Focusrite Clarett 2Pre USB - quick measurements

[paulraphael]

Come on, almost 5 V RMS @ 600 Ohms is quite a decent value for an audio interface headamp. Motu M4 has a 7 times lower value. After all, these are studio interface, so headamp should be used, mostly, with monitor cans, not with high impedance or planars.Clarett 2Pre - 4.8V RMS driving 600 Ohms resistive load, so about 38 38 mW

Yes, I agree. I don't know why I was running low on juice. I don't listen at very high volumes.

 

[paulraphael]

I was surprised to see Julian's results, considering that I've been getting great sound with no audible distortion from my 32 ohm Philips cans. I suspect it's because they're very efficient (published spec is 100db at 1mw), and I don't listen anywhere near that loud. I stuck my phone mic into the ear cup with a db meter running (no idea how inaccurate this is, but hoping it's ballpark). At my loudest listening levels it measured around 85 db.

If I'm doing the math right, this would be somewhere in the neighborhood of 100µΩ, which according to Julian puts the THD+N below -70db:

Still not super impressive, but I know from listening tests that I probably can't hear it.

I compared to the headphone output on my Mac Pro tower. At higher levels the computer output showed less distortion than than the Clarett, but it sounded significantly worse. Thinner and brighter. So for whatever reason the computer's output seems to be eq'ing the sound in an unfavorable way. This was not a blind test nor was it objectively level-matched. But I was getting the same strong impression from low to high output. And what I was listening for was distortion (which I couldn't hear from either source).

 

[Prismaspect]

Come on, almost 5 V RMS @ 600 Ohms is quite a decent value for an audio interface headamp. Motu M4 has a 7 times lower value. After all, these are studio interface, so headamp should be used, mostly, with monitor cans, not with high impedance or planars.

So what does this mean? If I get Beyer high impedance headphones with this interface it’s not a good idea? Because I was thinking of doing exactly that. I thought those Beyer cans ARE monitor cans? (Pretty sure “1170” was supposed to be “1770” the closed back model...I’m looking at the 1990’s) There’s an Amazon answer about Beyer high impedance headphones and this interface and they said it should be fine to drive them. Sorry if I’m not understanding all this I’m trying to keep up

 

[paulraphael]

So what does this mean? If I get Beyer high impedance headphones with this interface it’s not a good idea? Because I was thinking of doing exactly that. I thought those Beyer cans ARE monitor cans? (Pretty sure “1170” was supposed to be “1770” the closed back model...I’m looking at the 1990’s) There’s an Amazon answer about Beyer high impedance headphones and this interface and they said it should be fine to drive them. Sorry if I’m not understanding all this I’m trying to keep up

I wouldn't base your decision on my very brief experience. But maybe ask around and see if others can report on the combination. TRL is right, the specs say they should work fine. And Focusrite says the Clarett's good up to 600 ohms. I didn't find the Beyers to be incompatible; I was just surprised to find I had barely enough power for quieter tracks, especially if I was inserting effects that required the levels to be turned down 4 or 5 db.

 

[trl]

So what does this mean? If I get Beyer high impedance headphones with this interface it’s not a good idea? Because I was thinking of doing exactly that. I thought those Beyer cans ARE monitor cans? (Pretty sure “1170” was supposed to be “1770” the closed back model...I’m looking at the 1990’s) There’s an Amazon answer about Beyer high impedance headphones and this interface and they said it should be fine to drive them. Sorry if I’m not understanding all this I’m trying to keep up

RME ADI interfaces are having a beefier headphones amplifier, if want to try out some German sound. However, Clarett is having a beefier headamp than Focusrite Solo Gen3 and Motu M4, so it's up to you with what headphones you would like to pair it, but in my objective and subjective tests I had no troubles with Clarett 2Pre USB listening to AKG K701 or Hifiman HE-560 to satisfacatory levels.

 

[trl]

Yes, I agree. I don't know why I was running low on juice. I don't listen at very high volumes.

It's internal headamp is definitelly not built for high end audio and high volume listening levels, but I think it should do the job well. Just take care on the settings done on Focusrite Control panel, as headphones audio can be set from both potentiometers: the big master one and the small one for the headphones!

Meanwhile, you can check some other measurements I've done here: https://www.audiosciencereview.com/...analysis-and-few-in-house-measurements.18605/.

