• 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.

Musical Fidelity MX-DAC Review (Balanced DAC)

sarumbear

Major Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
1,787
Likes
1,603
Location
Southampton, UK
I thoroughly enjoy Amir's deep dives into products, but the comments section does tend to get a bit lively!
Ha! Try Reddit to see what lively comments mean. ASR comments are more sedate than the English House of Lords :)
 

Sheriff1972

Member
Forum Donor
Joined
May 29, 2019
Messages
24
Likes
23
Location
United Kingdom
Ha! Try Reddit to see what lively comments mean. ASR comments are more sedate than the English House of Lords :)
Well i normally just sit back and just enjoy reading the squabbling... (a bit like prime ministers questions on a Wednesday). Not sure why i jumped in on this thread, but i had a few mins spare - that will teach me....
 

pjug

Major Contributor
Forum Donor
Joined
Feb 2, 2019
Messages
1,271
Likes
1,057
Which experts are those? I have passed such tests and there is peer review published paper that indicates audibility.

Here are the public tests I have taken:

---
foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/08/02 13:52:46

File A: C:\Users\Amir\Music\Archimago\24-bit Audio Test (Hi-Res 24-96, FLAC, 2014)\01 - Sample A - Bozza - La Voie Triomphale.flac
File B: C:\Users\Amir\Music\Archimago\24-bit Audio Test (Hi-Res 24-96, FLAC, 2014)\02 - Sample B - Bozza - La Voie Triomphale.flac

13:52:46 : Test started.
13:54:02 : 01/01 50.0%
13:54:11 : 01/02 75.0%
13:54:57 : 02/03 50.0%
13:55:08 : 03/04 31.3%
13:55:15 : 04/05 18.8%
13:55:24 : 05/06 10.9%
13:55:32 : 06/07 6.3%
13:55:38 : 07/08 3.5%
13:55:48 : 08/09 2.0%
13:56:02 : 09/10 1.1%
13:56:08 : 10/11 0.6%
13:56:28 : 11/12 0.3%
13:56:37 : 12/13 0.2%
13:56:49 : 13/14 0.1%
13:56:58 : 14/15 0.0%
13:57:05 : Test finished.

----------
Total: 14/15 (0.0%)

As you see, 14 out of 15 right which is almost perfect.

And Mark's test tracks he produced for a test on AVS:
foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/07/10 18:50:44

File A: C:\Users\Amir\Music\AIX AVS Test files\On_The_Street_Where_You_Live_A2.wav
File B: C:\Users\Amir\Music\AIX AVS Test files\On_The_Street_Where_You_Live_B2.wav

18:50:44 : Test started.
18:51:25 : 00/01 100.0%
18:51:38 : 01/02 75.0%
18:51:47 : 02/03 50.0%
18:51:55 : 03/04 31.3%
18:52:05 : 04/05 18.8%
18:52:21 : 05/06 10.9%
18:52:32 : 06/07 6.3%
18:52:43 : 07/08 3.5%
18:52:59 : 08/09 2.0%
18:53:10 : 09/10 1.1%
18:53:19 : 10/11 0.6%
18:53:23 : Test finished.

----------
Total: 10/11 (0.6%)

Again, 10 out of 11.
Did you crank up low level sections to be able to do this? I thought I read that that is how you were able to tell the difference but I don't know if that is accurate. If that is how you did it, I would say it does show you were able to show an audible difference, but not as a practical matter. To do that you would have to use the same volume you would use playing the whole track.
 

Gorgonzola

Senior Member
Joined
Jan 27, 2021
Messages
344
Likes
420
Location
Ontario, Canada
well... sorry to tell you, but without proper scientific data your conclusion worths nothing at all.
No need to apologize.

I endorse science. Most scientific research evolves from conjecture, to hypothesis, to theory. As I admitted, my assertion is conjecture.
 

