Motu M2 Review (Audio Interface)

[Robbo99999]

This looks like a really good balanced DAC & headphone amp for a good price as long as you use it within a few of it's usage limits that have been exposed in this review, I'm really quite impressed, but you gotta research those limits exposed in this review to see if they'd affect you.

 

[Pearljam5000]

https://www.ifixit.com/Teardown/Solid+State+Logic+SSL+2+Teardown/135354

AKM AK4621EF

https://e-muzyk.eu/testy/1244/solid...s-audio-usb-24-bit-192-khz-do-domowego-studia

Thanks I remember from my days on Head - fi that Ess chips are more detailed but brighter, and AKM are more "musical" and organic and warm
Is this still the case today?

 

[AnalogSteph]

Thanks I remember from my days on Head - fi that Ess chips are more detailed but brighter, and AKM are more "musical" and organic and warm
Is this still the case today?

If this was ever a real thing at all (of which there can be some doubt), the only possible reason I can think of would be that the integrated ASRC in the ESS chips may not be fond of >0 dBFS levels (intersample-overs). (AKMs typically do not include one, which is why DACs with them tend to fare measurably worse in terms of jitter over SPDIF or HDMI.) There may be some indication of this in this thread, the M500 is an ES9038PRO job. I also remember reading that a CS4398 (on Xonar D1) would handle intersample-overs to some degree while a PCM1796 (on Xonar D2) clips at 0 dBFS flat. This potential problem area is easily reined in on a computer with a pinch of digital attenuation if need be, potentially already provided by level normalization like ReplayGain.

Aside from that, any two DACs with ruler-flat frequency response in the audio band, typical inaudible levels of distortion and noise, decent enough jitter, decently low periodic passband ripple relating to FIR reconstruction filters (pre-ringing/pre-echo) and passable out-of-band rejection should sound nothing short of identical when levels are matched. They literally are that good. Someone here has ABX'd an RME DAC and a USB dongle (about an order of magnitude of price difference, possible more) and utterly failed to tell them apart.

The AK4621 is a decent midrange chip, even though the DAC reconstruction filter side rather counts as a step back over the preceding AK4620B for playback purposes, as periodic passband ripple spec is no better than +/-0.06 dB even for the filter options of "traditional" latency (the old chip could offer +/-0.005 dB and a group delay of 28/fs which the newer ones reduces to 27/fs, not sure that's a good tradeoff - the added short delay options with 7/fs look more sensible). I don't really like to see more than +/-0.02 dB, and preferably +/-0.005 dB - some CD player filters were one or two orders of magnitude better than that 30 years ago already.
Now the spec may be bogus and missing a zero for the "traditional" delay options (a slight deterioration would seem way more plausible, the A/D side still manages essentially the same performance as well), but I suspect an audio interface would be using the short delay option anyway, given how big of a topic latency is in recording circles. 20/fs is about 0.4 ms at 48 kHz, by the time you're adding another 25/fs on the A/D side that's almost a full millisecond you can save on roundtrip latency. I guess there are cases where you can't or won't use a higher sample rate even though that would generally be the best solution.

 

[Pearljam5000]

If this was ever a real thing at all (of which there can be some doubt), the only possible reason I can think of would be that the integrated ASRC in the ESS chips may not be fond of >0 dBFS levels (intersample-overs). (AKMs typically do not include one, which is why DACs with them tend to fare measurably worse in terms of jitter over SPDIF or HDMI.) There may be some indication of this in this thread, the M500 is an ES9038PRO job. I also remember reading that a CS4398 (on Xonar D1) would handle intersample-overs to some degree while a PCM1796 (on Xonar D2) clips at 0 dBFS flat. This potential problem area is easily reined in on a computer with a pinch of digital attenuation if need be, potentially already provided by level normalization like ReplayGain.

Aside from that, any two DACs with ruler-flat frequency response in the audio band, typical inaudible levels of distortion and noise, decent enough jitter, decently low periodic passband ripple relating to FIR reconstruction filters (pre-ringing/pre-echo) and passable out-of-band rejection should sound nothing short of identical when levels are matched. They literally are that good. Someone here has ABX'd an RME DAC and a USB dongle (about an order of magnitude of price difference, possible more) and utterly failed to tell them apart.