 

[trl]

I was surprised to see Julian's results, considering that I've been getting great sound with no audible distortion from my 32 ohm Philips cans. I suspect it's because they're very efficient (published spec is 100db at 1mw), and I don't listen anywhere near that loud. I stuck my phone mic into the ear cup with a db meter running (no idea how inaccurate this is, but hoping it's ballpark). At my loudest listening levels it measured around 85 db.

If I'm doing the math right, this would be somewhere in the neighborhood of 100µΩ, which according to Julian puts the THD+N below -70db:

View attachment 100531

Still not super impressive, but I know from listening tests that I probably can't hear it.

I compared to the headphone output on my Mac Pro tower. At higher levels the computer output showed less distortion than than the Clarett, but it sounded significantly worse. Thinner and brighter. So for whatever reason the computer's output seems to be eq'ing the sound in an unfavorable way. This was not a blind test nor was it objectively level-matched. But I was getting the same strong impression from low to high output. And what I was listening for was distortion (which I couldn't hear from either source).

I don't understand how and why Julien gets these poor measurements while mine are way better. I will try to investigate this further and get back here.

P.S.: I guess you were referring to 100 µW instead of 100µΩ.

 

[trl]

Just to be 100% sure, I measured again with Motu M4 the headphones output of Clarett 2Pre USB at 2 V RMS @ 300 Ohms (13.33 mW) and I got a THD+N of 0.0011% that corresponds to a THD+N of -99.2 dB instead of -70 dB that appears in Julien's graph (the blue line). Headphones potentiometer was pretty close to max. position, same for the main digital one. I used a 300 Ohms pure resistive load, not real headphones.

Clarett 2Pre headphones output at 2 V RMS @ 300 Ohms (13.33 mW), THD+N: -99.2 dB

​

Another measurement below, but now for a 0.25 V RMS (0.2 mW @ 300 Ohms) output level; both pots at the same level as the first measurement, including the Motu M4 Line2 gain level:

Clarett 2Pre headphones output at 0.25 V RMS @ 300 Ohms (0.2 mW), THD+N: -88.4 dB

​

Same as above, but the Motu M4 Line2 gain level was increased to near 0 dB of the scale:

Clarett 2Pre headphones output at 0.25 V RMS @ 300 Ohms (0.2 mW), THD+N: -90 dB
​

You can notice the increase in noise due to Motu M4 gain/sensitivity increase. With an Audio Precision the noise difference would be lower and THD+N better.

However, my figures for 0.2 mW (200 µW) are very close with the approx. -90 dB found in the clip published above, but I wonder who and what did wrong for the > 10 mW THD+N from the blue line? Worth mentioning that with Hifiman HE-560 connected (about 40 Ohms) I got a THD+N of -84 dB @ 100 mW for 1 kHz sinewave, like 30 dB better than the figure published in the above movie from post #42 for the same 100 mW power/32 Ohms (the red line).

 

[paulraphael]

I don't understand how and why Julien gets these poor measurements while mine are way better. I will try to investigate this further and get back here.

P.S.: I guess you were referring to 100 µW instead of 100µΩ.

Yes, I meant 100 µW. Very curious to see what you discover, especially with lower impedance cans.

 

[Interference]

I finally found a Clarett 2Pre USB at a good price and decided to give it a try.

Usually I always test my devices with minimum input gain and maximum output level, if sufficient headroom is allowed. I am also testing the device with RCA-to-TRS adapters to have an idea of what I can expect from testing a 2 Vrms single-ended source.

First, I was surprised by the fairly high amount of distortion in loopback when using the input in "line" mode. Not sure if this is due to "low" impedance charging too much the output stage (but I do not remember this being so bad on my E-MU 0404).

In "instrument" mode I get -100 dB distortion and N+D roughly at -94/95 dB. This is not much better than the good old E-MU 0404.

I will try to reproduce @coastpunk results, where distortion is 10 dB lower, although even on their result I see N+D at -100 dBFS(A) with a signal at -6 dbFS, meaning again N+D at -94 dB.

 

[trl]

Use the input pot to around 08:30-09:00 o'clock for best results for 4V RMS input. It's all about finding the sweet spot.

 

[Interference]

Use the input pot to around 08:30-09:00 o'clock for best results for 4V RMS input. It's all about finding the sweet spot.

Thanks, it is a bit counter-intuitive that a raise in the gain would improve overall noise and distortion, isn't it?

(I think I should also exclude the output pot from the loop.)

 

[trl]

Not really, ot all depends on how the designer was using the inside chained gain stages from line inputs. At least in my case, with 4V RMS input signal I get the best results (around -115dB THD @1KHz) at around -2.1dB, which corresponds to near 9 o'clock input pot.