Gorgonzola

Senior Member
Joined
Jan 27, 2021
Messages
344
Likes
420
Location
Ontario, Canada
Yes, I doubt it.
1. Unless it's well over 0.5%, all the research any of us have seen says it's not likely audible. So -70 dB (0.03%) in a decent tube amp? sorry, nobody can hear it.
2. Anecdotes without data are fun but irrelevant.
3. I like it in my (own) electric guitar, but not in my (recordings of) cymbals and especially not in my violin solos.
I guess my picking on the Musical Fidelity DAC was poor instance. I agree that its distortion is real but so low that it is likely in audible to anyone. Your personal experience, however, (anecdotal as it is), implies there is some level at which some type of distortion is audible and agreeable.

You mention 0.5%, maybe so. But consider that many tube amps and some others exceed 1.0% especially at high output levels and higher frequencies.
 

rdenney

Major Contributor
Forum Donor
Joined
Dec 30, 2020
Messages
1,123
Likes
1,751
But again, the V90 from the same company, which is from 2013, performs better.
This is a big point, and the best argument against my request for a more moderate use of words in condemning this model. The v90 is older, cheaper, and performs 10 dB better. But it doesn't have the same feature set.

Rick "what panther would they have gotten with a SINAD of 110 dB?" Denney
 
Last edited:

sarumbear

Major Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
1,787
Likes
1,603
Location
Southampton, UK
This is a big point, and the best argument against my request for a more moderate use of words in condemning this model. It's older, cheaper, and performs 10 dB better. But it doesn't have the same feature set.

@amirm concluded with the following words. Which particular word you are disagreeing with?

Conclusions
The website description says, "The MX-DAC is a tremendous performer. It has ultra low distortion, very wide bandwidth and low noise. The technical performance of the MX-DAC is on a par with any other DAC at any price." At any price? There are DACs at less than 1/7th the price of MX-DAC which run circles around it in every objective score. Next sentence states a bunch of specs which as I explained, is not in sync with their one manual or what I measured. Regardless, this is embarrassing level of performance today. Company needs to retool this platform and build something competitive.

I can't recommend the Musical Fidelity MX-DAC.
 

rdenney

Major Contributor
Forum Donor
Joined
Dec 30, 2020
Messages
1,123
Likes
1,751
Okay, Sarumbear, I'll leave it to you. You are right and I am wrong.

Rick "feel better?" Denney
 

rdenney

Major Contributor
Forum Donor
Joined
Dec 30, 2020
Messages
1,123
Likes
1,751
What does this mean?

This product is very expensive and many DAC's much cheaper do the same job, yet they were able to obtain great performance for little cost.

MF should pull this model.



JSmith
Do not confuse price and cost. There are all sorts of influences on cost that are utterly unrelated to the quality of the product or the price it deserves in the market.

If this product is overpriced, it won't sell. But I agree that Musical Fidelity shouldn't lie about its capabilities, and sites like this provide measurements that put the truth to that.

Rick "the market needs accurate product information in addition to be competitive" Denney
 

rdenney

Major Contributor
Forum Donor
Joined
Dec 30, 2020
Messages
1,123
Likes
1,751
Yes, that’s a good point if you’re listening to the source untouched. There’s no real benefit in having more than 16bits in the source material.
But you want your DAC to provide headroom for DSP. You want to be sure that it will continue to provide full 16bit performance even after the data has been scaled down by EQ, ReplayGain and possibly a digital volume control. That’s the benefit of using a DAC that can handle 20bits or more. Of course this implies that you’re making up the gain in the analog domain, so that part of the system needs a low noise floor as well.
Okay, this is interesting. I'm asking because I don't know--my system is an old-school analog system, for the most part. I use a digital PEQ, but it connects in an analog processor loop in my preamp (with a noise floor in the 90's or better).

If I apply PEQ in my streaming software (let's say) does that increase the required bit depth in the DAC? Or does it just (in effect) merely remap the bits that are there to different output voltage levels? (In the photography world, for example, we routinely work at 24-bit color depth for each of the three primary colors plus black. But by the time it goes to a printer, it's downconverted back to 8-bit color, because that's all the printer can distinguish. But those 8 bits cover the full gamut capability of the printer. More bits provide more precision, but accuracy is a different parameter.)

If so, this is a use case that brings with it requirements for a DAC that provides enough bit depth to play the music, plus enough bit-depth headroom to accommodate processing at playback. I've just never seen it described that way.

