• 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

Is the difference really audible? What about using 48k, which will halve file sizes (for the same bit-depth)?
I guess what he meant was the decimation filter of some ADCs are rather shallow and soft, as shown on my previous post. Analog signal with frequencies higher than half of the recording sample rate (e.g. above 24kHz for 48kHz recording) may creep into the passband, as shown here, which is called aliasing. While "aliasing" is often used to describe DAC interpolation filters that go over fs/2, in more precise terms they should be called imaging for DAC, and aliasing for ADC.

Using a higher recording sample rate means with the same ADC filter, the effect of aliasing will be reduced as this aliasing will further away from the band of interest. So after you digitized the vinyl and applied whatever restoration you may need, then you can safely downsample to 16/48 or 16/44 with a high quality software resampler. So there would be no wasted storage space.

But then, if you can't hear this aliasing in the first place, of course, it is up to you. I don't want to imply one can or cannot hear the difference, just want to describe what is going on.
 
I see 4 line outputs. Does this mean that the dac can be used as an 4-channel dac?
If so, this would be one of the best DAC's for active two way speaker duties.
 
That's what I thought, too. However, for a DAC a little ultrasonic noise is not too harmful. For ADC, it might induce sampling artifacts. ADC was tested with 48 kHz, maybe we could verify if 44.1 performs the same.
Sampling theory demands a sharp cutoff at Nyquist to avoid aliases, so I'm not sure about "not too harmful".
 
Work entirely at 48kHz. Only ever convert to 44.1kHz at the very last moment, if you have to, for delivery on CD. Otherwise, stay at 48kHz.
Makes sense. Shame that CD settled on 44k instead of 48k, presumably for space/tech reasons at the time.
 
I bought one of these shortly after release and I love it. I run the balanced monitor outputs to my t.Racks DSP204 for my speakers and the unbalanced line outputs to my JDS Labs Atom, since the headphone amp simply isn't up to par for HD650s and AKG K712 Pros.

One cool thing is that the balanced and unbalanced outputs can all be used at the same time, so I could, e.g., use the balanced line outs for a balanced headphone amp concurrently with the unbalanced line outs running to my Bottlehead Crack (although this doesn't work in practice due to my PC ground loops...).
 
Sampling theory demands a sharp cutoff at Nyquist to avoid aliases, so I'm not sure about "not too harmful".
That's the ADC part. DACs would produce ultrasonic noise that's often not audible.
 
Shame that CD settled on 44k instead of 48k
There are indeed a great many stories as to why we ended up with the apparently bizarre sampling rate of 44.1kHz, with varying degrees of basis in fact. Obviously, the technology of the time predicated the use of a sampling rate considered just adequate for purpose i.e. to obtain a flat response to 20kHz. The only practical recording technology for digital audio then available was modified videotape equipment which, by definition, was designed to operate at either NTSC (US) or PAL (European) standards. It so happened that 44.1kHz was a suitable and practical sampling rate that was sufficiently compatible with both NTSC and PAL recorders.

When 48kHz was later introduced as the standard for professional digital audio, there is some evidence to support the view that this was intentionally chosen, amongst other reasons, to be sufficiently different from the domestic standard so as to render interchange between the two a tricky proposition (which, at that time, it was), more or less impossible outside specialist facilities.
 
When 48kHz was later introduced as the standard for professional digital audio, there is some evidence to support the view that this was intentionally chosen, amongst other reasons, to be sufficiently different from the domestic standard so as to render interchange between the two a tricky proposition (which, at that time, it was), more or less impossible outside specialist facilities.

A prime reason was also that it is very easy to resample to 32kHz, which was already used for data in FM broadcast. Until the advent of arbitrary rate sample reconversion algorithms, which were another decade off, resampling needed simple ratios. 3:2 was easy. 44.1:32 not so good.
But one feels sure that the industry got some sense of smug satisfaction that it was not possible to make make bit perfect copies of CDs onto early digital tape media. A short lived victory to say the least.
Back then there was a lot of cycling of recordings through analog anyway. An all digital chain for much work was still a way off.
 
That's the ADC part. DACs would produce ultrasonic noise that's often not audible.
Thanks for the correction, but aren't these images at the very least wasting amp power and stressing the speakers, and potentially generating IMD in the audible band?
 
A prime reason was also that it is very easy to resample to 32kHz, which was already used for data in FM broadcast. Until the advent of arbitrary rate sample reconversion algorithms, which were another decade off, resampling needed simple ratios. 3:2 was easy. 44.1:32 not so good.
But one feels sure that the industry got some sense of smug satisfaction that it was not possible to make make bit perfect copies of CDs onto early digital tape media. A short lived victory to say the least.
Back then there was a lot of cycling of recordings through analog anyway. An all digital chain for much work was still a way off.
Another example of industry kneecapping technology for the sake of protecting profits.
 
You know the Scarlett gear, at least the better ones, seem to be craftily made on the edge of inaudibility. Meaning some of them are good enough you'll likely hear no difference vs better gear or it will be a tiny very slight one heard only rarely. To me that is pretty good for low cost entry level gear. They did make a couple particular versions not quite that good. But mostly Scarletts are that good.

