• 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!

CEntrance Hifi-M8 V2 Review (DAC and headphone amplifier)

MerlinGS

Active Member
Forum Donor
Joined
Dec 28, 2016
Messages
131
Likes
271
Anecdotal I said, but are you trying to say that the example I gave was just pure luck? since we are in a science based place, what would be the probability of this?
When you consider the variables introduced by the room, speakers and source (I take it we are talking yrs back and the source was analogue), the differences between a Yamaha and a NAD were probably not as easy to determine as you suggest. As to you correctly guessing it, there are numerous explanations that can be proposed: a) NAD was popular among frugal audiophiles and you subconsciously picked up that this person cared about audio, but also cared about cost; b) you clearly liked the sound, NAD was popular so there was a chance that it would be NAD; c) someone may have suggested it or you overheard it but you did not consciously process it, but it surfaced when you had to guess (especially since you like NAD); d) the speakers you saw were frequently sold with NAD...I could continue speculating, but you get the gist of it, your guess has little scientific value, it is just a nice narrative that helped validate certain experiences and beliefs.

PS, please note that the above does not argue you could not differentiate both amps in a controlled test, it merely provides a number a valid alternative as to why your guess was correct. However, the concerns stated regarding the variables introduced by speakers, room and source, are reasonable considerations to question your conclusion that you could easily identify NAD equipment the described context. An unfamiliar environment (in the context of you testing NAD) in conjunction with the other varying unknown variables (i.e. speakers that you may know, but were not familiar with in that environment, the sonic signature of the source, etc.) will introduce far greater sonic differences than most competently designed amps of the period might display (especially in a noisy environment). The "house" sound (assuming there is one) you identify as differentiating the two brands is much smaller than the sound signatures introduced by the other variables.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,667
Likes
241,033
Location
Seattle Area
With all respect Amir, you can't have it both way, Every audio gear reviewer out there are trained listeners, but people don't come here for subjective evaluations.
None of those reviews are properly trained professionally to hear small impairments as I have. It used to be part of my job and career. And what I have demonstrated time and time again online in countless challenges. See: https://www.audiosciencereview.com/...-and-perceived-sound-quality.5431/post-146493

I have taken both Archimago and AIX Records double blind tests using foobar ABX plug in and passed most of them:

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.

This doesn't mean high-res sounds better than CD. It however means I know what to listen for and pass such difficult tests. :)
The reviewers you talk about have no prayer of passing such tests which is the reason they never participate in them. As such, none are trained listeners. They say they are, but that is a false claim not backed by any proof.

Importantly, I hope no one believes what you said about subjective evaluations. We absolutely, positively are here for subjective evaluations. They just need to be controlled tests, like the ones I listed above and especially so if the claims don't seem plausible.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,667
Likes
241,033
Location
Seattle Area
Way way back when I was a (poor) college student, maybe not a trained listener, but certainly already interested in audio, I was a big fan of NAD gear, I remember going to a party where the sound system was hidden in a cabinet. I was talking to the host and said, we are listening to a NAD right? Of course we where, why, because I had been living with NAD gear for years. I wouldn't have been able to identify any other sound system but I had a long time relation with NAD. I am a trained listener too, and hey there are many audio professionals in this community, and I honestly don't think that in a blind test I could identify reliably which sounds better between a 0.05 and a 0.02 THD, but I can certainly hear the difference between a classic NAD and a Classic Yamaha.
"With all due respect" :), no way, no how you can do that. Above was a lucky guess. It would have to be repeated many times to demonstrate proper skill. And any audible difference that may have been there for which we have no proof.

So please don't compare these kinds of anecdotes with proper training and skill. They are not the same. There is a lot of rigor in what I do, even when the test are not formal. They are not lucky guesses.
 

PeteL

Major Contributor
Joined
Jun 1, 2020
Messages
3,303
Likes
3,846
"With all due respect" :), no way, no how you can do that. Above was a lucky guess. It would have to be repeated many times to demonstrate proper skill. And any audible difference that may have been there for which we have no proof.

So please don't compare these kinds of anecdotes with proper training and skill. They are not the same. There is a lot of rigor in what I do, even when the test are not formal. They are not lucky guesses.
Well you are right, that’s why I mentioned it was anecdotal and not trying to demonstrate a theory, I was not trying to put forward my « skills ». I also fail to mention that this anecdote was more than 20 years ago. It could be demonstrated, according to your own metrics, that amps were not transparent back then, but sure, it could be luck, maybe... But could you guide us what is « proper » training and skill?I never doubted that you are. And you could even arguably be the most, but not the only. When saying to someone, my subjective evaluation matters, yours don’t, it also mean, I am trained and skilled, you are not, because that’s all there is if both subjective evaluations don’t follow proper science protocol.

