• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Motu M4 Audio Interface Review

OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,672
Likes
241,061
Location
Seattle Area
@amirm thank you for your review of the motu m4! But it seemed strange to me, why you are testing ADC Motu (m4, 624) in unbalanced mode?
I ran out of TRS to XLR cables with the right configuration! I am not into music production so have limited set of TRS cables. Post the review I purchased a few more so I am set from here on.

Note that the "unbalanced" output from Audio Precision is still floating (not ground referenced unless I force it) and is capable of very high output levels. So is a halfway solution toward fully balanced.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,672
Likes
241,061
Location
Seattle Area
@amirm Nice review, thanks again for all you do.

Any chance you could show SE performance on the DAC portion of this device? Interested to see if it differs at all from balanced...
Well, if you can come over and can 20 pounds of peppers like I did most of today, sure, I will test it that way. :)
 

shuri

Member
Joined
Nov 27, 2019
Messages
29
Likes
29
Location
Finland
I was considering this device a while ago, but I think the thing was that the headphone output was summed from all the 4 output channels?
So you couldn't have different EQ's on the headphone out and line out.
 

Atanasi

Addicted to Fun and Learning
Forum Donor
Joined
Jan 8, 2019
Messages
716
Likes
796
I was considering this device a while ago, but I think the thing was that the headphone output was summed from all the 4 output channels?
So you couldn't have different EQ's on the headphone out and line out.
According to the manual, the headphone output is the same signal as the monitor output, but with an independent volume control. So, outputs 3/4 are distinct.
 

thefsb

Addicted to Fun and Learning
Joined
Nov 2, 2019
Messages
796
Likes
657
According to the manual, the headphone output is the same signal as the monitor output, but with an independent volume control. So, outputs 3/4 are distinct.
That fits my experience. I use 4 inputs but only the monitor (i.e. 1/2) outputs and the headphone. So I suppose you could set up 3/4 with a differently equalized version of the same program that's on monitor/headphones.
 

SDX-LV

Active Member
Joined
Jan 11, 2020
Messages
135
Likes
143
Location
Sweden
I am wondering if M4 can output to all 4 outs simultaneously, to do some bass management with it? Can't fing any info if the 4 outs are actually usable or is it just for switching between 2 stereo pairs.
 

Atanasi

Addicted to Fun and Learning
Forum Donor
Joined
Jan 8, 2019
Messages
716
Likes
796
I am wondering if M4 can output to all 4 outs simultaneously, to do some bass management with it? Can't fing any info if the 4 outs are actually usable or is it just for switching between 2 stereo pairs.
It depends on drivers. M4 is a class-compliant four-channels in-and-out device internally. WDM drivers present stereo devices, ASIO or Jack on Linux use four channels in and out.
 
Last edited:

AllanDavidson

Member
Joined
Feb 10, 2020
Messages
90
Likes
34
This is one of the best pieces of hardware to ever come to the segment.

- Mic Preamps (w/Phantom Power)
- Instrument Preamps
- Good enough AKM ADC
- Decent ESS DAC
- Headphone Amp
- 5v Amp output (bus powered)
- LCD metering
- Loopback with Monitoring
- Well built
- USB-C

All this for ~$220. Explains why this unit is out of stock EVERYWHERE for months. I brought it on Guitar Center back in May, still waiting for the backorder.

_
I wish Motu would release a new M6/M8 line with the AKM/ES issues improved, plus raw digital audio (optical/coax) support @ 192/32 with self/auto clocking
 

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,309
Likes
1,475
I was thinking that this might work for a software crossover for a simple 2 + 1 subwoofer setup. I already use brutefir on my LMS server for DRC and a RPi4 running piCorePlayer for the streamer. What's not clear to me is whether LMS can stream the 3 needed channels to the RPi and whether squeezelite can then route the 3 channels to the Motu.
 

VeerK

Active Member
Joined
Jan 29, 2019
Messages
258
Likes
318
Location
NYC
I’m looking at this, the 624, and the Babyface Pro FS for usage with Oktava mics and vinyl digitization. Has anyone had any experience with these and what the 624 and Babyface do better? Seems like other than a few dab SINAD, not much to complain about here, especially given that it’s about $220
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,393
Likes
3,341
Location
.de
If you're looking at ADC THD(f), you can definitely tell the difference - distortion performance degrades at higher frequencies more so than in the 624, and the low end shows an increase in noise floor. You can also spot ADC digital filter periodic ripple on the RMAA FR graph with the naked eye, the AK5554 datasheet suggests about 0.06 dB peak-peak in single and double speed mode. (It reduces to about 0.02 dB peak-peak in quad speed and seems to reduce further at 384 and 768 kHz, but I'd consider those a tad inconvenient. The much older AK5385A has a +/-0.005 dB = 0.01 dB p-p passband ripple spec.) The Babyface with a similar but higher-grade chip (AK5574) does not seem to be doing this, perhaps they are leveraging some DSP power to do either downsampling or FIR filtering?