The AK4621 is a decent midrange chip, even though the DAC reconstruction filter side rather counts as a step back over the preceding AK4620B for playback purposes, as periodic passband ripple spec is no better than +/-0.06 dB even for the filter options of "traditional" latency (the old chip could offer +/-0.005 dB and a group delay of 28/fs which the newer ones reduces to 27/fs, not sure that's a good tradeoff - the added short delay options with 7/fs look more sensible). I don't really like to see more than +/-0.02 dB, and preferably +/-0.005 dB - some CD player filters were one or two orders of magnitude better than that 30 years ago already.
Now the spec may be bogus and missing a zero for the "traditional" delay options (a slight deterioration would seem way more plausible, the A/D side still manages essentially the same performance as well), but I suspect an audio interface would be using the short delay option anyway, given how big of a topic latency is in recording circles. 20/fs is about 0.4 ms at 48 kHz, by the time you're adding another 25/fs on the A/D side that's almost a full millisecond you can save on roundtrip latency. I guess there are cases where you can't or won't use a higher sample rate even though that would generally be the best solution.

Thanks for the comment.
"any two DACs with ruler-flat frequency response in the audio band, typical inaudible levels of distortion and noise, decent enough jitter, decently low periodic passband ripple relating to FIR reconstruction filters (pre-ringing/pre-echo) and passable out-of-band rejection should sound nothing short of identical"
While it does make sense, how do you explain DAC's that have more than one chip , like the Benchmark DAC 3 that has 4 of them , in theory it should have no benefits at all.

 

[AnalogSteph]

While it does make sense, how do you explain DAC's that have more than one chip , like the Benchmark DAC 3 that has 4 of them , in theory it should have no benefits at all.

Oh, it absolutely does - measurable ones anyway. The dynamic range of modern DACs and ADCs tends to be limited by their analog (uncorrelated) noise floor, so for each doubling of converters you can wring out up to 3 dB less noise. Which is why you will find up to 8-channel DACs used for stereo and the like. Common practice e.g. with ESS DACs or AKM ADCs (e.g. AK5578 with an 8-to-1 mode, expectedly boosting its 121 dB DR to 130 dB). Otherwise IC converters would not have reached the 130 dB vicinity. 20 years ago, this was the domain of hybrids in converter boxes with 5-digit price tags and analog levels of up to +28 dBu balanced.

Now no audio playback application requires an instantaneous dynamic range of more than about 110 dB(A) at any given time, so you could be happily amplifying and attenuating this as needed in the analog domain to accommodate different playback setups. That's why studio-level converters tend to have analog level settings, allowing them to cover a huge total dynamic range if need be (e.g. RME ADI-2 Pro FS: 123 dB(A) + 20 dB = 143 dB(A)). This is, however, costly and/or inconvenient, given that at the kind of performance levels required to not degrade converter performance you can't just use a PGA and old-fashioned switches / relays / jumpers and potentially trimpots tend to be used instead.
Up to a certain point, it tends to be more economical to just increase converter dynamic range and use a fixed-level analog stage, accommodating any analog level mismatch via digital attenuation instead. (The Asus Xonar soundcards were among the first to really push this concept.) And I mean, something like the Okto DAC8 stereo (another DAC channel combiner) covers an instantaneous dynamic range of 131 dB(A) already, that's more than a great many analog preamps. It works better for measurement purposes as well, which is really where a lot of the appeal lies these days. You can just "brute-force" your way more easily.

Now tradeoffs are not the same everywhere. High dynamic range DACs typically go along with substantial power consumption, so in a portable player it tends to be more advantageous to go with a less fancy DAC and a PGA for analog gain control still, even if modern designs like CS43130/131 with Class H analog stages have made considerable headway in the economy department.

 

[Jimster480]

Really great performance considering the price. It blows my UMC202HD out of the water and honestly doesn't cost THAT much more.

 

[AnalogSteph]

Really great performance considering the price. It blows my UMC202HD out of the water and honestly doesn't cost THAT much more.

Mind you, Behringer is rather expensive in the US these days (blame tariffs on Chinese goods, I guess). Here in Germany where it tends to be the exact opposite, the M2 (when available) typically cost 190€ (with dips to 179€) vs. 64...69€ for the UMC202HD, which obviously is not even close.

BTW, M2 pricing is now up to 198€, and if you want one that's actually in stock right now it's more like 219. My guess is that MOTU may have stockpiled units (or at least finished PCBs) in advance for the holiday season, and now supply is slowly drying up.
I wonder what their strategy is - wait until production at AKM is back up? Respin the board to take a different type of ADC? The number of suitable parts to replace an AK5552 in both performance and price is quite limited, PCM4202 might be the closest ($4.80 and change for a reel of 1000 rather than $2 and change, but I don't think it has any low-latency filter options and 28SSOP is a wee bit larger than 48QFN as well, not to mention that power consumption @ 48 kHz is 308 mW instead of 83 mW).

