or Connect
Head-Fi.org › Forums › Equipment Forums › Dedicated Source Components › USB to SPDIF converters shoot-out : EMU 0404 USB vs. Musiland Monitor 01 USD vs. Teralink-x vs. M2Tech hiFace
New Posts  All Forums:Forum Nav:

USB to SPDIF converters shoot-out : EMU 0404 USB vs. Musiland Monitor 01 USD vs. Teralink-x vs. M2Tech hiFace - Page 76  

post #1126 of 1712
Thank you for the respond. Between Musiland 02US and Hiface, which one in your opinion have the best sound quality? Or if there are other better option under $200?

Quote:
Originally Posted by Shahrose View Post
Sorry, I don't use Winamp anymore. I consider Foobar a lot more suitable for my needs and it also supports KS/WASAPI/ASIO/ASIO4ALL. In foobar, the HiFace unit works well with KS but is glitchy with WASAPI, and the 02 works equally well with both KS/WASAPI. BTW, I prefer the sound of WASAPI over KS, as subtle as the difference between them is.
post #1127 of 1712
Quote:
Originally Posted by Shahrose View Post
Testing the Musiland 02US before burn-in right now. I think if you already use this as a transport, there's no point getting the HiFace as the Musiland is not any worse overall IMO. The HiFace has the best bass quality of all the transports still, even after it got a bit looser and more plentiful after burn-in, though the Musiland is close. The Musiland has a better soundstage and its imaging sharpness and details are almost on par with the HiFace. I think the HiFace does a bit better in complex passages but the lack of harshness/in-your-face soundstaging makes the Musiland better for listening over the long run IMO. It has better bass quality and less smearing than my motherboard output, but less bass quantity than all transports.

I'm also impressed by the DAC of the 02. Not as good as the y2, but amazing for a $130 item. I might buy another one to give to family/friends. The amped output is decent too, although lacking in control (not volume) when compared to higher-end amps.

These findings may change after burn-in. So far my impression is positive, especially for the price paid.

I still think these reports may point to anomalies in the Hiface jitter spectrum. I know Slim is very high on the 1nS jitter spec but if all that jitter is under the digital recievers PLL cut-off point (unlikely) it could point to an issue. My guess is that the power supply upgrade would fix this problem. My point is Jitter can't be summed with a single number, you have to look at what frequencies the jitter is peaking at. A good transport will have the jitter spread almost equal across 20-20khz, this is much less audible than jitter centered at say 300hz.
post #1128 of 1712
I should be getting the Hiface in today (thanks Sharose!). I expect it to match well with my warm DA100 -> WA6 setup, but we'll have to see. I'll try to compare it to the only other USB/SPDIF transport I've tried so far, the uDAC.
post #1129 of 1712
I think the Musiland and Hiface are both good implementations. On paper the Hiface uses separate sample-rate-multiple clocks and the Musiland generate those clocks from a single 24Mhz crystal. It could be argued that the Hiface is a "better" implementation. However, after the signals are processed by the FPGA to generate SPDIF the resultant jitter may or may not be much different between the implementations (I have not seen jitter comparisons between the two yet).

It also depends on what kind of DAC you are using. If you use something like Sabre32, then whatever original jitter is minimized as the DAC has an internal ASRC and the original clock is thrown away.

And then there is also the "fun factor". If you like to tinker with the device but do not have the skills of Jkenny, then the Musiland is much easier to mod.
post #1130 of 1712
Thread Starter 
Quote:
Originally Posted by regal View Post
I still think these reports may point to anomalies in the Hiface jitter spectrum. I know Slim is very high on the 1nS jitter spec but if all that jitter is under the digital recievers PLL cut-off point (unlikely) it could point to an issue. My guess is that the power supply upgrade would fix this problem. My point is Jitter can't be summed with a single number, you have to look at what frequencies the jitter is peaking at. A good transport will have the jitter spread almost equal across 20-20khz, this is much less audible than jitter centered at say 300hz.
I think that except for the last few pages, the majority of the people described the hiface as being an excellent transport and better than CD based and computer based transports.

However, at the same time, some people like Shahrose and h.rav didn't like the sound of the hiface that much.

