audio Interface and looking for suggestions/recommendations.

I’ve been using a Focusrite Scarlett 2i2 3rd generation for several years now and been reasonably happy with it. However I’ve been disappointed with recordings I’ve been making for this month’s challenge. I’m using more gain than I normally use and the recorded signal doesn’t compare well to the sound I’m hearing from the guitar/amp when making the recording. The recorded signal sounds clipped, compressed, a bit quacky and no where near as rich harmonically. I’ve been careful to make sure that the signal isn’t too loud and only peaking to around 12db so that’s not the cause.

I’ve tried recording with different sources, e.g. a mic’d amp vs amp modellers. Even into different DAWs but all seem to exhibit the same characteristics albeit to different extents. So I’m beginning to suspect the Scarlett as being the cause.

I’ve read the Focusrite Scarlett 2i2 4th Gen uses an upgraded ADAC chip, the same one as used in their higher end models which would make it a worthy upgrade.

Has anyone been able to directly compare a 3rd Gen 2i2 to its 4th Gen cousin?

Are there any alternatives that others are using at a similar or slightly higher price point? I’m happy to spend more on a good solution within reason.

Have any of you noticed similar problems and know a way to improve the recorded quality?

I’ve noticed my sample rate is 44.1khz so I will experiment increasing this but I’m not convinced that alone will resolve the problem. Do people use/recommend higher rates e.g. 48k or higher?

Andy

[sunjamr]

Hi Andy – I’ve got 3 different audio interfaces, including the world’s most popular one: The Scarlett 2i2. I can’t tell any difference between any of them, to be honest. What does an interface actually do? I reckon its basic job is to just convert an analog sound input to a digital sound input, which is something sound engineers have been doing for over 70 years.

But I have certainly noticed the very issues you mention, and it only happens during direct input recording. For me, if I want the highest quality recording, I do have to mic my amp, either a tube amp or a modeling amp. What about this: The modeling amp is converting my analog guitar input to digital sound, which it then feeds to the speakers. My tube amp is just taking my analog guitar input, amplifying it, then sending it to the speakers as analog sound. So if I put a mic in front of the modeling amp, that mic is collecting a stream of digital 1s and 0s and sending them off to my Scarlett 2i2, which then wants to rearrange that sequence and forward it on to my DAW. But if I mic my tube amp, it puts out an analog signal, which is fed to my Scarlet, which converts it to a digital signal. And that’s what the Scarlett was designed to do.

Anyway, if you do decide to try out a different audio interface, be sure to let us know how it works out. More than 90% of all the musicians I know use the Scarlet 2i2.

Sunjamr Steve

[Andy N]

Many thanks for your thoughts Steve.

sunjamr wrote:

So if I put a mic in front of the modeling amp, that mic is collecting a stream of digital 1s and 0s and sending them off to my Scarlett 2i2, which then wants to rearrange that sequence and forward it on to my DAW.

I’m not sure that’s right. Mic’s and Speaker are analogue devices so I think the Scarlett in this situation is going to be doing exactly the same as it is for a Mic in front of a Tube amp speaker. Taking the analogue Mic signal and converting it to digital. The modelling amp, if it’s digital internally, would be converting to analogue before sending sound to the speaker.

There’s definitely a difference between direct inline and Mic’s though. I think the Mic signal goes through a preamp within the Scarlett which must potentially have an impact on its digital output to the DAW. I’ve had mixed results. Sometimes inline is better and sometimes the Mic, so I think the nature of the analogue signal being converted is a part of the puzzle.

The digital output from an audio interface is, by its very nature, always going to be an approximation to the original analogue. The more processing power and data handling capacity we can throw at the conversion the closer that approximation can get. The question is at what point does the digital approximation get so close that we can’t hear the difference between it and the real (analogue) thing? The fact that you’ve got three and can’t hear a difference suggests we could already be at that point but it’s possible to spend 1000s on an audio interface which begs the question why? Is it just more features? Marketing or do they do a better job!