The M4 is probably about as good as still-inexpensive bus-powered USB audio interfaces get these days. It gets fairly close to some higher-grade options but cannot entirely overcome the limitations of its parts and power budget in the end. There's no fancy DSP or anything either, just an XMOS USB interface. (Higher-end units often sport FPGAs with firmware and drivers developed in-house. Software support for third-party chipsets has often proved dodgy over the years, resulting in either annoying bugs or driver updates ceasing early.)

I suggest you RTFM for all three. This is very much needed for the Babyface anyway, which is a rather powerful beast.

I would assume that it's very much possible to iron out the passband ripple with some (presumably FIR) EQ in software, but this would be quite software-dependent. It may be nice to have this as a selectable driver option, if this is feasible at all.
 
Last edited:

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,693
The Babyface with a similar but higher-grade chip (AK5574) does not seem to be doing this, perhaps they are leveraging some DSP power to do either downsampling or FIR filtering?
These RME interfaces have FPGA anti-ripple filters after Amir complained about the ripple.
https://www.audiosciencereview.com/...0s-adc-comp-to-rme-adi-2-pro.3143/#post-78592
https://forum.rme-audio.de/viewtopic.php?pid=132166#p132166
https://www.audiosciencereview.com/...s-portable-interface-review.12313/post-363016

M4 on the other hand...
https://www.audiosciencereview.com/...u-m4-audio-interface-review.15757/post-503953

Just for fun, here are my Creative soundcard with TI PCM4220 ADC recording my mainboard's ALC892 line out:

44k
fr.png


48k
fr.png


96k
fr.png


The M4 is probably about as good as still-inexpensive bus-powered USB audio interfaces get these days. It gets fairly close to some higher-grade options but cannot entirely overcome the limitations of its parts and power budget in the end. There's no fancy DSP or anything either, just an XMOS USB interface. (Higher-end units often sport FPGAs with firmware and drivers developed in-house. Software support for third-party chipsets has often proved dodgy over the years, resulting in either annoying bugs or driver updates ceasing early.)
Yes, for example the higher-end MOTUs (e.g. 624, UltraLite) don't use XMOS. Some people here seem to misunderstand what these dedicated chips do, and think there are all for fancy hardware effects... No. Interfaces with these chips and dedicated drivers are much more likely to have true hardware multiclient support. It means for example, multiple ASIO applications can access the interface's internal digital mixer simultaneously without additional software routing like virtual cables and such, to ensure glitch-free ultra low-latency operation, playback and recording to and from any program, route each mono channel to multiple destinations and vice versa. I can always give up a few dBs of SINAD for these features.
 
Last edited:

forma

Member
Joined
Sep 22, 2020
Messages
8
Likes
2
So, my old TC Electronic Desktop Konnekt 6 is playing up lately (dying or driver conflict wiith W10, who knows), so have been looking for a replacement. The Moto M2 seems to be a good choice by the looks of it.

However, I went to read the old spec sheet for my Konnekt 6 here and it states at the bottom of the pdf:

Headphone Output:
Power @ 40 Ohm Load: 200 mW
Power @ 600 Ohm Load: 93 mW

Am i missing something here or did this oldie just happen to have a very powerful headphone out? It's making me second guess the upgrade. Using HD600's.
 

dwkdnvr

Senior Member
Joined
Nov 2, 2018
Messages
418
Likes
698
I was thinking that this might work for a software crossover for a simple 2 + 1 subwoofer setup. I already use brutefir on my LMS server for DRC and a RPi4 running piCorePlayer for the streamer. What's not clear to me is whether LMS can stream the 3 needed channels to the RPi and whether squeezelite can then route the 3 channels to the Motu.
As far as I know, the LMS / Squeezebox world is strictly 2-channel. I don't believe there is a way to get LMS to stream 3 channels. Your best option might be to look at something like CamilaDSP which is a system-wide EQ/Xover setup for Linux similar in principle to EqAPO on windows. You'd pipe the output of Squeezelite to CamilaDSP which would to the eq/xover. There is some discussion on the diyaudio thread on how to handle sample rate changes with this chain, but I'm not current on the state of things.

I'm actually considering more or less this exact setup - already have the PI and the Motu M4. I might end up starting with a 'simple' BruteFIR setup running analog-in to analog-out to bypass the sample rate problem to start, but OTOH basically everything I have is 44.1 so I could also just accept the fixed-rate implementation limitations to begin with.
 

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,309
Likes
1,475
As far as I know, the LMS / Squeezebox world is strictly 2-channel. I don't believe there is a way to get LMS to stream 3 channels. Your best option might be to look at something like CamilaDSP which is a system-wide EQ/Xover setup for Linux similar in principle to EqAPO on windows. You'd pipe the output of Squeezelite to CamilaDSP which would to the eq/xover. There is some discussion on the diyaudio thread on how to handle sample rate changes with this chain, but I'm not current on the state of things.