 

[Interference]

OK, so I figured two things:
1. the "monitor" potentiometer is not really a pot after all, it is rather an encoder. It controls the output level alternatively to the software (depending on the setup of Focusrite Control) by driving some kind of hardware volume control; maybe this is documented somewhere;
2. a few dB attenuation through the hardware volume control (either done via knob or software) seems to yield better results compared to applying the same attenuation to the digital signal.

All in all I get around 97 dB(A) of SINAD in loopback (unbalanced connection).

 

[trl]

the "monitor" potentiometer is not really a pot after all, it is rather an encoder

Not sure I get it. It is real potentiometer, but in the backend the input gain is still digitally controlled. Only the headphones output volume is controlled in the analogue domain.

Check the links from below and see if those articles are helpful or not:
https://www.audiosciencereview.com/...analysis-and-few-in-house-measurements.18605/
https://prosound.ixbt.com/interfaces/focusrite-clarett.shtml (Chrome + Right Click and choose Translate to English).

 

[AnalogSteph]

It is real potentiometer, but in the backend the input gain is still digitally controlled.

The old ADC trick. Little 8-bit low-speed ADCs are super common in micros and the like, and 255 levels should allow for more than enough volume steps.

 

[Interference]

Thanks @trl , and I apologise for not checking the available information in advance

I played a bit more with the interface and I was able to reproduce the figures of the first post, corresponding to a ~0.77 Vrms signal. Things look slightly worse when running at 2.1 Vrms, not sure if due to the DAC or the ADC. There is some benefit in reducing a bit the output level with the Clarett volume control rather than doing it in the digital domain, as I mentioned before.

Some of my other findings:

- the interface works OK under Linux, but "naturally" Focusrite Control is provided only for Windows and Apple platforms; in case anyone is wondering: my attempts to run it in a Win10 virtual machine with VirtualBox were unsuccessful as the interface is recognised but not detected by the software;

- input in line mode is OK when tested in loopback with a balanced cable, but distortion rises dramatically with an unbalanced connection (using TRS-RCA adapters); the instrument input is fine across the two modes; I will check if this is due to the DAC or the ADC with an external source;

- output 1 => input 1 gives the best loopback performance, the others have 2-3 dB inferior SINAD.

Not sure I am gonna keep it, in spite of having relatively lower distortion its usability as a poor man's analyser is not that much superior with respect to my old E-MU. Would still love to see it tested by Amir, though.

 

[paulraphael]

Not sure I am gonna keep it, in spite of having relatively lower distortion its usability as a poor man's analyser is not that much superior with respect to my old E-MU. Would still love to see it tested by Amir, though.

He'll test it if you send it to him!

 

[Interference]

He'll test it if you send it to him!

I am not in the US, otherwise I would seriously consider doing it

 

[trl]

I played a bit more with the interface and I was able to reproduce the figures of the first post, corresponding to a ~0.77 Vrms signal. Things look slightly worse when running at 2.1 Vrms, not sure if due to the DAC or the ADC. There is some benefit in reducing a bit the output level with the Clarett volume control rather than doing it in the digital domain, as I mentioned before.

Try feeding it with an external source, something you can rely on (a DAC with a good SINAD, at least 110 dB).

- input in line mode is OK when tested in loopback with a balanced cable, but distortion rises dramatically with an unbalanced connection (using TRS-RCA adapters); the instrument input is fine across the two modes; I will check if this is due to the DAC or the ADC with an external source;

Clarett 2Pre has no RCA sockets, right? If you're trying to get output signal from Clarett 2Pre via the TRS outputs to an external ADC that has RCA inputs, then you are probably short-circuiting V- from each output by doing this, hence the higher distortions (unless I'm misunderstanding your setup).

- output 1 => input 1 gives the best loopback performance, the others have 2-3 dB inferior SINAD.

Feel free to post here some of these measurements. Also, use a scope or a TrueRMS multimeter to measure the output voltage for each channel tested.

P.S.: Note sure what exactly you want to do with the Clarett 2Pre, but for 4 to 6 V RMS measurements, Motu M4 might be a better match, especially for measuring 5 W @ 8 Ohms outputs coming from amplifiers. If you need best measurements for 0.77 V RMS inputs, then let me know, maybe I'll have time to do again couple of measurements for both Motu M4 and Clarett 2Pre, while being fed from Topping D90 (120 dB SINAD). This will show the truth about their ADC inputs only, just in case their DAC-outputs might have SINAD troubles for lower volumes (like 0.77 V RMS).