That said, you shouldn't take this personal, and I believe my exemple was poorly chosen, but I felt it wasn't fair to completely dismiss the contributor opinion, based on the fact that long term evaluation aren't reliable. My opinion is that they are. The fact that we like blind processes that removes bias in evaluation, shouldn't make us forget the rest of the stuff we know about science. There is a little thing called sample size when it comes to evaluate a behavior. Here's a better exemple. I have two different systems which I use for music listening. Don't you think that if I have been listening to both for 10 years, would it be remotely possible that I am more apt to judge which of the two sets of speaker I prefer, after comparing them for that long, with many type of music, at different SPLs? Is this evaluation is rigorously less reliable than short blind testing? I'm not necessarily saying I'm more apt to judge the one with less distortion, but to judge which one perform better for my need, my and that's what matter no? Yes I've compared them blind, for fun, they are too different to mistake, but objectively, common sense, I think we can say it's quite rigourous to believe that the odds of mistaking them would have been higher on day zero, than on day 3000? You disagree with this?
 
Last edited:

radioman

Member
Audio Company
Joined
Oct 29, 2020
Messages
70
Likes
97
Location
Chicago
Amir, have you received CEntrance's data yet?
Note: In the US, many people, including our staff, had a longer weekend due to the Thanksgiving Holiday. Work resumes on Monday. We have not forgotten and will continue as we return to the office.
 

radioman

Member
Audio Company
Joined
Oct 29, 2020
Messages
70
Likes
97
Location
Chicago
A quick End-Of-Day update:

- We have not received the measured unit from Amir today. If it doesn't arrive tomorrow, we will ask for a tracking number in order to track it.

- We ran a bunch of tests today. We should have all needed measurements tomorrow and then we plan to present a comprehensive argument for re-evaluation.

- We regret that the review didn't isolate noise performance away from SINAD, which is a compound measurement, combining noise and THD into a single number. Amir can easily turn off the distortion and just publish the noise graphs (he has done that in the past). This would present HiFi-M8 in much better light. We plan to advocate this in our proposal, as one of the recommendations.

- Our customers requested pitch-black noise floor because hearing noise in the BAs (the most sensitive of IEMs) is more annoying than hearing distortion. When you hear background hiss it can make it harder to go to sleep to music, so we worked hard to satisfy these people. We are proud of what he have achieved in low noise, in addition to also featuring high power output.

- This may create a point of contention here, but we feel that at lower THD levels, noise is a much bigger issue than distortion, because noise is a rather obvious distraction. You hear it and it makes it hard to go sleep. Distortion, especially second order harmonics, doesn't work the same way. It's much less obvious and therefore, less annoying. And yes, many of our customers use our products on a bed stand. This may not be your particular use case, but all people are different. We respect diversity.

- We found a big "faux pas" with regards to the Windows ASIO driver. The driver on our website is an older version (8.0.2). It was released a year before HiFi-M8 started shipping. Guess what -- It actually doesn't include HiFi-M8 in the list of supported devices! No wonder this caused Amir so much frustration. We apologize that it slipped from our radar. Since Windows 10 already includes native support for USB audio, nobody had brought this to our attention up until now. Again, sorry for the omission. We have fixed the description on our website accordingly.

- We have been focused on our Version 9 ASIO driver, which is currently in its final beta test. It has lower latency, many other improvements, and it definitely supports HiFi-M8. We will announce it soon.

That is all for today. It's 11:30pm. More tomorrow.
 
Last edited:

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,403
Likes
18,364
Location
Netherlands
- We regret that the review didn't isolate noise performance away from SINAD, which is a compound measurement, combining noise and THD into a single number. Amir can easily turn off the distortion and just publish the noise graphs (he has done that in the past). This would present HiFi-M8 in much better light. We plan to advocate this in our proposal, as one of the recommendations.

SINAD literally means SIgnal, Noise, And Distortion. Of course it includes distortion!. Most of us look a bit furtherer than to just the number. The spectrum plot clearly shows the noise floor as well, which quite frankly is not that special either. And why would we change out whole scoring system just to make the M8 present itself in better light? Why don't we let every manufacturer dictate how we should score the devices under test... That would be a fun exercise..

- This may create a point of contention here, but we feel that at lower THD levels, noise is a much bigger issue than distortion, because noise is a rather obvious distraction. You hear it and it makes it hard to go sleep. Distortion, especially second order harmonics, doesn't work the same way. It's much less obvious and therefore, less annoying. And yes, many of our customers use our products on a bed stand. This may not be your particular use case, but all people are different. We respect diversity.

Why not have both? Is it really a tradeoff? Did you intentionally design the thing to have nigher THD so that noise performance is better? I don't think anyone will dispute the audibility (or better lack of) of the distortion shown here, and that is not even the point. Fact remains that it lands you in the bottom tier of all tested devices (at least for the currently tested device), which in itself should tell you something.

BTW, if you look at the low level scores, THD+N even drops to -43 for the low gain setting. Still think that is not an issue?
 
