HiFace, sensitive information
Jul 21, 2010 at 11:07 AM Thread Starter Post #1 of 425

shamu144

1000+ Head-Fier
Joined
Jan 5, 2007
Posts
1,151
Likes
38
I am myself a happy owner of a HiFace, bought back in february this year. I have nothing but the highest praise for this great little piece of gear.
 
However, recently and after easter, more and more users having just bought a HiFace raised some worries about its sonic performance, one of them beeing Chris Connaker, founder of computeraudiophile. It seems as if the HiFace did not perform as expected and created some tensions between HiFace users.
 
Regal - who was not overwhelmed with his HiFace performance - first pointed out in this forum that the 22.592Mhz clock in its HiFace was not the same as the original HiFace batches with larger MEC clock, and that it could have been a reason explaining the differences in opinions among HiFace users:
 
http://www.head-fi.org/forum/thread/446375/usb-24-192khz-m2tech-hiface/1215#post_6708753
 
Because a friend of mine was not happy with the performance of his HiFace, I recomended him to open up his HiFace and he discovered that his unit indeed came with a small oscillator as shown in Regal's picture above. See picture of my friend's Hiface below.
 

 
To avoid any doubt, he sent me over his HiFace unit and I have been able to A/B them with my own stock HiFace, bought back in february, with the larger MEC oscillator.
 
I had the opportunity to compare myself in my own system a "small clock" HiFace unit with my  "original MEC clock" unit, and there is no doubt it was indeed a night and day difference, the "small clock" performing very poorly in terms of perceived sonic qualities (thin, harsh sounding, no soundstage, etc...). I then sent him over both HiFace so he could also test them with its own system, and came to the exact same conclusions. I think Regal also had the opportunity to do the same comparison, and experienced the same results.
 
I then contacted Marco at M2Tech to get some clarifications, and he explained that the original oscillator supplier - MEC - was not able to cope with the demand and they experienced a shortage of oscillators. They did look for a "suitable" substitute in the form of the smaller oscillator and did produce a few batches, before switching back to the larger and original MEC oscillators. Hiface are now shipping with the MEC oscillator, as confirmed by Marco.
 
My point is that they seem to be some bad apples among HiFaces currently on the market. And apparently, the only way to find out wether your HiFace is affected or not is to open the device, or by listening... If you find your HiFace just sounds regular compared to your former transport, chances are that you have a bad one... The HiFace when working properly is trully a fantastic USB/SPDIF converter, and you can tell it from the very first moment, at least in my experience, and that is without mods...
 
In case your Hiface is coming with the smaller oscillator, I strongly recommed you to contact Marco at M2tech to look for the best solution to this issue. In my opinion, users should not be charged for parts or labor costs, since no comunication was done by the manufacturer on this specific matter and substitution costs should be their responsability.
 
I hope this information was usefull to some of you... For the rest who are already enjoying their HiFace, well, keep it that way :wink:
 
 
EDIT 08/31/2010: How to determine wether your Hiface is affected without opening the device and voiding the guarantee.
 
At this stage, I still believe the best way to determine wether you have or not a small clock unit is to be able to A/B with 44.1khz material your Hiface unit with another HiFace that you know for sure has a large clock. Differences will show up immediately if any because they are very noticeable. But of course, very few will have the opportunity to test that.
 
That is why an alternative method is the use of the SoX resampler in Foobar, said to be one of the best resampler available out there. That is what I have done earlier, following Xdanny suggestion, and I find it a very valid method.
 
Just chose some 44.1 material to play, and listen A. Then, load the resampler SoX in Foobar DSP panel to resample at 48 and listen B. Repeat as much as you want. Oh BTW, the Lavry allows to display the incoming sample rate, and I have been able to check that the resampler was indeed at work when selected. If A sounds better than B (and it does, even though someone said that the SoX resampler was absolutely transparent, it is not to my ears and in my system), you have a large clock. But if B sounds better than A (and it does as well), then you obviously have a small clock at 44.1khz.
 