Will certainly come back and post an update it if I upgrade and can do some back to back tests 👍

Cheers
Andy

[Andy N]

Just watched this video and learnt way more than was necessary about audio interfaces but interesting nonetheless.
It’s long so the executive summary is that with very few exceptions it doesn’t really matter which one you have.
For home recording they all do a good job and you’ll probably struggle to tell the difference. That matches your experience Steve 👍

It does mean the Scarlett probably isn’t the cause of my recording issues. One thing I was wrong about is that the inline/instrument inputs do go through the preamp, exactly the same as the Mics do. So there is really no significant difference between the signal path of both types of input.

[Jean-Michel G]

The signal that comes from a mic is ridiculously low – and it’s called mic level.
The one that comes from an instrument (electric guitar or active piezo pickup) is also rather low, although significantly hotter than mic level – it’s called instrument level.

In order to be processed by whatever other piece of equipment in the signal chain, these signals need to be raised to a voltage level called line level. That’s essentially the job of the preamps. Please note that line level is much higher than either mic level or instrument level, so the gain required by the preamp is significant.
Given the difference between mic level and instrument level, you don’t need the same preamp gain to reach line level; maybe that’s something you need to check in you gain staging. Too much gain on a hotter signal may cause distorsion.

Top quality preamps are able to achieve this very high gain (60 dB and more) without compromising the S/N ratio – which is quite an achievement. Another usual expectation from preamps is that they be as linear as possible.
It’s very easy to tell the difference between a high end preamp and an entry level one when you are given the opportunity to compare them side by side. Also, some specific usages expect the preamps to color the signal (affect the harmonics) in one way or another.
All together, this is why top end preamps easily cost a fortune.
But you are right: the preamps in the Scarlet interfaces are fit for the job, even if they don’t match the top end.

The sampling rate affects the A-to-D conversion within the interface, after the preamps have done their job. The higher the sampling rate, the better the signal reconstruction after sampling – in theory. In practice, the processor needs to be able to keep up the pace. You easily get all sorts of cracks and pops and drops if you go wild with the sampling rate.
A higher sampling rate is not going to change anything in your case, I think.

I personally use a Scarlet 18i8 3rd gen (which probably use the same preamps as the 2i2) at a sampling rate of 44.1 kHz, and I am also reasonably happy with it.

[Jean-Michel G]

So if I put a mic in front of the modeling amp, that mic is collecting a stream of digital 1s and 0s

This is not correct. A mic can only capture an analog signal, and whatever audible sound comes from an amplifier through speakers or monitors has been converted to analog.

[snakechisler]

I have a 212
I record with the input volume set fairly low to capture dynamics and harmonic output to avoid clipping of any sort and use my mixing software to digitally amplify the track to the right level for the mix.

I also use Audacity which is freeware and is written in C+ and has a small footprint to record any tracks

If you have a windows computer to record you may be suffering with software issues caused by resource management , my PC is a simple desktop while it has plenty of memory and an ssd + graphics card it still would struggle to run some of the extremely boated software out there in record mode.

If your on Apple and GarageBand ignore the software comment I’d still record low and digitally amplify though

[Andy N]

Hi Greg,

That’s interesting, I’d been doing a bit more research too and seen this recommendation elsewhere. Tried a test today and there’s definitely an improvement between recording ‘loud’ and recording ‘quiet’.
I think there may be some more experimentation to do to find the sweet spot though.

For ‘loud’ I recorded with the DAW peaking at around 0db but making sure it didn’t go over that. The Scarlett itself is giving me greens for recording so it shouldn’t be clipping. That in itself is a bit odd as peaking at a 0db should have the Scarlett on bordering Red and definitely in Yellow territory.

For the ‘quiet ‘ I had the level in the DAW between around -18db and -12db which is the both the level I’d seen recommended on the WibblyWobblyWeb and the levels the 4th Gen Scarlett will attempt to target with its Auto Gain facility (my Scarlett is a 3rd Gen so doesn’t have the feature).

Given the oddness around 0db and my ‘loud’ recording, I’m wondering if the levels reported by the DAW and what the Scarlett thinks it is, is different, hence the need for a bit more experimentation.

The Windows thing is not an issue for me, I switched to Mac as soon as I retired and could ditch Windows. It definitely was a pain when I was using Windows. Recording signals would drop all the time as well as odd clicks and pops when Windows would go off and harness gobs of CPU for some housekeeping other.