The M2 and M4 appear to be even better and nothing wrong with a little gilding the lily. MOTU doesn't yet have anything comparably priced to the Scarlett 18i20 with 8 microphone preamps. But maybe in the near future they'll have an M8 or something like that.

It could be just bad luck, but I had a 2i2 die on me. With a bit of support I believe I could have been able to repair it. They wouldn't give me anything. I was not obviously asking for schematics, just information about one part. I decided It was not an expensive enough product to waste time with it and just disposed of it. Again, this happens, I don't say Focusrite have reliability problems, but I tought their support past warranty was a bit stiff.
 
Thanks for the correction, but aren't these images at the very least wasting amp power and stressing the speakers, and potentially generating IMD in the audible band?

I hear from time to time a bunch of people mention this really becomes a problem with power amps.

Tbh, this is why I have all my EQ's start with a low pass and high pass (respectively at 15Hz and 18kHz since I can only hear 17.5kHz anyway), Don't need any lower frequencies since I don't use subs (not a big speaker guy) and headphones don't go down that low at the volumes I listen to. Likewise with the highs, just worthless junk after 18kHz if you ask me personally.

JRiver is great for this, with a strong 48dB per octave sloping. (I thought I could do it with my RME DAC, but for whatever reason the slope factor is extremely tame, takes too long to roll off, but alas better than nothing).

Motu could have had a better filter (ESS chips in past reviews demonstrate that they have the filter option, so a simply firmware upgrade could rectify this instantly). As for what's going on in the ADC portion (since it's AKM) idk what they would need to do in order to smooth some of the late rises in distortion. That's just wrong what I said.
 
Last edited:
Is the difference really audible? What about using 48k, which will halve file sizes (for the same bit-depth)?

With the gear I had at the time it was audible yes. One could use 88khz and 16 bit which likely would be enough. The effect at 48 khz wasn't huge, but in very careful listening you could hear a difference.

Now when I first ran across this I didn't think you would need more than 48 khz, but you did. And with advancing age maybe you no longer do need the extra bandwidth.
 
These lazy filters (flat to 21k not 20k, and reaching full attenuation at 24k instead of 22k) look like they're meant for 48k sampling and they couldn't be bothered to provide a proper one for 44k. Thoughts?

No... there's only one set of filter coefficients... at 48kHz, the stopband will begin at 0.54fs, i.e., 26.2kHz. Just as "lazy" in the sense of the word you're describing. Except it isn't lazy... it's an intentional design choice. At 44.1 and 48kHz fs, there is no possible digital filter that simultaneously provides flat frequency response to 20kHz, linear phase, fs/2 stopband, and good time domain performance. Filter designers have to weigh the options and pick a middle ground. This is why the very expensive AK4499 also doesn't include a sinc filter with fs/2 stopband, even though it would be exceptionally easy for the AKM engineers to include.

There's a lot of misinformation about digital filters on this forum unfortunately.
 
Last edited:
Motu M2 specs
https://motu.com/en-us/products/m-series/m2/specs/

[ 1x headphone out (driven by ESS converters) with independent volume control ]

It says nothing about a volume of the outputs. I suppose, as they have commented, that the level is fixed -> Does not work as cheap preamp / attenuator. Am I wrong?
The M2 outputs have volume control with the big knob.
 
I hear from time to time a bunch of people mention this really becomes a problem with power amps.

Tbh, this is why I have all my EQ's start with a low pass and high pass (respectively at 15Hz and 18kHz since I can only hear 17.5kHz anyway), Don't need any lower frequencies since I don't use subs (not a big speaker guy) and headphones don't go down that low at the volumes I listen to. Likewise with the highs, just worthless junk after 18kHz if you ask me personally.

JRiver is great for this, with a strong 48dB per octave sloping. (I thought I could do it with my RME DAC, but for whatever reason the slope factor is extremely tame, takes too long to roll off, but alas better than nothing).

Motu could have had a better filter (ESS chips in past reviews demonstrate that they have the filter option, so a simply firmware upgrade could rectify this instantly). As for what's going on in the ADC portion (since it's AKM) idk what they would need to do in order to smooth some of the late rises in distortion.
The M4 uses ES9016. IIRC until ES9028 or newer these ESS chips only have fast and slow filters and none of them fully attenuates at fs/2.
index.php
 
The M4 uses ES9016. IIRC until ES9028 or newer these ESS chips only have fast and slow filters and none of them fully attenuates at fs/2.


Why Motu, why >_<

Thanks for the correction, I had no idea.

EDIT: Any idea what's causing the rise in ultrasonics? Would've at least thought such a deep filter decimates properly, but it seems to lay off for later frequencies as hinted by the THD vs FR. Another inherent chip flaw of earlier chips or?
 
Curious about how the Scarlett Solo gen3 would fare against the Motu M2 now :)
 
Back
Top Bottom