Rick "finding (again) that understanding use cases is necessary for identifying requirements and evaluating designs" Denney
 
Last edited:

pjug

Major Contributor
Forum Donor
Joined
Feb 2, 2019
Messages
1,271
Likes
1,057
Shrugging.

Or piggybank raiding.
The Pro-ject Pre Box DS2 Digital measured 107dB SINAD, costs ~$900, and got the golfing panther. That has more functionality though, and I think Amir liked the all-around competence.
 

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,491
Likes
3,465
Location
Michigan
I keep seeing the comments referring that $9 apple dongle to so many other products in ASR.
When will you guys understand that the dongle is a different breed in comparison with let's say speaker amplifiers, mono amplifiers, preamps, desktop DACs, and more? I too think this is overpriced but the way some folks here bash the product(s) reminds me how some of the hate for ASR from outsiders is not unwarranted.

Also, this product was released several years ago, at that time the race toward flawless measurements hasn't kicked into gear yet. You can say that the ASR recommendation from Amir is in the current year and for the current market and audience then I will agree. But back then there simply weren't too many great products so it isn't like this DAC is garbage for it's time, or even in the current market.
Straw man.

I compare it to the Topping D90se as a fully featured product at the price point and then compare its SINAD to the dongle.

I would compare the whole product to a Schiit Modius, if you want to talk value, since Modius is way better.

The dongle does get you better engineering and implementation though.

I did make a mistake. The dongle was only $8. Its the Sony headphone thas $9.
 

charleski

Senior Member
Joined
Dec 15, 2019
Messages
402
Likes
665
Location
Manchester UK
If I apply PEQ in my streaming software (let's say) does that increase the required bit depth in the DAC? Or does it just (in effect) merely remap the bits that are there to different output voltage levels?
It's the first one. Your DAC has no knowledge of what sort of manipulation you've done upstream of it. All it can do is decode the bits that are presented to it.

Let's say you want to add a 6dB bass shelf (6dB = 1bit). Since digital signals can't go over 0dBFS what the EQ actually does is reduce everything by 6dB, then boost the bass only up 6dB (most decent EQs have a preamp function where you can set this specifically). Let's say your music uses the full 96dB dynamic range offered by 16bits. This has now changed to a 17bit signal with a dynamic range of 102dB to accommodate the extra 6dB in the bass. If you fed this to a DAC that only has 16bits' worth of SNR then the quietest passages would now get lost in the noise floor. The EQ will be working at 32bits internally, and your player software should be sending out a 24bit signal to the DAC. The bottleneck in terms of SNR happens at the DAC: a SOTA one can manage 22bit decoding, but it's easy to find DACs in the $100-200 range that can do 20bits.

There's also the issue of ReplayGain, which can easily apply 10dB or more of gain reduction, and this happens on top of whatever you have set in your EQ. Of course heavily-compressed tracks that end up with a lot of RG correction are probably not using the full dynamic range offered by 16bits, but it's nice to know my system has the headroom needed to accommodate it.
 

rdenney

Major Contributor
Forum Donor
Joined
Dec 30, 2020
Messages
1,123
Likes
1,751
It's the first one. Your DAC has no knowledge of what sort of manipulation you've done upstream of it. All it can do is decode the bits that are presented to it.

Let's say you want to add a 6dB bass shelf (6dB = 1bit). Since digital signals can't go over 0dBFS what the EQ actually does is reduce everything by 6dB, then boost the bass only up 6dB (most decent EQs have a preamp function where you can set this specifically). Let's say your music uses the full 96dB dynamic range offered by 16bits. This has now changed to a 17bit signal with a dynamic range of 102dB to accommodate the extra 6dB in the bass. If you fed this to a DAC that only has 16bits' worth of SNR then the quietest passages would now get lost in the noise floor. The EQ will be working at 32bits internally, and your player software should be sending out a 24bit signal to the DAC. The bottleneck in terms of SNR happens at the DAC: a SOTA one can manage 22bit decoding, but it's easy to find DACs in the $100-200 range that can do 20bits.