[Jean-Michel G]

Recording “hot” or “loud” is a thing of the past, when sound engineers had to fight the noise coming from analog devices by achieving a high S/N ratio.
In the digital world, noise has become almost completely irrelevant. The big problem in the digital world is clipping.

As soon as a Scarlett 3gen input level turns green, that means you have reached a signal level of -24 dBFS; that corresponds to 0 dBu and basically says that the signal is audible, that you do record “something”.
As you keep on increasing the input gain on the Scarlett, the visual control will still report green, but you should see the level increase in the DAW.
When the Scarlett visual signal turns red, you have reached the point where the digital signal gets clipped – so you should stay below that level and make sure that the loudest parts of the signal never enter the red zone.

When you reach -20 dBFS in the DAW, you are at +4 dBu, which is the nominal level. On an analog VUmeter you would read 0 VU.
At -12 dBFS you are at +12 dBu, which is the nominal peak level.
At 0 dBFS in the DAW, the signal gets clipped.

When recording, input levels on each track should be around -18 to -12 dBFS to provide enough headroom for when they all get summed into the stereo bus. I personally often keep them at around -20 dBFS while still maintaining an adequate dynamic range.

Keep in mind that with 24 bits you have a dynamic range of about 1500 dB – that’s more than adequate to encode even the quietest parts of the signal. So in the digital world, recording you be at low levels, never at high levels.

[Andy N]

Hi Jean-Michael,

Many thanks for the summary, that brings a lot of clarity and that ties up with my experiments except for one thing. The thing that still confuses me is the difference between 0db in the DAW and what the Scarlett thinks is 0db and turns its Red clipping warning on. The fact that there’s two different units (dbFS and dBu) might be the cause of this confusion and is why I’ve deliberately used ‘db’ in the second sentence above.

From the documentation, the Scarlett will clip and turn the halo red at 0db which I’m assuming is 0dbFS?
But at that point the meter in the DAW is well above 0db. I’m assuming that too must be 0dbFS since 0dBu from your description above is actually very quiet and barely audible.

If I adjust things so that the DAW is hitting 0db(FS?) then the halo on the Scarlett is green and must be well enough below that level or it would be blipping into yellow at the very least.

So I can’t square that circle in head. The DAW doesn’t seem to be measuring what I think it should be.

But that aside, I’ll definitely be recording at lower volumes in the future.

One other thing, when I say the recording sounds ‘clipped’ above. It doesn’t sound clipped in the very recognisable way that a signal sounds clipped when the Scarlett is showing red. It’s more that the recorded sound lacks detail and the transients are getting muted. It still may be clipping but it’s a much more subtle sound.

Cheers
Andy

[Gary W]

Hi Andy. there’s so much to experiment with on this topic & managing gain is a very engaging aspect. I had your interface & upgraded to the Clarret (4 Pre) s couple of years ago. I’ve found this to be superb. The quality of the preamps is better. Tbh I mainly upgraded for additional inputs and double headphone monitoring as i was recording a bit with others at the time. The advantages in sound quality (in what’s essentially just another sound card/converter) has been noticeable – I think this is down to pre-amp quality, but i fully take on board the probability that at my ‘consumer level’, there’s little difference to be had.

Gary

[Andy N]

Many thanks for the info Gary. That’s very helpful.

If I could achieve recordings of the quality of yours, I’d be a very happy bunny! I feel I’ve still a way to go to drag myself up the learning curve.

I’ll take a look at the higher end Scarlett models as from the discussion so far, the quality of pre-amps is coming out as the single most significant factor. I’d thought it might be the quality of the chipset doing the actual Analogue to Digital conversion but seem to be a secondary factor.

I think I will invest in an upgrade, I don’t need lots of inputs and I know the 4th Gen 2i2 shares the AD/DA chipset with the higher end models. I don’t know if that extends to the pre-amps so I’ve still a bit of homework to do!

[Bill]

I’ve been happy with my UAudio interface. It competes in the same space as the Scarlett.

[Andy N]

Many thanks Bill, I’ll take a look at that too. Half the fun of shopping is looking up all the options!

[Author]

Posts