I don't think that the distribution of the jitter spectrum is the cause. I am pretty confident that if we measure it on a good computer, it will give better figure than CD spinners or other soundcards (even on the low frequencies).
I have thought about what might be the origin and I have came up with a few ideas/speculations:

1. Type of the DAC: upsampling vs. non upsampling, de jitter functions, DSP...
2. type of computer: motherboard, usb sockets...
3. Power supply to the computer
4. Vibration control of the computer

Point 1 - the type of the DAC
When people use any transport with a DAC that has an ASRC or a DSP like the one used in the audio-gd ref-1, results are unpredictable. If you read the specs and understand how they work, the jitter of the source should not be relevant at all since they don't treat the data in real time but they buffer it.
If you look at jitter measurements of the Benchmark DAC1 which uses an ASRC to "reclock" the data, you will see that even high amounts of jitter at any frequency do not affect the jitter measurements taken by an Audio Precision system 2 at the analog outputs. So any jitter in the Benchmark DAC1 would be generated after the ASRC. As long as we have bit perfect data, there should be no measurable difference in jitter figures whit most usb deviced (even the poorly constructed ones).
So if we assume the AMB Y2' ASRC is properly built (which I think it is), it wouldn't show any measurable difference in jitter figures regardless of transport.
The same thing apply to DSPs such as the one used in the Audio-gd Ref1 and audio-gd DAC19dsp. In theory, the transport shouldn't matter as long as it is bit perfect.
However, listening tests seem to indicate that there are other factors because different transports and digital cables affect the sound of thoses DACs.
Funny thing, Kingwa (the designer behind the audio-gd gear) encourages people using a high quality transport despite the fact that theory and some measurements say it shouldn't matter.
To sum up, I don't think we should make assumptions about the jitter spread of a transport when listening to such equipment. At least, in my opinion, it is not a safe assumption to speculate that by miracle all the jitter of the hiface is centered at 300hz. If that was the case, everyone who has listened to the hiface wouldn't have liked it.
I am not saying that the hiface is perfect, but it is highly unlikely that it has some highly unusual spread of jitter just because a few isolated people didn't like it their system.
So far, the only measurement we have is the 1ns figure and it is a good one and speculating that the hiface has an exceptionally weird jitter spectrum is just trying to find an excuse to comply with the listening experience of a few people.
I think that by doing that, we are looking at the problem from the wrong side, I think that there are a other factors that affects the sound of a transport beside the jitter figure.

So what are those factors? Honestly I don't know. Even Kingwa (the designer of audio-gd) acknowledges that there are differences between transports besides the jitter measurement. But he didn't go into too much details about it.
Point 2/3 & 4 are just based on my own listening experience: all of them made noticeable changes to the sound of the hiface. So by proxy, I am assuming that different opinions on the hiface (a lot of people liking it and a few not so much) are partly due to those points.

Point 2 - different computers
I have tried the hiface on 2 different notebooks connected to the same dac with the same digital cable and it sounded slightly different. (The emu 0404 usb showed however more differences between the 2 notebooks).
Also, it has been reported by many that different usb sockets sound differently.

Point 3 - Power supply
The power supply to the computer is important. I am plugging my notebook to a separate power filter than the rest of my chain. I have also noticed than running the notebook from battery changed the sound.

Point 4 - Vibration control
I never thought a computer performance in streaming audio would be affected by vibration. It happened to me by accident when I was switching the place of my notebook from a glass shelf (in my living room) and I put it temporarily on my wooden floor. The sound changed and it took me a while to understand what was happening.
I will spare you the details of I tried after but I have finally settled on Herbies's Audio Tenderfoot under the notebook. I am also using a thick acrylic shelf (instead of glass).
Those small tweaks made a big and noticeable improvement on the hiface. I have had a lot of friends laugh at the decoupling of the notebook when they first saw it. But usually, as soon as they listen to the sound, they forget about those weird "tweaks".

So my speculation is based on how the sound changed with those tweaks. I can imagine (or even create) a situation where the hiface is plugged in the wrong usb socket, where the power supply to the computer is not (or is poorly) filtered and where it suffer from a lot vibration coming from the computers.

Contrary to many other converters, the hiface is directly "hard wired" to the computer which must make it a lot more sensitive to some of the things I described.
post #1131 of 1712
Thread Starter 
Quote:
Originally Posted by glt View Post
I think the Musiland and Hiface are both good implementations. On paper the Hiface uses separate sample-rate-multiple clocks and the Musiland generate those clocks from a single 24Mhz crystal. It could be argued that the Hiface is a "better" implementation. However, after the signals are processed by the FPGA to generate SPDIF the resultant jitter may or may not be much different between the implementations (I have not seen jitter comparisons between the two yet).

It also depends on what kind of DAC you are using. If you use something like Sabre32, then whatever original jitter is minimized as the DAC has an internal ASRC and the original clock is thrown away.

And then there is also the "fun factor". If you like to tinker with the device but do not have the skills of Jkenny, then the Musiland is much easier to mod.