 

[txbdan]

Wish someone would create a model with a stronger headphone amp, which is the only thing holding these devices back. Even some devices that require a wall plug only seem to achieve 20 mW.

Is it possible to control the RCA and TRS outputs separately in software, or are both always controlled by the Monitor knob? Guessing you would need something with dedicated line outputs like the M4 to bus a separate full-output send to a headphone amp.


What kind of computer do you have? Some models have known DPC latency issues with USB audio interfaces, like Dell XPS and T2-based Macs.

DPC Latency Rankings (anecdotal, but a good starting point):
https://www.notebookcheck.net/DPC-L...ablets-offer-the-lowest-latency.504376.0.html

List of T2-based Macs:
https://support.apple.com/en-us/HT208862

My understanding is that the T2 USB was solved a year or so ago. I had a T2 2018 Mac mini and it had the USB issues. I ended up going with a thunderbolt interface (MOTU 624 coincidentally), but later found that the issue was fixed and could have gone with a much cheaper interface.

 

[infinitesymphony]

My understanding is that the T2 USB was solved a year or so ago. I had a T2 2018 Mac mini and it had the USB issues. I ended up going with a thunderbolt interface (MOTU 624 coincidentally), but later found that the issue was fixed and could have gone with a much cheaper interface.

That would be great. Do you have more info about the patch? Just want to make sure I'm up to date.

 

[txbdan]

That would be great. Do you have more info about the patch? Just want to make sure I'm up to date.

I haven’t kept up with the issue, but gearslutz has the scoop I’m sure.

 

[Damian]

This and the very similar Scarlett 2i2 are both good. How stable are the drivers?

Coming from 2i2 3rd Gen to Motu M4, I would say Scarlett has less problems... in my Motu I had issues with Zoom and what not, I have to use specific bitrates and sometimes there is no audio output after the PC coming from Sleep or what not and I need to turn off and on, then is ok. Also the scarlet had more volume output somehow to power my monitors.

 

[daftcombo]

Coming from 2i2 3rd Gen to Motu M4, I would say Scarlett has less problems... in my Motu I had issues with Zoom and what not, I have to use specific bitrates and sometimes there is no audio output after the PC coming from Sleep or what not and I need to turn off and on, then is ok. Also the scarlet had more volume output somehow to power my monitors.

I too have to change the sample rate manually to 96.000 Hz in the Focusrite panel to use Webex with the Scarlett 2i4 2nd Gen. Zoom works fine though (I set 44.000 Hz by default in Windows and the Focusrite panel).

Some time ago I also had to unplug / plug again the soundcard when the PC went out of sleep (I don't use sleeping mode now).

 

[AnalogSteph]

Also the scarlet had more volume output somehow to power my monitors.

That's odd, on paper they should be near identical (+15.5 dBu vs. +16 dBu max). Some sort of settings (mixer) issue?

 

[Damian]

That's odd, on paper they should be near identical (+15.5 dBu vs. +16 dBu max). Some sort of settings (mixer) issue?

God knows, now I'm passing it through EQ APO and I have plenty of volume again. I take windows could be the culprit in all problems if I take it as a comparison with my macbook. I'm thinking to switching to a macbook mini with M1.. will see how it goes.

 

[GGekko]

This is a review and detailed measurements of the Motu M2 Audio Interface (DAC, ADC and headphone amplifier). It is on kind loan from a local member and costs US $179. I previously reviewed Motu M4 which is its larger brother.

The M2 has the look of other Motu products:

View attachment 109298

I like the color LED bar graphs as if you don't get sound, you know whether it is or is not getting to the product. Front panel sockets are for input sans the 1/4 headphone jack. The back has the only set of outputs (both balanced TRS and RCA):

View attachment 109299

The overall feel of the product was good. Compatibility was great without me having to use their control panel or driver except to set the ADC sample rate as needed.

M2 is bus powered using that USB-C connector which was nice.

