• 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 Clarett 2Pre USB - quick measurements

Joined
Dec 15, 2019
Messages
5
Likes
17
#1
Up until recently I was using the Digilent Analog Discovery 2 to perform audio measurements on my DIY audio projects - with 14-bits of resolution, it's a surprisingly powerful and cost effective little device. And while still an exceptional and blisteringly fast scope - I honestly can't see myself using a "traditional" oscilloscope when these insane USB devices exist! - the bottleneck is the noise floor, which doesn't get much lower than around -80dB. So while I could verify that my projects were relatively clean, I couldn't really plumb the depths of their noise characteristics.

I recently picked up the Clarett 2Pre to take over FFT/spectrum duties, freeing up the Analog Discovery as a full-time scope (and logic analyser) and giving me much more dynamic range to test out designs.

Apart from the abysmal software - "Focusrite Control" - I couldn't be happier. Great build quality, and effortless to integrate into my setup. My chain is as follows:
DUT -> Dummy Load -> Passive Attenuator -> Clarett 2Pre

While I'm still fairly new to best practices regarding sound interfaces and loopback measurements, here are some quick results:

Setup

  • Balanced 1/4" to balanced 1/4" cable from Line Out 3 to Input 1
  • Focusrite Control set to Direct Routing (the only way I could find to prevent a feedback loop - still getting to grips with the software)
  • Line Output at -6dB in Focusrite Control
  • Input pot at 10-o-clock
  • Mono only - didn't have a second cable on hand :)

REW @ -6dBFS 1KHz sine:

REW Loopback Clarett 2Pre USB.PNG


That third harmonic hops around a little, somewhat larger than I'd like. Although massive nitpicking (it's ~125dB down!).

And for what it's worth, RMAA:

RMAA Loopback Clarett 2Pre USB.PNG


fr.png


noise.png


Very happy with the results! Not quite the rock-bottom distortion figures seen elsewhere in the Clarett range (I've seen -116dB noise figures), but an absolutely razor sharp frequency response. Lovely.

If anyone has any tips about how best to perform loopback tests, software choices, setup (particularly the godawful Focusrite Control software) I'd appreciate it!
 

pkane

Major Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
1,485
Likes
2,227
Location
North-East
#2
Up until recently I was using the Digilent Analog Discovery 2 to perform audio measurements on my DIY audio projects - with 14-bits of resolution, it's a surprisingly powerful and cost effective little device. And while still an exceptional and blisteringly fast scope - I honestly can't see myself using a "traditional" oscilloscope when these insane USB devices exist! - the bottleneck is the noise floor, which doesn't get much lower than around -80dB. So while I could verify that my projects were relatively clean, I couldn't really plumb the depths of their noise characteristics.

I recently picked up the Clarett 2Pre to take over FFT/spectrum duties, freeing up the Analog Discovery as a full-time scope (and logic analyser) and giving me much more dynamic range to test out designs.

Apart from the abysmal software - "Focusrite Control" - I couldn't be happier. Great build quality, and effortless to integrate into my setup. My chain is as follows:
DUT -> Dummy Load -> Passive Attenuator -> Clarett 2Pre

While I'm still fairly new to best practices regarding sound interfaces and loopback measurements, here are some quick results:

Setup
  • Balanced 1/4" to balanced 1/4" cable from Line Out 3 to Input 1
  • Focusrite Control set to Direct Routing (the only way I could find to prevent a feedback loop - still getting to grips with the software)
  • Line Output at -6dB in Focusrite Control
  • Input pot at 10-o-clock
  • Mono only - didn't have a second cable on hand :)

REW @ -6dBFS 1KHz sine:

View attachment 43810

That third harmonic hops around a little, somewhat larger than I'd like. Although massive nitpicking (it's ~125dB down!).

And for what it's worth, RMAA:

View attachment 43811

View attachment 43812

View attachment 43813

Very happy with the results! Not quite the rock-bottom distortion figures seen elsewhere in the Clarett range (I've seen -116dB noise figures), but an absolutely razor sharp frequency response. Lovely.

If anyone has any tips about how best to perform loopback tests, software choices, setup (particularly the godawful Focusrite Control software) I'd appreciate it!
Looks very good! I prefer to use a different FFT window than Hann (Blackman-Harris 7 or Chebyshev 200). You may also want to set a low pass filter to, say, 22.05kHz or 24kHz, so that the statistics reflect the audible range, excluding ultrasonics. Since you are using a high sampling rate, you can also try to increase the FFT size to dig further into the noise floor showing more of the structure that may be hiding below.