It is not true that the clock of the hiface has to go through the FPGA. The hiface uses true audio clocks which means that jitter in the data stream is directly dependant on the jitter value of the clocks and not the FPGA.
While it would be a waste to put ultra low jitter clocks in the Musiland (since the signal would have to go through the FPGA to generate some frequencies as you said), it would be more intersting to put them on the hiface.
We don't have jitter measurements of the Musiland but some high end DACs couldn't even lock properly to the 44.1 frequency which was generated by the FPGA. When a digital receiver cannot lock, it means very high jitter levels. have a look at the spdif output of the Musiland here: Computer Audio Asylum - Musiland Monitor 01 USD - late to the party, but... - Thorsten - December 17, 2009 at 10:35:40
post #1132 of 1712
Quote:
Originally Posted by duckymcse View Post
Thank you for the respond. Between Musiland 02US and Hiface, which one in your opinion have the best sound quality? Or if there are other better option under $200?
I would say the HiFace is a bit better, everything taken into account. The only area where the Musiland bests it is soundstaging/smoothness, which just happens to be important for me.
post #1133 of 1712
Quote:
Originally Posted by slim.a View Post
It is not true that the clock of the hiface has to go through the FPGA. The hiface uses true audio clocks which means that jitter in the data stream is directly dependant on the jitter value of the clocks and not the FPGA.
The HiFace clocks are selected and divided inside the Xilinx CLPD, not a PGA.

Quote:
While it would be a waste to put ultra low jitter clocks in the Musiland (since the signal would have to go through the FPGA to generate some frequencies as you said), it would be more intersting to put them on the hiface.
Still interesting to improve the clock of the Musiland IMO, even though it is synthesized.

Steve N.
Empirical Audio
post #1134 of 1712
Yea, this whole area of computer sound is finally getting some serious attention & hopefully the various factors that effect the sound will be gradually teased out.

Trying to tease this out, at the moment, is hampered by the number of different parameters that are impinging on the final sound. In my opinion the only way to begin to address this is to try to minimise the number of variables.

That's one of the reasons I did my modded USB cable - it gave me a way to test the sonic effect that the 5V PS has in isolation from the PCs PS. Now the Pcs PS may well still have an effect but it gives us a better tool perhaps for teasing this out.

I really haven't had any time to do these tests in a formal way as slim.a does (I've been too busy & too lazy).

My take on the Musiland Vs Hiface is that I've listened to both (& modified both ). I found the Hiface to sound much better than the Musiland 01 US. When I modded the Musiland as far as I could take it, I reckoned (not a scientific term, agreed ) that it sounded like a stock Hiface. This was not a very controlled test & is simply my opinion on my system.

Anyway, technical issues & driver gymnastic issues eventually put me off the Musiland but I still think it's great value for money.
post #1135 of 1712
Quote:
Originally Posted by audioengr View Post
The HiFace clocks are selected and divided inside the Xilinx CLPD, not a PGA.



Still interesting to improve the clock of the Musiland IMO, even though it is synthesized.

Steve N.
Empirical Audio
Tried it - didn't make any noticeable difference to my ears, etc
post #1136 of 1712
Thread Starter 
Quote:
Originally Posted by jkeny View Post
My take on the Musiland Vs Hiface is that I've listened to both (& modified both ). I found the Hiface to sound much better than the Musiland 01 US. When I modded the Musiland as far as I could take it, I reckoned (not a scientific term, agreed ) that it sounded like a stock Hiface. This was not a very controlled test & is simply my opinion on my system.
I had similar results with the Teralink-X. Once I upgraded the usb cable (wireworld ultraviolet), used better ASIO drivers (Ploytec), and run the notebook from battery, the sound of the Teralink got closer to the sound of the stock hiface (but still not on the same level).
That is one of the reasons I have repeatedly said the hiface was clearly a better and more accurate usb converter than the other ones I have tested.

Also, once I started using the hiface, it was a lot easier for me to pick sonic differences between digital cables and between recordings.

In fact, it is astonishing the differences in sound quality you can get between different CDs. When I talk about soundstage or timber of instruments, I use recordings from Linn Records, HDtracks, 2L, ... that were recorded properly.
To give an example, let's say that my headphones' maximum soundstage size is 100.
With the Teralink-X, I would always have a "big" soundstage that is always between 70 and 90.
With the Hiface, the soundstage can change from 50 to 100 depending on the recording itself. But in fact, the 100 figure doesn't give an accurate description. Besides the big size in well recorded material, you get a "being there" feeling with the hiface that you simply cannot get with lesser converter because they can't dig deep in the (good) recordings.