Motu M2 Measurements: DAC
Measurements of the M2 started on a sour note using the unbalanced output with one channel being 8 dB worse than the other due to noise. Strangely, as it sat there for a few minutes, the noise subsided fair bit and improved that channel to 104 dB SINAD:
View attachment 109300

Fortunately there was no such issue when using the TRS balanced output:

View attachment 109301

This is quite a competent performance for such a budge multifunction device:
View attachment 109302

Dynamic range was good:

View attachment 109303

As was jitter:

View attachment 109304

Unfortunately even though Chinese manufacturers have figured out how to get rid of the increase in intermodulation distortion when using ESS DAC chips, the news has not gotten to M2 yet:

View attachment 109305

Unlike pro interfaces with their own power supplies, there is not much more output to be had than the 4 volts I test:
View attachment 109306

I was pleasantly surprised to see a proper reconstruction filter that chops everything off at 22 kHz rather than the common 24 kHz:
View attachment 109307

Linearity was almost perfect:
View attachment 109308

Distortion rises with frequency but not enough to be a concern:
View attachment 109309

View attachment 109310

Motu M2 Measurements: ADC
For this testing, I used TRS input in the front:

View attachment 109311

Specification is confusing but seems to imply if you use TRS, you have lower gain and reference would be 16 dBu. That did not happen with my testing. Clipping occurred at a low 2.5 volts as opposed to minimum of 4 volts. You can see that in this sweep as well:

View attachment 109313

Notice also how distortion rises well before clipping.

Overall ranking is average and below Motu M4:
View attachment 109316


We also see rising distortion at both ends of frequency spectrum:
View attachment 109314

Fortunately our hearing thresholds are higher in those extremes so this is less of an audible problem than pure engineering.

Frequency response was wide enough at 192 kHz sampling:

View attachment 109315

Motu M2 Measurements: Headphone Amplifier
Power is usually an issue with these types of headphone outputs so let's jump right into that:
View attachment 109317


View attachment 109318

Not much power but what is there is good prior to clipping. So if you use a sensitive headphone, this may be a workable solution.

Conclusions
The DAC implementation in Motu M2 is quite good especially given the price. Clearly attention was put in to produce a good design. On ADC side, looks like they wanted to tier this down below the M4 and performance there is just adequate. Headphone amplifier is barely adequate from power point of view but otherwise, not bad.

Still, the package is well worth considering over a DAC-only solution. You get nice VU meters, mic and analog in ADC and passable headphone output for an extra $80. Good trade in my book.

Overall I am happy to recommend the Motu M2. If you can though, spend the other $60 and get the M4 as that has better ADC and two more channels.

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

Appreciate any donations using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

I'm surprised the M2's ADC SINAD is this low. According to Motu's website, the M2 and M4 should perform the same is this regard:

https://motu.com/en-us/products/m-series/m2/specs/
has THD+N (unweighted) at -129dBu EIN
https://motu.com/en-us/products/m-series/m4/specs/
also has THD+N (unweighted) also at -129dBu EIN and other specs are the same

I also don't understand why the M2's ADC performance seems to be outmatched by the Scarlett 2i2 3rd gen, when it's official specs (https://focusrite.com/en/usb-audio-interface/scarlett/scarlett-2i2 showing Noise EIN-128dBu (A-weighted)) indicate the contrary. Also, other measurements show the contrary, where the M2 seems to beat the 2i2 in that regard, for instance here:
https://musicalvs.com/2020/06/11/motu-m2-vs-focusrite/
and here :

I'm by no means an expert when it comes to this so I'd be glad if someone could enlighten me about these apparent discrepancies

 

[Tks]

I'm surprised the M2's ADC SINAD is this low. According to Motu's website, the M2 and M4 should perform the same is this regard:


I'm by no means an expert when it comes to this so I'd be glad if someone could enlighten me about these apparent discrepancies

Yeah kinda weird. Aside from SINAD, even the other plots show better results for the M2 (none of that weird rise in both low and high level output). And even the filter is different (oddly enough the M2 has the best filter I've seen in almost any all-in-one device).

Who knows what's going on here..

 

[Blumlein 88]

I'm surprised the M2's ADC SINAD is this low. According to Motu's website, the M2 and M4 should perform the same is this regard:

https://motu.com/en-us/products/m-series/m2/specs/
has THD+N (unweighted) at -129dBu EIN
https://motu.com/en-us/products/m-series/m4/specs/
also has THD+N (unweighted) also at -129dBu EIN and other specs are the same

I also don't understand why the M2's ADC performance seems to be outmatched by the Scarlett 2i2 3rd gen, when it's official specs (https://focusrite.com/en/usb-audio-interface/scarlett/scarlett-2i2 showing Noise EIN-128dBu (A-weighted)) indicate the contrary. Also, other measurements show the contrary, where the M2 seems to beat the 2i2 in that regard, for instance here:
https://musicalvs.com/2020/06/11/motu-m2-vs-focusrite/
and here :

I'm by no means an expert when it comes to this so I'd be glad if someone could enlighten me about these apparent discrepancies

SINAD as Amir measures it will be via the line level inputs. EIN specs refer to going thru the microphone preamp section. One doesn't tell what to expect with the other.