I have done this experiment with both unit of the HiFace in my possession, the small clock unit and the large clock unit. I choosed Diana Krall, The Look Of Love, S'wonderful, for I find it a good recordings technically speaking with Diana Krall's voice very easy to isolate, and an accompanying orchestra.
 
With the small clock unit, the resampled version at 48khz sounds better, with Diana Krall's voice fuller and richer, with more weight and better articulation. You will also notice that the violins are more realistic sounding, with better pace and with details and intonations that you simply miss at 44.1khz.
 
With the large clock unit, the resampled version at 48khz sounds worst than at 44.1khz by a fair margin. Diana Krall's voice sounds thinner, more aggressive, lacking emotion. Violins in the background will make you feel unconfortable, as they stop abruptely, unnnaturally, lacking decay.
 
I shall say that with both units, the resampled version at 48khz sounded identical to my ears. So it really goes like this:
 
large clock 44.1 > large & small clock resampled SoX 48 > small clock 44.1
 
 
EDIT 08/31/2010: users reports
 
Users so far who have reported relevant differences either in A/B two different units, or through the SoX resampler tests:
 
Regal, Shamu144, Picodeloro, USG


 
Jul 21, 2010 at 11:42 AM Post #2 of 425
I am also a happy owner of a (modded) hiface but I hope that m2tech can solve the problem (free of charge) for those who are stuck with such units.
 
This issue might explain why there were so many disagreements about the sound of the hiface. People were just describing (almost) 2 different devices.
 
Jul 21, 2010 at 12:36 PM Post #4 of 425
Robscix, I think Shamu DID A/B the two versions?
 
The clock in the photo is the 22.5792MHz which is responsible for audio at 44.1KHz, 88.2KHz & 176KHz - Shamu, can you or did you check if the 48KHZ, 96KHz & 192KHz sounds different between the two Hiface units? Obviously as they are both operating off the same MEC clock (24.576MHz) for these speeds they should sound the same. Similarly, 48KHz audio should sound better than 44KHz audio on the device with the small clock. Just trying to tease out that the clock is the culprit before going any further.
 
I did notice the change of clocks a while back & spoke to Marco but he told me these were just as good. I also listened & heard no difference but maybe my listening wasn't as critical or maybe I listened to the unit after it was modified, I can't remember now.
 
One other thing strikes me & maybe Regal can come in here - did Regal change this clock or just the PS to it & now is very happy with the sound?
 
Jul 21, 2010 at 12:42 PM Post #5 of 425

 
Quote:
Robscix, I think Shamu DID A/B the two versions?
 
The clock in the photo is the 22.5792MHz which is responsible for audio at 44.1KHz, 88.2KHz & 176KHz - Shamu, can you or did you check if the 48KHZ, 96KHz & 192KHz sounds different between the two Hiface units? Obviously as they are both operating off the same MEC clock (24.576MHz) for these speeds they should sound the same. Similarly, 48KHz audio should sound better than 44KHz audio on the device with the small clock. Just trying to tease out that the clock is the culprit before going any further.
 
I did notice the change of clocks a while back & spoke to Marco but he told me these were just as good. I also listened & heard no difference but maybe my listening wasn't as critical or maybe I listened to the unit after it was modified, I can't remember now.
 
One other thing strikes me & maybe Regal can come in here - did Regal change this clock or just the PS to it & now is very happy with the sound?

Yes but what are we talking about, a quick listen?  What type of testing was done, any measurements?
Companies use different components that are identical in function, this happens all the time.  I bought one awhile back, so heres hoping I got the good version.  Everything sounds good when I have used it with my DAC's.
 
 
Jul 21, 2010 at 1:17 PM Post #6 of 425
No problem finding the datasheet for the MEC part, so if someone could share the manufacturer and/or identifying number for the smaller part it would help for an initial evaluation before moving forward.

Also, could someone give a shot or two of the plastic parts that need to come apart so we can see how they fit together/ come apart?
 