The same can be said about the tonal quality: While the Teralink-X is pleasing to the ears most of the time, it simplifies good quality recordings.

Personally, when I listen to a Mahler's symphony or to Norah Jones, I don't want to have the same experience. When choosing a colored converter, the risk is to have everything sounding the same but it is sometimes what some people might be looking after in a transport.

Speaking of "colored" converters, I received the Teralink-X2 yesterday and have liked what I heard so far. While waiting for the dac19dsp (and for the hiface back), I tried different DACs, usb converters and different digital cables (optical and coaxial) and no combination sounded musically satisfying. That was until yesterday when I tried the Teralink X2 with the Audio-gd FUN.
So far, the following chain: Foobar --> SoX upsampling to 96K --> Wireworld Ultraviolet --> Teralink X2 --> Oyaide DB-510 --> Audio-gd FUN (version A, Fet output) gave me good subjective results but not nearly as transparent or satisfying as my former reference (hiface+dac19mk3).
post #1137 of 1712
Quote:
Originally Posted by slim.a View Post
Speaking of "colored" converters, I received the Teralink-X2 yesterday and have liked what I heard so far. While waiting for the dac19dsp (and for the hiface back), I tried different DACs, usb converters and different digital cables (optical and coaxial) and no combination sounded musically satisfying. That was until yesterday when I tried the Teralink X2 with the Audio-gd FUN.
So far, the following chain: Foobar --> SoX upsampling to 96K --> Wireworld Ultraviolet --> Teralink X2 --> Oyaide DB-510 --> Audio-gd FUN (version A, Fet output) gave me good subjective results but not nearly as transparent or satisfying as my former reference (hiface+dac19mk3).
Does the X2 improve upon the first version? I've only heard the X2.
post #1138 of 1712
Thread Starter 
Quote:
Originally Posted by Shahrose View Post
Does the X2 improve upon the first version? I've only heard the X2.
Yes, according to the listening I have done so far, the X2 improves indeed on the X1. And since it adds the 24/96 capability compared to the 16/48 of the X1, it is a no-brainer choice for someone hesitating between the slightly cheaper X1 and the X2. The slightly higher price is worth it in my opinion.

However, the only problem I encountered so far was playing native 24/88 files. Did anyone verify it could play 24/88 without resampling?
post #1139 of 1712
Quote:
Originally Posted by slim.a View Post
However, the only problem I encountered so far was playing native 24/88 files. Did anyone verify it could play 24/88 without resampling?
I play 24/88 files with my X2 without any noticeable problems. However I don't know, whether there arrives a 24/88 or a re-/upsampled 24/96 stream
at my DAC (since my DAC hasn't a display of the choosen samplerate).

Kind regards
Fujak
post #1140 of 1712
Thread Starter 

Teralink X2's software seems to resample 88.2 to 96K

Quote:
Originally Posted by Fujak View Post
I play 24/88 files with my X2 without any noticeable problems. However I don't know, whether there arrives a 24/88 or a re-/upsampled 24/96 stream
at my DAC (since my DAC hasn't a display of the choosen samplerate).

Kind regards
Fujak
Hi Fujak,

Here is what happened to me:

I set up the SoX resampler in Foobar to upsample everything to 96K except for 88.2 files and 96 files. Everhting played fine until I played a native 24/88 file from Linn Records. It played fine but last time I played it with the Musiland it sounded much better.
So I deactivated all DSP plugins and the file still didn't sound good.

Up to that point my assumptions were:
a) There was some resampling going on specifically on the 88.2 files
b) The clock was poorly generated for 88.2K frequencies
c) a little of both a) and b)

Then I checked the control panel of the Teralink X2 and it showed 24/96 while playing different 24/88. So I got even more suspicious. The control panel showed correctly the 44.1, 48 and 96 frequencies but not the 88.2.

Then when you talked about having a DAC which displays sample rates, it made me remember that the control panel of the emu 0404 usb had that function: it displays the digital input sample rate.

So finally I repeated the test with Emu 0404 usb and it confirmed what I suspected: 44.1, 48 and 96 are played fine but not 88.2 which is resampled to 96.

Overall, it is not a big problem for me but I thought it was worth mentioning it. For now I have set the SoX upsampler to upsample everything to 96K and the SoX plugin does sound better than the software upsampling that seems to be going on with the Teralink X2.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Dedicated Source Components
This thread is locked  
Head-Fi.org › Forums › Equipment Forums › Dedicated Source Components › USB to SPDIF converters shoot-out : EMU 0404 USB vs. Musiland Monitor 01 USD vs. Teralink-x vs. M2Tech hiFace