Agree about Focusrite software -- took me days to figure out how to set it up :)
 

Panelhead

Active Member
Forum Donor
Joined
Apr 3, 2018
Messages
279
Likes
101
#3
The results do look very good. The 4Pre I use has line inputs and mic (adjustable gain) inputs.
Did my testing with REW. The REW test signals look different than using analog signals. The analog are lower THD.
Compared to the published specs does it match or exceed?
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
9,980
Likes
8,411
Location
Riverview FL
#4
The REW test signals look different than using analog signals. The analog are lower THD.
Did you back off the REW generator by 0.2dB or run it full scale?
 

Panelhead

Active Member
Forum Donor
Joined
Apr 3, 2018
Messages
279
Likes
101
#5
I used many REW output levels. Ran the output stage wide open. Since the Clarett 4Pre has fixed gain line inputs the level of the REW signal and the analog sine wave generator was varied to set output level.
Tested the Clarett first, many variables.
1. Power supply - stock switcher, iFi iPower, Jays linear, a crappy beast 7 amp variable supply.
2. Thunderbolt cables - Apple, generic cables, Corning fiber optic.
3. Interfaces - stock, modded one with decoupling caps and FET input opamps.
4 Input - REW signal generator and a 2 ppm 1KHZ analog generator from Akita.
After so many tests I was totally confused started on amplifier testing. 6 ohm non-inductive load and all power levels.
Since my system has high efficiency horns, crossover and a sub the levels of interest were really less than a watt.
Ran those with several amps. Then replaced the capacitor coupled single ended inputs with balanced to single ended input transformers with -12 dB gain in two amps and reran the testing.
Tested out. Hide the test cables. Gone back to listening.
No desire to restart testing, cannot stop.
Lessons learned.
REW is a great software package. RTA is a throw in. Made three donations to them, it is free. I hope Amir does not read this.
Clarett is not SOTA, but I read here that 16 bit performance good enough. Way exceeds that. More like 19 - 20 bit for ADC and DAC.
My amps can show -130 dB noise baseline and measure -115 dB THD+N. Even with 105 dB efficient horn it is dead silent with ear in the mouth of all three horns.
And the stock Clarett measured better than the one with 20 NJR5532 replaced with OPA2604. Just Different THD spectrum. The copper TB cable measured the same as the fiber optic cable with complete galvanic isolation, the iFi iPower was slightly cleaner than the Focusrite supply. But a dB or so. Never hear it. Both tested better than the two linear power supplies. And Cinemag input transformers handle high signal better than Lundahl transformers
Time to go play some tracks.
 

nanook

New Member
Joined
Mar 31, 2019
Messages
2
Likes
0
Location
Berlin, Germany
#6
I'm thinking about the Clarett 2 Pre or Clarett 4 Pre as an audio interface for my hobby electronics. Currently I'm using a really cheap interface that only has RCA- in- and outputs and that has a noise-shaping that significantly raises the noise floor above 50kHz when operated at 192kS/s. I'm using REW; this software does all I need, specifically sweeps and the RTA with a lot of features like calibrated "noise voltage density" vertical scale, a flexible multi-tone generator and "coherent averaging" to dig deeper into the noise floor when analyzing distortion.

Can somebody help me with these questions?
1) Loopback Frequency response:
I'm wondering if Focusrite used the steep or slow roll-off for the ADC. Also it would be interesting if the analog DAC filter already introduces quite some attenuation at say 80kHz and if the high-pass filter of the ADC- buffer is reasonably low for a measurement setup.
--> Did anyone run a loop-back sweep at 192kS/s (0Hz ... 96kHz)?
2) Noise spectrum (Line input, shorted, min. gain, 192kS/s, 2Hz ... 96kHz)?
--> Does anyone have a noise voltage density plot?
3) ADC and DAC chips used in Clarett 2 Pre ?
4) ADC and DAC chips used in Calrett 4 Pre ? (I read somewhere that it uses 2x AK5388 ADC and 1x AK4413 DAC -> is this right?)
(I'm fully aware, that this is only part of the whole story, but that would allow me to look up the digital filter frequency responses, the digital High-Pass- filter and if these can be changed via a dedicated pin)
5) The inputs 1+2, resp. 1-4 seem to be Microphone inputs (XLR, differential, low input impedance), Line inputs (TRS, differential, relatively low input impedance) or Instrument (TS, single-ended, high impedance) -> is this right?
6) Clarett 4 Pre: are the Line- inputs 5-8 the same quality as the Line inputs 1-4 (except that they have a fixed gain)?
7) The gain pots, do they have a mechanical detent (-> reproducible gain settings)?
8) Does Focusrite provide ASIO drivers (I use these within REW)
9) Does the driver run stable (the RTA in REW seems to be quite demanding)

Last not least: I'm open for suggestions in case there is an audio interface around with a similar price- tag, that is suited better for this task. The RME ADI-2 PRO fs would be a very nice candidate, but price-wise this would be over the top; in the end it's just a hobby.

