• 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.

Measurements Of SMSL M8 DAC

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
2,121
Likes
863
Location
Zg, Cro
#41
As long as they are not used in debates to convince the validity of such, it is OK to state them.
Are you saying we should accept somebodies claim to hear the difference between modern DACs as a fact?

Just to remind you, here's the claim: "The SMSL M8A v3 is bright / lacking in bass. "
And yet, no frequency response curve to backed that up..
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
16,836
Likes
13,190
Location
Seattle Area
#42
Are you saying we should accept somebodies claim to hear the difference between modern DACs as a fact?
No. If the statement is not used to counter objective data, or disprove science, then let's not make a battle out of it.
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
2,121
Likes
863
Location
Zg, Cro
#43
No. If the statement is not used to counter objective data, or disprove science, then let's not make a battle out of it.
Basically, you've got the point - as with flat earthers, you can't really prove to the people they can't hear the difference between DACs, so better leave them alone as long as they are not trying to make you and everybody else believe the earth is flat.

The only problem with that logic is, when they make such post in a public forum, they actually are trying to make us believe that.
 
Joined
Jun 20, 2018
Messages
313
Likes
102
#44
Just to remind you, here's the claim: "The SMSL M8A v3 is bright / lacking in bass. "
And yet, no frequency response curve to backed that up..
This was posted as part of my description labeled as my subjective finding relative to the sound of the D2. I did not make that claim out of this context (or at least tried not to). And neither have I meant to say that it is scientific fact that the M8A v3 is bright / lacking in bass. All I claim is that it is clearly audible to me that my M8A v3 is brighter and has less bass than my D2. I'll make another claim, my M8A v3 also sounds brighter and with less bass than my HP Probook laptop audio out sound to me.
To me it seems like my particular M8A v3 is off in frequency response (or something else which gives a perceived deviation from a flat frequency response), but who knows perhaps both my D2 and laptop out are off, not likely to me but could be.
I only wanted to post this as a warning to other forum members that it's quite possible that other M8A v3 out there have the same thing and that possibly the M8A v3 does not measure as well as the M8A v1 did. Absolutely not sure, but my personal experience makes me think it could be the case..
Sorry if I wasn't clear enough about this and indeed didn't want to start a personal debate about but my apologies for continuing it and won't post more on the matter.
And no hard feelings btw, I like questioning subjective audio claims in general, perhaps I took it a bit too personal :)
Btw, I don't have any ADC here now. Can get one but tomorrow my M8A v3 goes back in the mail as a return. Otherwise I would have tried to measure it all. (well actually, I have a bad "mic in" input on my laptop, perhaps I can give it a quick go if I have time this evening)
Cheers!
 

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
2,121
Likes
863
Location
Zg, Cro
#45
This was posted as part of my description labeled as my subjective finding relative to the sound of the D2. I did not make that claim out of this context (or at least tried not to). And neither have I meant to say that it is scientific fact that the M8A v3 is bright / lacking in bass. All I claim is that it is clearly audible to me that my M8A v3 is brighter and has less bass than my D2. I'll make another claim, my M8A v3 also sounds brighter and with less bass than my HP Probook laptop audio out sound to me.
To me it seems like my particular M8A v3 is off in frequency response (or something else which gives a perceived deviation from a flat frequency response), but who knows perhaps both my D2 and laptop out are off, not likely to me but could be.
I only wanted to post this as a warning to other forum members that it's quite possible that other M8A v3 out there have the same thing and that possibly the M8A v3 does not measure as well as the M8A v1 did. Absolutely not sure, but my personal experience makes me think it could be the case..
Sorry if I wasn't clear enough about this and indeed didn't want to start a personal debate about but my apologies for continuing it and won't post more on the matter.
And no hard feelings btw, I like questioning subjective audio claims in general, perhaps I took it a bit too personal :)
Btw, I don't have any ADC here now. Can get one but tomorrow my M8A v3 goes back in the mail as a return. Otherwise I would have tried to measure it all. (well actually, I have a bad "mic in" input on my laptop, perhaps I can give it a quick go if I have time this evening)
Cheers!
This is only audio talk so no hard feelings from my side M8, hope it's the same with you. :)

Cheers!
 