Jul 21, 2010 at 1:29 PM Post #7 of 425


Quote:
No problem finding the datasheet for the MEC part, so if someone could share the manufacturer and/or identifying number for the smaller part it would help for an initial evaluation before moving forward.

Also, could someone give a shot or two of the plastic parts that need to come apart so we can see how they fit together/ come apart?

I think both the MEC & other oscillators are custom orders so I don't know about datasheets. Anyway, what will the datasheet tell us about the sound? I believe you are better off doing the comparison tests I outlined to check if it truly is the oscillator that is the issue with the sound difference.
 
You won't be able to take the plastic apart without some breakage of it (depending on how well it is stuck together) - the base is tuck to the rest of the case.
 
 
Jul 21, 2010 at 3:20 PM Post #8 of 425
Beyond sound, I wanted some specs to find a suitable replacement.
 
wrt datasheets, I wanted to compare some other oscillators and to see if they have the same pinout, power consumption, and RMS jitter(jitter is something omitted from the MEC datasheet btw). It can wait until I get home when I will see exactly what part my HiFace has and take some actual frequency and voltage measurements.
 
The first oscillator that comes to mind would be the Crystek used in the Gamma2, which I already have in the parts bin, and appears at first glance to spec very well.
 
Jul 21, 2010 at 4:37 PM Post #10 of 425
Gentlemen, rather than dragging this thread in a sterile debate over measurments, specs or datasheets, I believe it would be a lot more constructive for the community to actually hear feedback of users who have been able to A/B both versions of the HiFace, in its stock form.
 
The "bad" HiFace can be easily recognized as it sounds much thinner in the mids, especially on voices with lot's of grain, highs show traces of harshness, leading to an overall agressive sound, with a poor separation of instruments and a collapsed soundstage. The fantastic continuity and smoothness than the original HiFace can provide is simply non existent. It reminded me in some ways the Musiland 01 USD I also had the opportunity to try in my system.
 
So please share with us your subjective listening experience between both HiFace versions, you could be of great help to this community. I actually firmly believe that the HiFace has set a new standard for the USB/SPDIF converters which we all could benefit from if done properly.
 
Jul 21, 2010 at 6:07 PM Post #12 of 425


Quote:
Shamu, I agree & to this end can you do the 44.1KHz Vs 48KHz within & between these different Hiface units?


I remember suggesting this test to Regal in the original thread, but nothing came of it.  I do remember he fitted a battery and suddenly was over the moon about the sound, and of course a battery won't make up for any shortcomings with the oscillator itself.
 
Jul 21, 2010 at 6:27 PM Post #13 of 425
Yes, Dan, I remember you suggesting this too - I don't think that Regal changed the "offending" oscillator, do you? And as you say the PS change would not necessarily compensate the oscillator! Maybe he could say something about it here? 
 
How are those batteries going? I'm just wiring up a Sabre based AckoDAc to run off battery (only 4 at the moment) - should be interesting.
 
Jul 21, 2010 at 10:20 PM Post #14 of 425


Quote:
I think both the MEC & other oscillators are custom orders so I don't know about datasheets. Anyway, what will the datasheet tell us about the sound? I believe you are better off doing the comparison tests I outlined to check if it truly is the oscillator that is the issue with the sound difference.
 
You won't be able to take the plastic apart without some breakage of it (depending on how well it is stuck together) - the base is tuck to the rest of the case.
 


Hi
 
Since you see more HiFaces than anybody else, can you guesstimate when the change back to the large clocks took place? 
 
(For instance, my HiFace was shipped from TweekGeek  May 10th, so in your estimation, did the change back to the large clocks occur before or after that date?)
 
Thanks
 
USG
 
Jul 22, 2010 at 1:00 AM Post #15 of 425
my developer PCBs have the offending smaller clock as well, but I dont have the opportunity to AB with the old. although i'm not sure how meaningful it is for me anyway since I am only using this lot with sabre dacs, which throw out MCLK anyway. all the same I have searched for a suitable substitute, but that clock speed is really quite rare it seems
 

Users who are viewing this thread

Back
Top