Also you need to keep in mind SINAD with the ADC is different than with the DAC.

Maybe back up and tell me which you are comparing to which here.

Also EIN (equivalent input noise) doesn't tell you what many people think it tells you. It isn't saying the ADC/mic pre has 129 db of dynamic range.
What it is telling you is with 60 db of gain the noise level with 150 ohm output impedance mic feeding it is equivalent to -129 db noise level with no gain. Or that you have a noise level of -69 dbu with 60 db of gain. There is a limit of -131 dbu that is possible as the 150 ohm impedance creates noise at this level itself.

 

[GGekko]

SINAD as Amir measures it will be via the line level inputs. EIN specs refer to going thru the microphone preamp section. One doesn't tell what to expect with the other.

Also you need to keep in mind SINAD with the ADC is different than with the DAC.

Maybe back up and tell me which you are comparing to which here.

Also EIN (equivalent input noise) doesn't tell you what many people think it tells you. It isn't saying the ADC/mic pre has 129 db of dynamic range.
What it is telling you is with 60 db of gain the noise level with 150 ohm output impedance mic feeding it is equivalent to -129 db noise level with no gain. Or that you have a noise level of -69 dbu with 60 db of gain. There is a limit of -131 dbu that is possible as the 150 ohm impedance creates noise at this level itself.

Thank you for your reply. I'm not interested in the DAC function (I have a better DAC already) I'm just interested in the mic in/preamp performance of the M2 vs the M4 and the 2i2 3rd gen.

 

[Blumlein 88]

Thank you for your reply. I'm not interested in the DAC function (I have a better DAC already) I'm just interested in the mic in/preamp performance of the M2 vs the M4 and the 2i2 3rd gen.

Well Amir mentioned on the M2 he had SINAD of 94 db in the mic inputs. It was 104 db over line. He didn't mention mic inputs on the M4, but did obtain a slightly lower SINAD of 94 for line inputs.

So you really don't have the information you need on this. I would think the M2 and M4 should be the same within whatever unit to unit variation they have. The claimed specs are very slightly better than claimed specs for the 2i2, but the difference is quite small. EIN of 1 db less and DR of 4 db less on the 2i2. 2i2 claimed SINAD would be 98 db. I don't think any of those are earth shatteringly different in use.

What meaningfully differs among various mic pres is the EIN over the 20-40 db gain range, and I don't know of anyone that tests that regularly. The better units won't have EIN degrade until you get gain below 20-25 db. I have a 1st gen 18i20 Scarlett, and if I'm not remembering wrong it is good to right around 30 db gain before EIN drops. An Antelope Audio unit I have is good to about 25 db before EIN degrades. Neither will matter except with the very quietest microphones you can get as the microphone self noise will be an issue by then. And on top of that most venues you will record in if they aren't a quiet treated studio are louder than the microphone self noise.

These units in this price range are all pretty good values. Other factors probably would be more important in owning one than the slight test results as none of the three are horrible and of course none are SOTA.

 

[GGekko]

Well Amir mentioned on the M2 he had SINAD of 94 db in the mic inputs. It was 104 db over line. He didn't mention mic inputs on the M4, but did obtain a slightly lower SINAD of 94 for line inputs.

So you really don't have the information you need on this. I would think the M2 and M4 should be the same within whatever unit to unit variation they have. The claimed specs are very slightly better than claimed specs for the 2i2, but the difference is quite small. EIN of 1 db less and DR of 4 db less on the 2i2. 2i2 claimed SINAD would be 98 db. I don't think any of those are earth shatteringly different in use.

What meaningfully differs among various mic pres is the EIN over the 20-40 db gain range, and I don't know of anyone that tests that regularly. The better units won't have EIN degrade until you get gain below 20-25 db. I have a 1st gen 18i20 Scarlett, and if I'm not remembering wrong it is good to right around 30 db gain before EIN drops. An Antelope Audio unit I have is good to about 25 db before EIN degrades. Neither will matter except with the very quietest microphones you can get as the microphone self noise will be an issue by then. And on top of that most venues you will record in if they aren't a quiet treated studio are louder than the microphone self noise.

These units in this price range are all pretty good values. Other factors probably would be more important in owning one than the slight test results as none of the three are horrible and of course none are SOTA.

ok thank you for this clarification