Thanks a lot and Kind Regards
Mathias
 

Panelhead

Active Member
Forum Donor
Joined
Apr 3, 2018
Messages
279
Likes
101
#7
I can answer a couple.
1. Steep filtwrs
3. I have read the chips in the 2Pre and 4Pre are the same. The upscale Red line uses these same chips.
 

Blumlein 88

Major Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
9,663
Likes
12,925
#9
3) ADC and DAC chips used in Clarett 2 Pre ?
4) ADC and DAC chips used in Calrett 4 Pre ? (I read somewhere that it uses 2x AK5388 ADC and 1x AK4413 DAC -> is this right?)
(I'm fully aware, that this is only part of the whole story, but that would allow me to look up the digital filter frequency responses, the digital High-Pass- filter and if these can be changed via a dedicated pin)

Thanks a lot and Kind Regards
Mathias
From what I've read you have the correct chips which also are supposedly used by the Babyface. Those were the same chips used in the Focusrite Forte. I previously owned a Forte. Nothing wrong with it, I only sold it because I had more interfaces than I had need for. I believe the same chips are used throughout the Clarett line.
 
Joined
Jul 26, 2020
Messages
17
Likes
19
#10
i know this is a very generic question but would this make a good cheap audio analyzer? starting to get into DIY audio stuff and need something to quantify what i am doing.

I am looking at the SOLO aswell but the 2Pre is on sale at my local audio store

My goals are 16 Bit 48Khz.
 

Blumlein 88

Major Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
9,663
Likes
12,925
#11
i know this is a very generic question but would this make a good cheap audio analyzer? starting to get into DIY audio stuff and need something to quantify what i am doing.

I am looking at the SOLO aswell but the 2Pre is on sale at my local audio store

My goals are 16 Bit 48Khz.
Yes. Should serve you well.
 

Panelhead

Active Member
Forum Donor
Joined
Apr 3, 2018
Messages
279
Likes
101
#12
Both should be fine. The Focusrite equipment usually exceeds the published specs. The sound card in the computer may be the limiting factor.
The 2Pre has lower noise and better resolution.
 

nanook

New Member
Joined
Mar 31, 2019
Messages
2
Likes
0
Location
Berlin, Germany
#13
Thanks a lot for your feedback.
I just came across a russian website, that clarifies a lot of the hardware side:
https://prosound.ixbt.com/interfaces/focusrite-clarett.shtml
The interface that I'm using currently (codec CS42426) has such a hefty noise shaping (the DAC even more than the ADC), that I would love to see a loopback spectrum at e.g. 1kHz, -60dBFS with 192kS/s. (The distortion of the CS42426 b.t.w. is excellent, as long as you stay several dBs below FS)
Also a loopback frequency response would be great. If the ADC uses the steep digital-filter, thats good news, but they could use an early roll-off in the analog DAC filter.
 
Joined
Sep 14, 2020
Messages
15
Likes
5
Location
Turin (Italy)
#14
I have a question for all "measurers" here.

Focusrite Clarett have a "AIR" function on their front pre. It should mimic the sound of prestigious Focusrite pre mic...

What does it do to the sound actually? Can you come out to it thought your measurements?

Thank you in advance, wether too can fulfill our not my curiosity
 

Blumlein 88

Major Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
9,663
Likes
12,925
#15
I have a question for all "measurers" here.

Focusrite Clarett have a "AIR" function on their front pre. It should mimic the sound of prestigious Focusrite pre mic...

What does it do to the sound actually? Can you come out to it thought your measurements?

Thank you in advance, wether too can fulfill our not my curiosity
I seem to recall that is just a frequency boost at and above 8 khz. If so yes, it could be measured.
 

Blumlein 88

Major Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
9,663
Likes
12,925
#17

Blumlein 88

Major Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
9,663
Likes
12,925
#19
I thought of something more complex, kind a secret sauce, than a sheer hi eq boost....:confused::confused::confused:
That is what they want you to think. Same ideas are pushed about microphones that cut thru the mix. Most have an uptilted response in the upper mids so voice is heard more clearly in the midst of music. And that is about all the secrete sauce there is in them if otherwise decently made.
 
Joined
Sep 14, 2020
Messages
15
Likes
5
Location
Turin (Italy)
#20
Ear is damn stupid and bias-affected... :p:p

Indeed, I am very happy with my Focusrite Clarett 4 pre, despite this buzz feature.

(Coming from a Scarlett 1st gen is like ascending from hell to highest heaven :D)
 

Similar threads

Top Bottom