There's also the issue of ReplayGain, which can easily apply 10dB or more of gain reduction, and this happens on top of whatever you have set in your EQ. Of course heavily-compressed tracks that end up with a lot of RG correction are probably not using the full dynamic range offered by 16bits, but it's nice to know my system has the headroom needed to accommodate it.
Okay, thanks. I had not thought all that through.

My DSP-based PEQ is an old commercial-sound Yamaha YDP2006, and has analog inputs and outputs. The digital part has 20-bit resolution, but it's making the analog conversion going in and coming out, and expecting me to adjust the overall gain structure to prevent clipping. If I've lost anything in the noise floor of that device, I sure can't tell it using only my ears.

Rick "who therefore provides EQ downstream from the DAC in the playback chain" Denney
 

KSTR

Major Contributor
Joined
Sep 6, 2018
Messages
1,547
Likes
3,297
Location
Berlin, Germany
-0.5 dB at -115 dBFS is about -139.5 dBFS, which is getting awfully close to the limits of 24 bits. Besides, we are talking multibit delta-sigma converters here, and those are trying hard to eliminate their converter nonlinearity but it may not be 100% perfect. If you look at the AD1955 datasheet, linearity curve TPC 12 even shows a minimal negative deviation if you look closely, and it's actually even worse at about -1 dB at -120 dBFS. In 2002 when this part came out, perfect low-level linearity still was not a given.
Yeah, Plot #TPC12 shows some microscopic drop at -120dBFS but it is hard to say if this is representative. At -125dBFS we're already in the noise limit of the instrumentation of the day which most probably was an AP System Two model (23xx or 25xx), at least when using the default parameters for the linearity test (the raw data of which is the source of this plot)
Only way to get down to bit 24 is true bit step test with tons of time-domain averaging to get deep down into the noise... might do that when time allows. You are right, in 2002 the multibit D/S-DACs probably didn't use the elaborate digital weighting averaging schemes of today to trade linearity for noise and then some low-level nonlinearity (and distortion) is plausible.
 

Swtoby

Active Member
Joined
Jun 13, 2018
Messages
136
Likes
192
Just because you cannot doesn't mean it is inaudible. There are other humans on the planet besides you. Besides, where do you see the word broken?
This is exactly the point of the expensive cable manufacturers. And the power conditioner people, the Shakti stone maker, the carbon fiber platform purveyors, etc, etc.;)
 

nyxnyxnyx

Active Member
Joined
May 22, 2019
Messages
126
Likes
113
Straw man.

I compare it to the Topping D90se as a fully featured product at the price point and then compare its SINAD to the dongle.

I would compare the whole product to a Schiit Modius, if you want to talk value, since Modius is way better.

The dongle does get you better engineering and implementation though.

I did make a mistake. The dongle was only $8. Its the Sony headphone thas $9.
I agree about value, the rule of thumb about value is that the cheaper it is, the better value ratio it has, then actual performance comes second. Boutique and TOTL products will never be able to compare the p/p ratio with 99% of other products in the spectrum, mainly because they were created with a hint of fanciness in mind, or something to attract the rich fellas.

Every comparison of this DAC with other desktop DACs is fair, I'm just a bit annoyed with the apple dongle talk since people bring it up to compare so often. Secondly, I don't get why people actually bash this actual product when it was kinda neat back in the days it first came up. Of course, in 2021 it doesn't hold a candle with the great bargains we have now but it still lands in the "fine" category which isn't all too shabby.

I found one thing even weirder that I've read this thread and only see people bashing the performance or value ratio, but not the false advertising/ false measurements from the company. To me, that's the biggest problem of all this, since the real product is fine but it was hyped up insincerely.
 

MaxBuck

Addicted to Fun and Learning
Forum Donor
Joined
May 22, 2021
Messages
563
Likes
625
Location
SoCal, Baby!
In re all the discussions of "audibility," I think it's likely that some sounds that are inaudible may nonetheless be perceptible. Otherwise why would DAC filters that don't even engage within the audible range affect the perceived sound quality? And even @amirm has acknowledged that filters can affect how a DAC "sounds" (which I think involves things other than those in the 10Hz-18kHz range).
 
Top Bottom