I'm actually considering more or less this exact setup - already have the PI and the Motu M4. I might end up starting with a 'simple' BruteFIR setup running analog-in to analog-out to bypass the sample rate problem to start, but OTOH basically everything I have is 44.1 so I could also just accept the fixed-rate implementation limitations to begin with.

Thanks for the comments. Yeah, I'd already been informed on the slimserver forums that squeezelite is only 2 channel. Though I still wonder if there's a way to send the 2-channel squeezelite output to the input of the Motu, then back thru brutefir and then to the outputs of the Motu, or am I crazy?

It's sad, though, that the squeeze architecture still can't handle multichannel in 2020.

For dealing with different sample rates, you could just upsample everything.

http://archimago.blogspot.com/2017/12/howto-musings-playing-with-digital_23.html

I've come to the conclusion that anything above 48 kHz is pointless when my filters are generated from 48 kHz impulse responses and cut off at 24 kHz. I'd rather have everything resampled to 44.1 kHz than do without DRC.
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,486
Likes
4,113
Location
Pacific Northwest
... I've come to the conclusion that anything above 48 kHz is pointless when my filters are generated from 48 kHz impulse responses and cut off at 24 kHz. I'd rather have everything resampled to 44.1 kHz than do without DRC.
If you're going to resample, it would make sense to resample to either 44.1 or 48, whichever is an integer multiple of the original rate. This should make the resampling computationally simpler and sonically transparent.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,393
Likes
3,341
Location
.de
So, my old TC Electronic Desktop Konnekt 6 is playing up lately (dying or driver conflict wiith W10, who knows), so have been looking for a replacement. The Moto M2 seems to be a good choice by the looks of it.

However, I went to read the old spec sheet for my Konnekt 6 here and it states at the bottom of the pdf:

Headphone Output:
Power @ 40 Ohm Load: 200 mW
Power @ 600 Ohm Load: 93 mW

Am i missing something here or did this oldie just happen to have a very powerful headphone out? It's making me second guess the upgrade. Using HD600's.
Almost +20 dBu on the headphone output is in fact a lot. Presumably they have like +/-12 V internal supplies. That's the advantage is not being bus powered, your power budget is a fair bit more generous. The M2 and M4 are getting about as much out of bus power as possible in their price range. I suspect internal supplies may be +/-7..8 V tops, and they're using low(ish)-noise parts and design to achieve their dynamic range instead.

Now HD600s should be quite happy on an output that does 2 Vrms, so I wouldn't be expecting any major issues with the even higher +12.5 dBu on the M2. Do you ever get anywhere near maximum output on the Konnekt 6?
 

forma

Member
Joined
Sep 22, 2020
Messages
8
Likes
2
Thanks for the response. With the Konnekt 6, i go to about half way max. It's as loud as i'm personally willing to go on any setup. But any louder than this and it just seems to become strained anyway.

I've discovered a fair amount this past day of browsing these forums and am now eying up the Magni Heresy, Atom, and L30 as potential purchases. Then I'll look into what i can do on the DAC side. It's s hame i need a mic preamp somewhere in my setup or I'd just get a dedicated DAC.

Anyhow, it will be interesting for me to see if a well respected headphone amp with a lot more power can improve the HD600s over what i currently have.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,393
Likes
3,341
Location
.de
Anyhow, it will be interesting for me to see if a well respected headphone amp with a lot more power can improve the HD600s over what i currently have.
Let's just say I wouldn't be holding my breath. Cans like that are pretty noncritical. They're generally fine with around 2 Vrms max output and <25 ohms of output impedance, plus easy on the output stage at 300 ohms. (Keep in mind they're rated 97 dB SPL @ 1 mW. You need all but 4 mW for 103 dB and 20 mW for 110 dB SPL. That's 2.45 Vrms = +10 dBu. At the same time, they're not too fussed about 10-20 µV of output noise.) The only thing even less critical I can think of would be an ATH-R70x, similar sensitivity but at 470 ohms nominal.

So generally speaking, HD600s are among the last cans you need more than a basic amplifier for. Something MAX97220 based is just fine. The M2's output ought to be perfectly adequate, it measured well in Julian Krause's testing. The only thing one might need an extra amplifier for is more gain for material with unusually low digital levels.

It is always easy to be tempted by better performance and higher output and whatnot, but at some point output impedance is low enough for a <0.3 dB deviation in frequency response, maximum output is loud enough and noise and distortion are inaudible. If you get to this point, the gear is transparent. A better amplifier may achieve this over a wider range of loads (you might have both super sensitive BA IEMs and an AKG K340 or a Hifiman HE-6 in your arsenal, now that gets tricky), but with a lone HD600 you should find that a lot of solutions are 100% equivalent, with differences in performance being measurable but not audible.

In the upgradeitis-infested lands of audiophilia, there is never such a thing as good enough, of course. ;)

For some real audible changes, I might give parametric EQ a shot; you'll find some presets for the HD600s out there.
 
Top Bottom