• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Focusrite Scarlett 18i8 Audio Interface Reviewed and Measured

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,598
Likes
34,798
Location
Seattle Area
#1
This is a review and detailed measurements of the Focusrite Scarlett 18i8 Version 2 Audio Interface (analog to digital converter). It is on kind loan from a member. The current version seems to have moved to version 3. Assuming its cost has not changed, it goes for US $297 with free shipping on Amazon.

Matching the name of the unit, this is one bright box:

focusrite Scarllet 18i8 ADC Audio Reviewed.jpg

Power is provided through an external power supply. There are a lot of inputs and outputs. I limited my testing to line inputs you see in the back:

focusrite Scarllet 18i8 ADC Back Panel Connectors Audio Reviewed.jpg
Let me tell you that I have been to hell and back on this device. I have had it for a few months trying every which way to get it working on my Windows 10 desktop. On my old system, I tried every trick in the book: multiple install/uninstall of the drivers. Different drivers. Booting, rebooting, installing, rebooting, etc. I literally spent days trying to get it work given the popularity of the brand and its low cost. I finally gave up.

Fast forward to today when I realized that it might work on my brand new Windows 10 system. I installed the drivers, plugged the unit it and to my pleasant surprise, its control panel popped up saying it detected it!

I ran my dashboard measurement and all was well. Then I noticed as I was documenting the output in photoshop that it was showing garbage data being captured. It was a sequence of sqarewaves at 750 Hz. I took a snapshot of it, put it in photoshop and then unplugged the input from it but it made no difference at all. I then hit the soft reset button in the Asio interface and bam! I was greeted with the good old blue screen of death:

focusrite Scarllet 18i8 ADC crash.jpg


Wdf01000.sys is the windows driver frame work. I looked at the unit with anger and it was as red as it could be, proving that the 18i8 was the culprit!

Using every ounce of self control I had, I hard reset the machine and booted. Sigh. Back to square one with neither Windows nor Focusrite control panel able to see the hardware. Same problem I had before. Multiple power cycles, reboot, USB input change, etc., nothing has fixed it. It has become totally non-functional as it was on my old PC.

Went through the troubleshooting page Focusrite has and everything is fine as far as services and such.

So, the measurements you see, is all the measurements I can produce.

ADC Audio Measurements
At first I started testing using the microphone input. There, even with zero gain, max input was 1.9 volt without clipping. The rear line inputs allow 4 volt input but that did not generate 0 dBFS as it should. Anything higher would make it clip so I had to go with what I could get:

focusrite Scarllet 18i8 ADC Line Input Audio Measurements.png


We have a jungle of noise and spikes everywhere in the FFT spectrum. Distortion is low however at -105 dB or so. SINAD (signal over noise and distortion) is much lower at 93 dB due to high noise floor. Here is where the 18i8 ranks among consumer and professional interfaces we have tested so far:

ADC Audio Interface Best Reviewed.png


DAC Measurements (Edit)
With my old PC I managed to make a couple of DAC measurements before it stopped working then.
Focusrite Scarlett 18i8 Audio Interface DAC Audio Measurements.png


I don't have it in the spreadsheet but it would fall just below second tier of all DACs tested.

Focusrite Scarlett 18i8 Audio Interface DAC Frequency Response Audio Measurements.png


Conclusions
I don't care what performance a device has if it crashes your system or flat out doesn't work. If I were charging even minimum wage I would not want to take jobs from clients with this type of reliability. I hear their interfaces work better on Mac and the owner says he is OK with it. But come on. I have tested countless pieces of hardware on my two machines with almost no trouble. Certainly never having blue-screens even on a brand new Dell computer and fresh Windows 10.

Besides the software issue, we have the self-oscillation or whatever it was that caused the unit to spin its wheels capturing garbage. Imagine monitoring that and having it go to your speaker at high volume. Not good. Not good at all.

Even putting all of that aside, the SINAD tells us it can't clear good old 16 bit audio. The consumer card EVGA Nu Audio runs circles around it.

Instead of making three versions of this interface, make one that works please.

------------
As always, questions, comments, recommendations, etc. are welcome.

After any of these negative reviews, I need to do something to feel better. In this case, I am thinking about getting some expensive ice cream but can't afford it. Please donate some money so I don't stay depressed for long using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
 
Last edited:
Joined
Dec 15, 2018
Messages
42
Likes
192
Location
Fremont, CA
#2
It's not *that* bad for the price, if it's just used for ADC and MIDI stuff...