Joined
Jun 20, 2018
Messages
313
Likes
102
#46
I tried to measure relative frequency response using my HP Probook line/mic input but alas I can't get the input to work :(

I have done another subjective A/B comparison this time using my nearfield speakers (small ones, "analytical" sounding, severely lacking bass below 80Hz or so.) Level matched and was listening only to frequency differences. And I have to say.. I have exaggerated the difference.
The difference is there for me but it is subtle. The M8A v3 sounds slightly more "forward" in the whole trebble range, the D2 slightly more "relaxed" to me. As for the midbass the M8A v3 sounds slightly more "lean" to me and the D2 slightly "fuller". Subbass I couldn't tell with my nearfields.
It's more of a slightly different "presentation".. Perhaps both measure flat and the difference in sound is due to I don't know what.. difference in Jitter perhaps? Difference in DAC chip perhaps if they have an inherent sound as many people seem to think? Btw tested both with 176.4KHz upsampling.
With the difference being subtle, I do think it is an important difference for me personally. A subtle thing which matters psychoacoustically for me.
That said, I even slightly preferred the M8A v3 with some tracks. Have some old digitzed tape tracks which are a bit muffled sounding with also a big tape bump in the bass. With the presentation of the M8A they sound warm but not so much "muffled" to me, with the D2 the slightly different presentation tells me the bass bump is a bit too much and there's a bit of a muffled sound.
Ah well.. in any case. Apologies for the fuzz coming from my exaggerated initial post.
 
Joined
Jun 20, 2018
Messages
313
Likes
102
#48
How exactly do you perform your "subjective A/B comparison"? :)
I had both the M8A v3 and D2 on my desk one on top of the other. Both at full volume. 2Vrms output. I listen through Foobar2000 ASIO output and use the Foobar2000 volume control (roughly -24dB for most tracks). I had both of them connected to a short cinch to female stereo mini jack and have a male stereo mini jack to cinch to my powered nearfield speakers.
To switch I lean foward, pull the male stereo mini jack out of one DAC and connect it to the other, then select the ASIO output of that DAC through Foobar2000 (leave the selection window open) and click apply. Making sure my head is in the same position. The total switch takes me less than 5 seconds.
Sometimes I left the track running sometimes I repeated the beginning.
Perhaps it is the different DAC design I'm hearing.. And think that perhaps the D2 is slightly forward in the upper mids now. I can imagine what I'm hearing being the effect of the M8A v3 running one channel of the DAC per channel and the D2 running 4 channels of the same sounding DAC per channel and the way they combine gives it a different "sound" / perspective. The reason I say this is because in some very remote way the to me audible difference reminds me of working with synthesizers and combining several slightly detuned oscilators vs one oscillator. Don't expect anybody to take this description serious :) I mean, of course I don't hear any phasing like when combining oscillators, but it does soften up the trebble in a certain way and makes it bigger. Just reminds me of that in a very remote way.. And I think now that I heard it in the Yulong D200 and Melokin DA1 as well. It does sound better to me in a way yet perhaps also less "pure" in some other way.
Yes you may laugh ;)

Edit: I don't know much about DAC design. But could it be that a delta sigma DAC like the ESS DACs actually have many random (or semi random but correlated) errors in their output, well above human hearing threshholds, which average out in a particular way (over a short timespan) to give a very low error on a sustained signal like a sine? Could well be? But if something like this is the case then there are serious errors in the transients of a musical signal. And it could well be that this is audible to the human ear, and then it could also well be that it is audible if several of these signals with audible transient errors are combined all with different errors like possibly is the case with the 4 internal DACs per channel of the full ESS chips (not mobile versions) like the ES9018 in the case of the D2? Just speculating here what could be the root cause of the to me audible difference..