Last edited:
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,667
Likes
241,033
Location
Seattle Area
- We have not received the measured unit from Amir today. If it doesn't arrive tomorrow, we will ask for a tracking number in order to track it.
Oh, I thought the owner had given you the tracking code. I got notification that UPS had delayed delivery by a day. It says it will be delivered on 12/1 by 9:00pm.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,667
Likes
241,033
Location
Seattle Area
- We regret that the review didn't isolate noise performance away from SINAD, which is a compound measurement, combining noise and THD into a single number. Amir can easily turn off the distortion and just publish the noise graphs (he has done that in the past).
I actually ran the test but due to truncation in the driver path, I did not want to post it as it is limited by the dynamic range of 16 bit data path:

CEntrance HiFi-M8 V2 Measurements USB Dynamic Range.png


With the proper ASIO driver, I can test this.
 

radioman

Member
Audio Company
Joined
Oct 29, 2020
Messages
70
Likes
97
Location
Chicago
With the proper ASIO driver, I can test this.
Right. Artificial truncation in software to 16-bit definitely obscures actual hardware capability. We plan to send you the driver that supports the product so you can measure noise correctly.
 

radioman

Member
Audio Company
Joined
Oct 29, 2020
Messages
70
Likes
97
Location
Chicago
Oh, I thought the owner had given you the tracking code. I got notification that UPS had delayed delivery by a day. It says it will be delivered on 12/1 by 9:00pm.
Looking forward to it. Thanks for the update.
 

SIY

Grand Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
10,511
Likes
25,350
Location
Alfred, NY
With the new Version 9 driver, have you fixed the problem of the CEntrance ASIO driver crashing AP software? And crashing ARTA?
 

radioman

Member
Audio Company
Joined
Oct 29, 2020
Messages
70
Likes
97
Location
Chicago
UPS says it was delivered earlier today.
Yes, we have it (thank you). By the time the box got to the engineers, the Audio Precision was already occupied with other tests, so we will work to check it out tomorrow.

Meantime we have one graph to share today -- a noise spectrum plot for HiFi-M8 V2. This was measured on the "cool side" of the unit and it relates to what we have referred to as HiFi-M8's "low noise floor". We compared this graph to your measurements of the Topping DX3 Pro, which has your endorsement, and we believe that HiFi-M8 V2 performs favorably in this context.

Let us know your opinion about this performance and whether it makes sense to re-test this on your side. We feel that including noise measurements would provide a more comprehensive overview of the product. Note that this measurement was made while the unit was running on battery power. Therefore, a connected request is to include HiFi-M8 in the list of portable DAC/Amp products, which you maintain as a separate list from the desktop DAC/Amp products on your site.

Let us know if you have any questions or comments. We will continue our research and plan to share additional info later this week.
M8V2-Cool-Amp-LowGain-Jitter-and-Noise.jpg
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,667
Likes
241,033
Location
Seattle Area
Let us know if you have any questions or comments. We will continue our research and plan to share additional info later this week.
An FFT is not instructive by itself without further documentation of the FFT length and bandwidth. Unless you match my parameters exactly, the graphs cannot be compared.
 

radioman

Member
Audio Company
Joined
Oct 29, 2020
Messages
70
Likes
97
Location
Chicago
An FFT is not instructive by itself without further documentation of the FFT length and bandwidth. Unless you match my parameters exactly, the graphs cannot be compared.
Today we measured at 32k and 16 averages. If you feel that 256k would provide a dramatic change to the results, we can remeasure.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,667
Likes
241,033
Location
Seattle Area
Today we measured at 32k and 16 averages. If you feel that 256k would provide a dramatic change to the results, we can remeasure.
The FFT in the dashboard uses 32K, 22.4 kHz bandwidth and 3 averages.

Importantly, it is NOT in dbV. But rather, the 0 dB is set to reference so that the noise floor is always consistent and can be read relative to that:

1606893256846.png


Your graph is at -18 dBV which means you need to visually raise the noise floor by the same amount to make it comparable to above graph.

Also, please select thicker traces as I have done above so everything is easy to read.
 

radioman

Member
Audio Company
Joined
Oct 29, 2020
Messages
70
Likes
97
Location
Chicago
The FFT in the dashboard uses 32K, 22.4 kHz bandwidth and 3 averages.

Importantly, it is NOT in dbV. But rather, the 0 dB is set to reference so that the noise floor is always consistent and can be read relative to that:

View attachment 96899

Your graph is at -18 dBV which means you need to visually raise the noise floor by the same amount to make it comparable to above graph.

Also, please select thicker traces as I have done above so everything is easy to read.

When do you select 12kHz vs. 1kHz? Referring to this plot:

1606893851951.png

It is late so will discuss more with the team tomorrow. Thank you for your replies so far.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,667
Likes
241,033
Location
Seattle Area
Top Bottom