At least they are honest about their specs!
 

splattened

Member
Patreon Donor
Joined
Sep 15, 2019
Messages
57
Likes
89
Location
St. Louis
#5
I have an old Scarlett 6i6 2nd gen (discontinued now I think) and recently auditioned a Clarett 4Pre USB, mainly trying them as ADC and general pre-amp for routing all my sources into my MacBook Pro. I aborted that approach because subjectively the ADC* sounded bad even when direct routing the inputs to outputs to cut my laptop out of the loop. Whether that was placebo or something more subtle going on that I wasn't educated enough to diagnose, I have no idea. I certainly wasn't expecting the Clarett to sound bad to me. I wish I was set up to ABX.

* Edit: I posted this late. And on second thought, at least in the case where I was direct routing the analog inputs to analog outputs, it isn't clear to me if the ADC was in the loop or not. I think it is though. I think everything goes through the software mixer.
 
Last edited:

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,598
Likes
34,798
Location
Seattle Area
#6
I wonder if the unit you have is broken. I've two Focusrite interfaces including a Scarlett and getting them to work with Windows has been dead easy from 7 thru 10. Did the owner get it to work?
Yes, he said it worked fine for him. Even if the box is broken, it should not crash windows. The darn thing also installs a ton of background junk on the PC include a service. Why or why???
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,598
Likes
34,798
Location
Seattle Area
#7

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,598
Likes
34,798
Location
Seattle Area
#8
It's not *that* bad for the price, if it's just used for ADC and MIDI stuff...
If it works and measures well, then it is a great bargain. It is just that it did neither for me.
 

Blumlein 88

Major Contributor
Joined
Feb 23, 2016
Messages
7,250
Likes
7,977
#10
Yes, he said it worked fine for him. Even if the box is broken, it should not crash windows. The darn thing also installs a ton of background junk on the PC include a service. Why or why???
Well I've got a 1st gen. It installed Focusrite Mix Control, and nothing else. I too hate devices that have to have a crap ton of other stuff installed instead of just what makes it work. I thought even in the 2nd gen you could install Focusrite Control skipping the bundled DAW and plug ins altogether.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,598
Likes
34,798
Location
Seattle Area
#11

Rja4000

Addicted to Fun and Learning
Joined
May 31, 2019
Messages
524
Likes
354
Location
Liège, Belgium
#13
He is using a single signal/asynchronous measurement mode:



That is the reason for the junk towards 20 kHz. There is no reason to use that measurements scheme with this device. The results are not very accurate using their method.

Using them as it is, the peak THD+N is around 0.002%.
What do you mean by "He is using a single signal/asynchronous measurement mode"
That he measures one frequency at a time?
Or that he's using a sweep?
What's best method?
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
22,598
Likes
34,798
Location
Seattle Area
#14
What do you mean by "He is using a single signal/asynchronous measurement mode"
That he measures one frequency at a time?
It is a method where you play a single signal and based on that signal, the analyzer attempts to extract a bunch of measurements at once. Signal processing is to detect that signal with accuracy, without which you get accuracy errors as shown in that graph (autocorrelation is the technical word).

It is useful/necessary when you can't control the device under test.

But if you do have control as is the case of an interface connected to the PC, then you play one signal at a time and measure what comes back with 100% accuracy. All the "sweep" graphs you see me posting are done this way.

The first method can be called "open loop" and the second, "closed loop." The best method is definitely closed loop since there is no need for cross correlation to detect the signal sequence.

The open-loop method is also used in manufacturing test because it is so fast. You play a tone and in just a few milliseconds you have a bunch of measurements you can make to give a "go/no-go" on whether the device is working. Speed is everything there and accuracy is not needed. So it is a fine method.

In our case, we want accurate measurements are not in that kind of hurry so closed loop method is by far preferred.
 

Rja4000

Addicted to Fun and Learning
Joined
May 31, 2019
Messages
524
Likes
354
Location
Liège, Belgium
#15
Thanks !

This website is one of the most famous for musicians and amateur/semi-pros SE.
I'm a long time user.

I like the fact they started to measure the interfaces with an AP, but I find really disappointing that they limit their measurements to some low-meaning ones. And this adds up to the rest.

This website, in general, is still very informative though.
 
Last edited:

restorer-john

Major Contributor
Joined
Mar 1, 2018
Messages
3,077
Likes
5,508
Location
Gold Coast, Queensland, Australia
#18
You are not alone with BSODs with the Focusrite. I have a 2i2 version 2 and it fabulous, except random BSODs during boot-up with the focusrite **usb.sys driver before Win10 has even started up. (Ver 1903)

Rare, but annoying nonetheless. I think the USB cable and/or the type B rear socket could be a culprit, but I really can't be bothered looking for the problem. A reboot fixes it for weeks. There's plenty on the internet about people with exactly the same problem.
 
Last edited:

SDC

Member
Joined
Oct 3, 2019
Messages
24
Likes
26
Location
S.Korea
#19
Guess behringer is a better choice? It's cheap, didn't kill the computer in the past measurement, so not bad?
I never saw measurements of 404hd,1820 though.
 
Top Bottom