Edit 2: What happens when we test a DAC with white or pink noise instead of a 1KHz sine? I'd really love to know this. Will we get roughly the same distortion figures? Why have I never seen a DAC spec tested with white or pink noise? Pink noise in particular is much closer than a sine to an actual musical signal it seems to me? Please someone do this! :)
 
Last edited:
Joined
Jun 20, 2018
Messages
313
Likes
102
#49
What happens when we test a DAC with white or pink noise instead of a 1KHz sine? I'd really love to know this. Will we get roughly the same distortion figures? Why have I never seen a DAC spec tested with white or pink noise? Pink noise in particular is much closer than a sine to an actual musical signal it seems to me? Please someone do this! :)
Thinking about it some more, I think it's rediculous this isn't being done yet!
Sure you probably wouldn't be able to tell some forms of distortion as well as with a sine test, but you could see other things you don't see well with sines. And it seems to me it is fully relevant to the actual performance of a DAC with music.
You can run a white or pink noise test signal and say what the maximum error is in % or dB, and what the average error is in % or dB and who knows what kind of math can be let loose on the test result showing certain correlations for instance.
I think it's quite possible, hell likely even, the results would be shocking and paint a very different yet usefull picture in addition to the current sine tests.
 
Last edited:

andreasmaaan

Major Contributor
Patreon Donor
Joined
Jun 19, 2018
Messages
2,469
Likes
1,569
#50
Thinking about it some more, I think it's rediculous this isn't being done yet!
Sure you probably wouldn't be able to tell some forms of distortion as well as with a sine test, but you could see other things you don't see well with sines. And it seems to me it is fully relevant to the actual performance of a DAC with music.
You can run a white or pink noise test signal and say what the maximum error is in % or dB, and what the average error is in % or dB and who knows what kind of math can be let loose on the test result showing certain correlations for instance.
I think it's quite possible, hell likely even, the results would be shocking and paint a very different yet usefull picture in addition to the current sine tests.
White and pink noise consist of information at all frequencies. How would you discern non-linear distortion from signal in the output?
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,064
Likes
2,050
Location
Riverview, FL
#51
Thinking about it some more, I think it's rediculous this isn't being done yet!
I suspect it has not been found to be useful.

Just because it isn't used, and it popped into your head as a possible test signal, doesn't mean it hasn't been tried, and rejected.
 
Joined
Jun 20, 2018
Messages
313
Likes
102
#52
White and pink noise consist of information at all frequencies. How would you discern non-linear distortion from signal in the output?
Well perhaps there is a way to single this type of distortion out, perhaps there is not.
No problem if there's not. It doesn't have to show all the same types of distortion as a sine test. It is a complimentary test.
Test both with sine and then also with white or pink noise?
Music is not one or the other but a combination of the two.
 
Joined
Jun 20, 2018
Messages
313
Likes
102
#53
I suspect it has not been found to be useful.

Just because it isn't used, and it popped into your head as a possible test signal, doesn't mean it hasn't been tried, and rejected.
I don't think it's likely that it's not useful in one way or another.
Perhaps it is harder to implement? Perhaps it shows DACs in a very unfavorable light? Perhaps everybody is following the standards? etc.
 

dc655321

Senior Member
Patreon Donor
Joined
Mar 4, 2018
Messages
320
Likes
217
#54
Thinking about it some more, I think it's rediculous this isn't being done yet!
Please consider learning a little of the theory and practice of how signals and systems function.
There are likely some really good introductory sources available on The Web.
If you care to share what your education/experience is with STEM, I may be able to recommend some reading.

And then think about it some more.
 

Blumlein 88

Major Contributor
Joined
Feb 23, 2016
Messages
4,993
Likes
3,566
#55
I don't think it's likely that it's not useful in one way or another.
Perhaps it is harder to implement? Perhaps it shows DACs in a very unfavorable light? Perhaps everybody is following the standards? etc.
I regularly use white noise to check frequency response in addition to other methods. I could show some plots, but noise is not a good signal for much of any other purpose.
 
Joined
Jun 20, 2018
Messages
313
Likes
102
#56
Well I'd be really curious to simply see a white or pink noise signal tested and a frequency response and THD+N based on this.
Why would it not be useful for a THD+N measurement? I can imagine jitter / any timing error throwing it off completely giving a very high THD+N. Still I'd like to see it.
And I can imagine one can compensate for any timing errors (this may take considerable time and effort though to code for this, perhaps this is a major reason why it doesn't exist yet?) and then I'd be very curious to see the THD+N after this compensation.. May very well not be the same as with a sine at all with DACs. I actually suspect the result will be far worse purely based on listening experience and some naive logic :)
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
6,064
Likes
2,050
Location
Riverview, FL
#57
I thought "how can a swept sine wave reveal transients?"

So I digitally created, played back, and recorded a single sample impulse and a stepped waveform in the air.

The calculated-from-swept-sine results for Impulse and Step response matched the recorded-from-actual-signal results.

See here...

I wouldn't underestimate the value of the tools that are commonly used.
 
Last edited:

Blumlein 88

Major Contributor
Joined
Feb 23, 2016
Messages
4,993
Likes
3,566
#58
Well I'd be really curious to simply see a white or pink noise signal tested and a frequency response and THD+N based on this.
Why would it not be useful for a THD+N measurement? I can imagine jitter / any timing error throwing it off completely giving a very high THD+N. Still I'd like to see it.
And I can imagine one can compensate for any timing errors (this may take considerable time and effort though to code for this, perhaps this is a major reason why it doesn't exist yet?) and then I'd be very curious to see the THD+N after this compensation.. May very well not be the same as with a sine at all with DACs. I actually suspect the result will be far worse purely based on listening experience and some naive logic :)
Your logic is indeed naive. Take a look at what Ray posted.

If we have all frequencies at once (white noise) and we look for THD+N, then we have plenty of N while THD would have to be horrendous to show up. I suppose it would show up as a rising frequency response on the noise which doesn't show up in sine wave sweeps. Each frequency creating harmonics at higher frequencies would mean frequencies would pile up at the high end. If bad enough it would convert white noise to violet noise. But we don't see that. I have a few DACs I fed white noise to and recorded. I can show a spectrum of the noise and a spectrum of a swept sine and they are for all practical purposes identical.
 
Joined
Jun 20, 2018
Messages
313
Likes
102
#59
I thought "how can a swept sine wave reveal transients?"

So I digitally created, played back, and recorded a single sample impulse and a stepped waveform in the air.

The calculated-from-swept-sine results for Impulse and Step response matched the recorded-from-actual-signal results.

See here...

I wouldn't underestimate the value of the tools that are commonly used.
This doesn't show what I'm talking about at all?
And I'm not surprised about the impulse response similarity. The two techniques have indeed been used for a long time already. Impulse response and swept sine to measure and generate an impulse response of a room etc.
The two are widely known to not give the exact same sounding result though. The result is similar though just as your test showed a very similar but not identical result. (though this is likely due to the difference in devices and their radiating characteristics used to generate a sine sweep vs an impulse)
Edit: what I mean is if you were able to show a THD+N of your two results it would be very high, I can clearly see the differences in the waveforms.
 
Last edited:
Joined
Jun 20, 2018
Messages
313
Likes
102
#60
Your logic is indeed naive. Take a look at what Ray posted.

If we have all frequencies at once (white noise) and we look for THD+N, then we have plenty of N while THD would have to be horrendous to show up. I suppose it would show up as a rising frequency response on the noise which doesn't show up in sine wave sweeps. Each frequency creating harmonics at higher frequencies would mean frequencies would pile up at the high end. If bad enough it would convert white noise to violet noise. But we don't see that. I have a few DACs I fed white noise to and recorded. I can show a spectrum of the noise and a spectrum of a swept sine and they are for all practical purposes identical.
What you say would only hold for harmonic distortion. It would not hold at all for intermodulation distortion and it would not hold for semi random correlated errors of a DAC either (different from noise) it seems to me. Neither would it hold for timing errors / jitter.
(edit: and harmonic distortion is probably the least audible type of distortion there is.. and sine tests fit it perfectly don't see any advantage white noise based test could give over testing harmonic distortion but nice that you argue it could be distinguished from other distortion types even with a white noise test. In fact, there are reasons to think that strong harmonic distortion with music is mostly perceived pretty much like you describe by the human ear, as a rising frequency response :) )
 
Last edited:
Top Bottom